Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Tools and applications for DNA sequencing in the clinical management of viral infectious diseases : examples… Brumme, Chanson Joachim 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_november_brumme_chanson.pdf [ 9.35MB ]
Metadata
JSON: 24-1.0166706.json
JSON-LD: 24-1.0166706-ld.json
RDF/XML (Pretty): 24-1.0166706-rdf.xml
RDF/JSON: 24-1.0166706-rdf.json
Turtle: 24-1.0166706-turtle.txt
N-Triples: 24-1.0166706-rdf-ntriples.txt
Original Record: 24-1.0166706-source.json
Full Text
24-1.0166706-fulltext.txt
Citation
24-1.0166706.ris

Full Text

TOOLS AND APPLICATIONS FOR DNA SEQUENCING IN THE CLINICAL MANAGEMENT OF VIRAL INFECTIOUS DISEASES:  EXAMPLES FROM HIV-1 AND HCV  by Chanson Joachim Brumme B.Sc., Queen’s University, 2001 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Experimental Medicine) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) September 2015 © Chanson Joachim Brumme, 2015   ii  Combination antiretroviral therapy has transformed Human Immunodeficiency Virus (HIV) infection from what was once a fatal diagnosis to a manageable chronic condition. Similarly, new direct-acting antivirals offer a potential cure for individuals infected with Hepatitis C Virus (HCV). Treatment of these two infectious diseases is now routinely guided by genotypic drug resistance testing: portions of the viral genome are sequenced and analyzed for mutations in order to select drug combinations best suited to treat each individual’s unique viral population. The primary aim of this thesis is to develop new methods to personalize therapies for HIV and HCV using a variety of DNA sequencing technologies. First, manual review of Sanger sequences is highly subjective, leading to potential bias in the detection of resistance mutations in diverse viral populations. Automated sequence analysis software that provides standardization between users and laboratories is presented. Second, HIV treatment in resource-limited settings is compromised by insufficient access to resistance testing. To facilitate individual-level monitoring, a low-cost resistance test, whereby hundreds of samples are simultaneously sequenced on a next-generation instrument, is proposed and validated. Third, novel screening and drug resistance tests are required to assess the efficacy of new antivirals. For example, certain regimens containing the protease inhibitor simeprevir are less effective in treating individuals infected with HCV harboring a common polymorphism. Two independent sequencing assays that test for this polymorphism are described and validated.  A secondary aim is to measure HIV evolution under immune and drug selection pressures using these same sequencing methods. The hypothesis that effective treatment suppresses viral replication and retards viral evolution is supported by evidence from longitudinal sequences from patients on antiretroviral therapy. A second study demonstrates limited selection of drug resistance mutations in patients with low-level viremia, supporting the hypothesis that many results from an approved clinical test are false positives. Finally, next-generation sequencing is used to quantify HIV variants in cultured virus in order to measure their relative replicative fitness. iii  This thesis provides evidence that new and existing assays and bioinformatic tools will remain invaluable in the clinical management of HIV and HCV as DNA sequencing technologies continue to evolve. iv  Versions of Chapters 2 through 7 have been published as original research in the scientific literature, and/or presented at scientific conferences. The candidate contributed substantially to all five of the published manuscripts and the two unpublished data chapters either as lead (Chapters 3 and 5.2), co-lead (Chapter 2), collaborating (Chapter 5.3), or last author (Chapters 4, 6 and 7). The contents of published manuscripts have been reprinted with permission from their respective journals. This is to certify that the candidate was a major contributor to study design, data collection, and/or supervision of the work presented. The candidate performed the majority of the data analysis and interpretation for all studies presented and wrote or co-wrote all seven manuscripts. The specific contributions of collaborating authors are detailed below. Chapter 2 has been previously published as:   *Woods CK, *Brumme CJ, Liu TF, Chui CKS, Chu AL, Wynhoven B, Hall TA, Trevino C, Shafer RW, Harrigan PR. Automating HIV drug resistance genotyping with RECall, a freely accessible sequence analysis tool. J Clin Microbiol, 2012; 50(6):1936-42.  Copyright © 2013, American Society for Microbiology. All Rights Reserved. doi:10.1128/JCM.06689-11.  *CKW and CJB contributed equally. CJB, RWS and PRH conceived and designed the study. CKW, CJB, ALC, BW and TAH contributed to RECall software development. TFL, CT, RWS provided data. CJB analyzed and interpreted the data. CJB and CKSC wrote the manuscript. All authors participated in manuscript editing.  Chapter 3 has been previously published as:  Brumme CJ, Swenson LC, Wynhoven B, Yip B, Skinner S, Lima VD, Montaner JSG, Harrigan PR. Technical and regulatory shortcomings of the TaqMan version 1 HIV viral load assay. PLoS One, 2012; 7(8):e43882.   v  Reprinted under the terms of the Creative Commons Attribution License. © 2012 Brumme et al. doi:10.1371/journal.pone.0043882.  CJB, VDL and PRH conceived and designed the study. SS provided additional study samples. CJB analyzed and interpreted the data. CJB and LCS wrote the manuscript. All authors participated in data collection and manuscript editing. Chapter 4 has been previously published as:  Knapp DJHF, Brumme ZL, Huang S, Wynhoven B, Dong W, Mo T, Harrigan PR, Brumme CJ. Increasingly successful highly active antiretroviral therapy delays the emergence of new HLA class I-associated escape mutations in HIV-1. Clin Infect Dis, 2012; 54(11):1652-9.   The right to reproduce this manuscript has been granted by Oxford University Press (RightsLink license numbers 3616101491276, 3616110115870) © 2012 Knapp et al. doi: 10.1093/cid/cis253.  DJHFK, ZLB, PRH and CJB conceived and designed the study. DJHFK and CJB performed data analysis. DJHFK, ZLB, PRH and CJB interpreted data and wrote the manuscript. CJB was supervisory author on this project. All authors participated in data collection and manuscript editing. Chapter 5 synthesizes two previously published manuscripts:  Brumme CJ, Huber KD, Dong W, Poon AFY, Harrigan PR, Sluis-Cremer N. Replication fitness of multiple nonnucleoside reverse transcriptase-resistant HIV-1 variants in the presence of etravirine measured by 454 deep sequencing. J Virol, 2013; 87(15):8805-7.   Copyright © 2013, American Society for Microbiology. All Rights Reserved. doi:10.1128/JVI.00335-13.  and  Sluis-Cremer N, Huber KD, Brumme CJ, Harrigan PR. Competitive fitness assays indicate that the E138A substitution in HIV-1 reverse transcriptase decreases in vitro susceptibility to emtricitabine. Antimicrob Agents Chemother. 2014; 58(4):2430-3.   Copyright © 2014, American Society for Microbiology. All Rights Reserved. doi:10.1128/AAC.02114-13.  vi  These two manuscripts were conceived as a single set of experiments. NSC and PRH conceived and designed the studies. Site-directed mutagenesis and viral culture were performed by KDH under the supervision of NSC. CJB and WD performed 454 sequencing. CJB and AFYP analyzed the data. CJB and NSC interpreted results and wrote the manuscripts. All authors participated in manuscript editing. Chapter 6 has been previously published as:  Lapointe HR, Dong W, Lee GQ, Bangsberg DR, Martin JN, Mocello AR, Boum Y, Karakas A, Kirkby D, Poon AFY, Harrigan PR, Brumme CJ. HIV drug resistance testing by high-multiplex “wide” sequencing on the MiSeq Instrument. Antimicrob Agents Chemother. 2015; 53(11) pii:AAC.01490-15 [Epub ahead of print]  Copyright © 2015, American Society for Microbiology. All Rights Reserved. doi:10.1128/AAC.01490-15.  PRH and CJB conceived and designed the study. DRB, JNM, ARM, YB contributed additional samples and clinical data. DK and AFYP contributed data analysis software. HRL and CJB collected MiSeq data and analyzed data, interpreted results and wrote the manuscript. CJB was supervisory author on this project. All authors participated in data collection and manuscript editing. Chapter 7 has been previously published as:  Chui CKS*, Dong WWY*, Joy JB, Poon AFY, Dong WY, Mo T, Woods CK, Beatty C, Hew H, Harrigan PR, Brumme CJ. Development and validation of two screening assays for the hepatitis C virus NS3 Q80K polymorphism associated with reduced response to combination treatment regimens containing simeprevir. J Clin Microbiol, 2015; 53(9): 2942-50.  Copyright © 2015, American Society for Microbiology. All Rights Reserved. doi:10.1128/JCM.00650-15.  *CKSC and WWYD contributed equally. CKSC, CB, PRH and CJB conceived and designed the study. AFYP, CKW and CJB analyzed data. CB and CJB interpreted the results and wrote the manuscript. vii  CJB was supervisory author on this project. All authors participated in data collection and manuscript editing. Ethical approval for these studies was granted by the University of British Columbia - Providence Health Care Research Ethics Board under certificates H04-0276, H04-50276, H07-03006, H08-00962, H10-01778, H11-01160, H11-01642 and/or H13-00395. viii  Abstract ................................................................................................................. ii Preface ................................................................................................................... iv Table of Contents ................................................................................................. viii List of Tables ....................................................................................................... xiii List of Figures ...................................................................................................... xiv List of Abbreviations ............................................................................................ xvi Acknowledgements ............................................................................................ xviii Chapter 1: Introduction and Thesis Objectives ....................................................... 1 1.1 Background ........................................................................................................................... 1 1.1.1 The Current State of the HIV/AIDS Epidemic ........................................................................... 1 1.1.2 HIV Structure and Replication .................................................................................................... 2 1.1.3 HIV Origins and Genetic Diversity .............................................................................................. 7 1.1.4 Natural History of HIV Infection in Humans ............................................................................. 9 1.1.5 Host and Viral Genetic Factors Affecting HIV Disease Progression and HIV Sequence Variation ..................................................................................................................................... 10 1.1.6 Hepatitis C Virus: A Brief Primer .............................................................................................. 12 1.2 Clinical Management of Viral Infectious Diseases ............................................................. 17 1.2.1 HIV Treatment Options: Antiretrovirals and Highly Active Antiretroviral Therapy ............. 17 1.2.2 HCV Treatment Options: From Interferon to Direct-Acting Antivirals ................................. 20 1.2.3 Markers of Disease Progression and Treatment Outcome ...................................................... 24 1.2.4 Viral Evolution and Drug Resistance ........................................................................................ 26 1.3 Sequencing Technologies ................................................................................................... 30 1.3.1 Sanger Sequencing .................................................................................................................... 30 1.3.2 Massively Parallel “Next-Generation” Sequencing Platforms ................................................. 33 1.3.2.1 Roche 454 Pyrosequencing ......................................................................................... 34 1.3.2.2 Illumina Sequencing-by-Synthesis ............................................................................. 36 1.3.2.3 Other Platforms ........................................................................................................... 37 1.4 Thesis Organization and Objectives ................................................................................... 39 Chapter 2: Automating HIV Drug Resistance Genotyping with RECall, a Freely Accessible Sequence Analysis Tool ..................................................... 42 ix  2.1 Background and Introduction............................................................................................ 42 2.2 Materials and Methods ...................................................................................................... 44 2.2.1 Laboratory Methods ................................................................................................................... 44 2.2.2 Sequence Analysis Using RECall ............................................................................................... 44 2.2.3 RECall Nucleotide Mixture Calling and “Marking” of Potentially Problematic Bases ........... 45 2.2.4 RECall Pass/Failure Criteria ...................................................................................................... 46 2.2.5 The RECall Web Application ..................................................................................................... 47 2.2.6 Data Analysis ............................................................................................................................. 48 2.3 Results ................................................................................................................................ 48 2.3.1 RECall Performance Characteristics ........................................................................................ 48 2.3.2 Nucleic Acid Sequence Concordance ........................................................................................ 49 2.3.3 Amino Acid Sequence Concordance .......................................................................................... 52 2.3.4 Antiretroviral Susceptibility Scoring ......................................................................................... 52 2.4 Discussion and Conclusions ............................................................................................... 55 Chapter 3: Technical and Regulatory Shortcomings of the TaqMan Version 1 HIV Viral Load Assay ................................................................................. 59 3.1 Background and Introduction............................................................................................ 59 3.2 Materials and Methods ...................................................................................................... 60 3.3 Results ................................................................................................................................ 62 3.3.1 Concordance of Amplicor and TaqMan pVL Results at Low Viral Load Strata ...................... 62 3.3.2 Detectability by TaqMan v1 does not Predict Short-Term Virologic Failure or Antiretroviral Resistance ................................................................................................................................... 65 3.3.3 Mismatches in the TaqMan v1 Primer Binding Regions in HIV gag Result in Systematic Underestimation of Viral Load in a Subset of Patients ........................................................... 68 3.4 Discussion and Conclusions ............................................................................................... 70 Chapter 4: Increasingly Successful Highly Active Anti-Retroviral Therapy Delays the Emergence of New HLA Class I-Associated Escape Mutations in HIV-1 .................................................................................................. 74 4.1 Background and Introduction............................................................................................ 74 4.2 Materials and Methods ....................................................................................................... 75 4.2.1 Cohort Description ..................................................................................................................... 75 4.2.2 HLA and HIV Genotyping.......................................................................................................... 77 4.2.3 Classification of HLA-Associated Polymorphisms and Drug Resistance Mutations ............. 77 4.2.4 Statistical Methods ..................................................................................................................... 78 x  4.3 Results ................................................................................................................................ 79 4.3.1 Baseline Prevalence and Post-HAART Incidence of Immune Escape and Drug Resistance . 79 4.3.2 Poorer Response to HAART is Associated with Continued Immune Escape .........................83 4.3.3 Drug Resistance, but not Immune Escape, Accompanies Viral Breakthrough during HAART ..................................................................................................................................................... 87 4.3.4 Continued Immune Escape is Not Affected by the Protein Targeted by HAART .................. 88 4.3.5 Reversion of Immune Escape Occurs Rarely ........................................................................... 88 4.4 Discussion and Conclusions ............................................................................................... 89 Chapter 5: Measuring HIV Fitness by Simultaneous Co-Culture and 454 Deep Sequencing ......................................................................................... 91 5.1 Background and Introduction............................................................................................. 91 5.2 Replication Fitness of Multiple NNRTI-Resistant HIV-1 Variants in the Presence of Etravirine Measured by 454 Deep Sequencing .................................................................. 92 5.2.1 Simultaneous Competitive Culture and 454 Deep Sequencing ............................................... 92 5.2.2 The Y181V Mutation Confers a Clear Fitness Advantage over Other NNRTI Resistant Mutations in the Presence of etravirine .................................................................................... 94 5.3 Competitive Fitness Assays Indicate that the E138A Substitution in HIV-1 Reverse Transcriptase Decreases in Vitro Susceptibility to Emtricitabine ..................................... 98 5.4 Discussion and Conclusions ............................................................................................. 104 Chapter 6: HIV Drug Resistance Testing by High-Multiplex “Wide” Sequencing on the MiSeq Instrument ....................................................................... 106 6.1 Background and Introduction.......................................................................................... 106 6.2 Materials and Methods .................................................................................................... 108 6.2.1 Plasma Samples and RNA Extraction ..................................................................................... 108 6.2.2 RT-PCR Amplification of Protease-RT and Sanger Sequencing............................................ 109 6.2.3 Sanger Sequence Data Processing ........................................................................................... 109 6.2.4 MiSeq Library Preparation and Sequencing ........................................................................... 110 6.2.5 MiSeq Data Processing .............................................................................................................. 111 6.2.6 Analysis of Concordance of Nucleotide Sequences and Resistance Interpretations ............. 111 6.3 Results ............................................................................................................................... 112 6.3.1 Sanger and MiSeq Sequencing Success Rate ...........................................................................112 6.3.2 Sanger and MiSeq Sequence Concordance .............................................................................. 115 6.3.3 Concordance in Drug Resistance Interpretations .................................................................. 119 6.3.4 Sensitivity Analysis of MiSeq Coverage and Mixture Cutoffs ................................................ 120 xi  6.4 Discussion and Conclusions .............................................................................................. 122 Chapter 7: Development and Validation of Two Screening Assays for the HCV NS3 Q80K Polymorphism Associated with Reduced Response to Combination Treatment Regimens Containing Simeprevir ............... 128 7.1 Background and Introduction........................................................................................... 128 7.2 Materials and Methods .....................................................................................................130 7.2.1 Sample Sets ............................................................................................................................... 130 7.2.2 RNA Extraction ........................................................................................................................ 130 7.2.3 One-Step RT-PCR ......................................................................................................................131 7.2.4 Amplification of HCV NS3 by Nested PCR and Sanger Sequencing ..................................... 132 7.2.5 Illlumina MiSeq Sample Preparation and Sequencing .......................................................... 133 7.2.6 Assessment of Performance Characteristics ........................................................................... 134 7.3 Results ............................................................................................................................... 135 7.3.1 Assay Accuracy – Sanger Sequencing Concordance .............................................................. 136 7.3.2 Assay Accuracy – MiSeq Results ............................................................................................. 139 7.3.3 Assay Precision – Repeatability of PCR and Sequencing ....................................................... 140 7.3.4 Assay Precision – Reproducibility of PCR and Sequencing ................................................... 141 7.3.5 Assay Sensitivity – Determination of the Assay Limit of Detection ...................................... 144 7.3.6 Assay Specificity – No PCR Amplification of HCV-Negative Samples .................................. 146 7.3.7 Assay Specificity – No Interference by Potential Coinfecting Viruses .................................. 146 7.3.8 Assay Specificity – Lack of Cross-Reactivity with Other HCV Genotypes ............................ 146 7.4 Discussion and Conclusions .............................................................................................. 146 Chapter 8: General Discussion and Conclusion .................................................. 150 8.1 Thesis Summary ................................................................................................................ 150 8.2 Impact, Applications and Future Directions .................................................................... 151 8.2.1 Discussion of Chapter 2 ............................................................................................................ 151 8.2.2 Discussion of Chapter 3 ........................................................................................................... 153 8.2.3 Discussion of Chapter 4 ........................................................................................................... 155 8.2.4 Discussion of Chapter 5............................................................................................................ 157 8.2.5 Discussion of Chapter 6 ........................................................................................................... 159 8.2.6 Discussion of Chapter 7 ............................................................................................................ 161 8.3 Limitations ........................................................................................................................ 163 Bibliography ....................................................................................................... 167 xii  Appendices ......................................................................................................... 217 Appendix I: HLA-Associated Polymorphisms in HIV-1 Protease and Reverse Transcriptase ....... 217 Appendix II: HIV-1 Reverse Transcription and PCR Amplification Primers for MiSeq and Sanger Sequencing ...................................................................................................................... 221 Appendix III: Index Sequences for 1152-Fold Multiplex MiSeq Sequencing......................................222 Appendix IV: Sanger and MiSeq Sequencing Success Rate Stratified by Plasma Viral Load and Cohort .............................................................................................................................. 225 Appendix V: Distribution HIV Subtypes in Successfully Sequenced Samples .................................226 Appendix VI: Prevalence of NNRTI and NRTI Drug Resistance Mutations Detected by MiSeq and Sanger Sequencing .......................................................................................................... 227 Appendix VII: Sanger and MiSeq Nucleotide Sequence Concordance Stratified by Plasma Viral Load and Cohort .......................................................................................................................229 Appendix VIII: Effect of Minimum Read Coverage Thresholds on Sensitivity and Specificity of MiSeq in Detecting Resistance Mutations Observed by Sanger Sequencing ......................... 230 Appendix IX: HCV NS3 Reverse Transcription and PCR Primers for Primary and Secondary Sanger Sequencing Methods ....................................................................................................... 232 Appendix X: HCV NS3 Sanger Sequencing Primers .......................................................................... 233 Appendix XI: Comparison of Nucleic Acid Base Calls between HCV NS3 Sequencing Assays and an External Laboratory Control ..........................................................................................234 Appendix XII: Sequencing Coverage across HCV by a Near-Whole-Genome HCV Next-Generation Sequencing Assay ............................................................................................................ 237  xiii  Table 2.1: Configuration Variables for Nucleotide Mixture Calling and Base “Marking” for Clinical Drug Resistance Genotyping ............................................................................................ 46 Table 2.2: Sequence Rejection Criteria Used by RECall ................................................................... 47 Table 2.3: Sierra Drug Resistance Interpretation Concordance between Human- and RECall-Analyzed Sequences ......................................................................................................... 54 Table 3.1: Characteristics of 279 HAART-Treated Patients with Low-Level Viremia Followed Longitudinally .................................................................................................................. 66 Table 4.1: Cohort Characteristics prior to HAART Initiation ........................................................... 76 Table 4.2: Predictors of Drug Resistance and Immune Escape ........................................................84 Table 4.3: Reversion of Pre-Therapy HLA-Associated Mutations after HAART Initiation ..............89 Table 5.1: Population Sequencing of Viral Culture Supernatants .................................................... 97 Table 5.2: Susceptibility to Etravirine and Nevirapine of Select NNRTI-Resistant HIV-1 Viruses ..98 Table 5.3: Susceptibility to Emtricitabine, Lamivudine, Tenofovir, Rilpivirine, and Etravirine of HIV-1 Viruses Containing Substitutions at Codon E138 ............................................... 103 Table 7.1: HCV NS3 Sanger Sequencing Assay Repeatability ........................................................ 140 Table 7.2: HCV NS3 Sanger Sequencing Assay Reproducibility .................................................... 142 Table 7.3: Lower Limit of Detection of the HCV NS3 Sanger Sequencing Assay ........................... 145  xiv  Figure 1.1: HIV-1 Replication Cycle and Antiretroviral Drug Targets ................................................. 3 Figure 1.2: Global Distribution of HIV-1 Subtypes .............................................................................. 8 Figure 1.3: HCV Replication Cycle and Targets of Direct-Acting Antivirals ...................................... 15 Figure 1.4: Genetic Diversity of HIV and HCV .................................................................................. 16 Figure 1.5: Timeline of HIV-1 Antiretroviral Drug Development and Approval ............................... 19 Figure 1.6: Sustained Virologic Response Rates of HCV Treatment Regimens ................................. 21 Figure 2.1: Concordant and Discordant Nucleotide Base Calls in Protease and Reverse Transcriptase Sequences Analyzed Manually and by RECall ................................................................. 50 Figure 2.2: Chromatograms Illustrating Discordant Base Calls between Human and RECall Sequence Interpretations ................................................................................................. 51 Figure 3.1: Plasma Viral Load Testing and Reporting Protocol in British Columbia between October 2009 and April 2010 ........................................................................................................ 61 Figure 3.2: Poor Concordance between TaqMan v1 and Amplicor v1.5 at Low Plasma Viral Load Levels (40-250 HIV RNA Copies/mL by TaqMan v1) ..................................................... 63 Figure 3.3: Bland-Altman Plot of Results from Parallel Testing of Viral Load Samples by TaqMan v1 and Amplicor v1.5 ............................................................................................................ 64 Figure 3.4: The Proportion of Samples Detectable by Amplicor v1.5 Increases as a Function of the TaqMan v1 Value .............................................................................................................. 65 Figure 3.5: Low-Level Viremia by Taqman v1 Does Not Predict Short-Term Virological Failure ..... 67 Figure 3.6: Example of Systematic Underestimation of Plasma Viral Load by TaqMan v1 in Longitudinal Samples from a Representative Patient ..................................................... 69 Figure 4.1: Prevalence of Immune Escape Mutations at HLA-Associated Codons Pre- and Post-HAART ............................................................................................................................ 80 Figure 4.2: Prevalence of Pre- and Post-HAART Drug Resistance Mutations ................................... 82 Figure 4.3: Successful HAART Delays HLA Escape and Drug Resistance .........................................86 xv  Figure 4.4: Escape and Resistance Prior to Suppression and at Time of First Failure ...................... 87 Figure 5.1: Concentrations of p24 in Viral Culture Supernatants...................................................... 93 Figure 5.2: Replication Fitness of Multiple NNRTI-Resistant HIV-1 Variants in the Presence of Etravirine or Nevirapine .................................................................................................. 96 Figure 5.3: Replication Fitness of Multiple NNRTI-Resistant HIV-1 Variants in the Presence of Etravirine and/or Emtricitabine .................................................................................... 100 Figure 5.4: Replication Fitness of Multiple NNRTI-Resistant HIV-1 Variants in the Presence of Rilpivirine and/or Emtricitabine ................................................................................... 102 Figure 6.1: Distribution of MiSeq Sequencing Coverage .................................................................. 113 Figure 6.2: Success Rates of Sanger and MiSeq Sequencing Stratified by Plasma Viral Load .......... 115 Figure 6.3: Base Calling Differences by the Illumina MiSeq and ABI 3730xl Sanger Sequencing Methods .......................................................................................................................... 117 Figure 6.4: Phylogenetic Tree of Samples Successfully Sequenced by MiSeq .................................. 118 Figure 6.5: Proportion of Samples with Detectable Resistance Mutations with Varying Mixture Calling Thresholds ......................................................................................................... 120 Figure 6.6: Effect of MiSeq Mixture-Calling and Minimum Coverage Thresholds on Sequencing Accuracy and Success Rates ............................................................................................ 121 Figure 6.7: Sanger and MiSeq Sequencing Concordance Stratified by Plasma Viral Load .............. 122 Figure 7.1: Summary of Nucleotide Sequence Discordances between HCV NS3 Sequencing Assays Developed by BCCfE and Results Obtained by Janssen Diagnostics ............................. 137 Figure 7.2: Analysis of the Number of Mixed Bases per Sequence Suggests no Systematic Bias in Amplification of Minority Species .................................................................................. 138 Figure 7.3: Repeatability of Q80K Measurements across Five Independent Replicates of a Whole-Genome HCV MiSeq Sequencing Assay ......................................................................... 143  xvi  3TC – lamivudine AIDS – Acquired Immune Deficiency Syndrome BC – British Columbia BCCfE – British Columbia Centre for Excellence in HIV/AIDS bp – Base pair CA – Capsid protein CCR5 – C-C motif Chemokine Receptor 5  CD4 – Cluster of Differentiation 4 CTL – Cytotoxic T Lymphocyte CXCR4 – C-X-C motif Chemokine Receptor 4 DNA – Deoxyribonucleic Acid dNTP – deoxyribonucleotide triphosphate ddNTP – dideoxyribonucleotide triphosphate EC50 – Half Maximal Effective Concentration  Env – Envelope EQA – External Quality Assessment ETV – etravirine FDA – United States Food and Drug Administration FTC – emtricitabine gp41 – Glycoprotein 41 gp120 – Glycoprotein 120 gp160 – Glycoprotein 160 Gag – Group-specific Antigen GT – Hepatitis C Virus Genotype HAART – Highly Active Antiretroviral Therapy HIV or HIV-1 – Human Immunodeficiency Virus type 1 HCV – Hepatitis C Virus HLA – Human Leukocyte Antigen HR – Hazard Ratio IC50 – Half Maximal Inhibitory Concentration IFN – Interferon IN – Integrase protein IAS-USA – International Antiviral Society USA IU – International Units xvii  LMIC – Low- and middle-income countries LTR – Long Terminal Repeat MA – Matrix protein MHC – Major Histocompatibility Complex MSM – Men who have Sex with Men mRNA – Messenger Ribonucleic Acid NC – Nucleocapsid protein Nef – Negative Factor NGS – Next-Generation Sequencing NRTI – Nucleoside Reverse Transcriptase Inhibitor NNRTI – Non-Nucleoside Reverse Transcriptase Inhibitor NS – Hepatitis C Virus Non-Structural protein NVP – nevirapine PAT – Parenteral Antischistosomal Therapy PCR – Polymerase Chain Reaction PegIFN – Pegylated Interferon PI – Protease Inhibitor PIC – Pre-Integration Complex Pol – Polymerase pVL – Plasma Viral Load Q1 – First quartile Q3 – Third quartile RBV – ribavirin Rev – Regulator of expression of viral proteins RNA – Ribonucleic Acid RPV – rilpivirine RT – Reverse Transcriptase  RT-PCR – Reverse Transcription Polymerase Chain Reaction SIV – Simian Immunodeficiency Virus  SNP – Single Nucleotide Polymorphism SVR – Sustained Virologic Response Tat – Trans-Activator of Transcription UARTO – Uganda AIDS Rural Treatment Outcomes VIDUS – Vancouver Injection Drug User Study Vif – Virion Infectivity Factor Vpr – Viral Protein R Vpu – Viral Protein Unique WT – Wild-Type xviii  I will forever be indebted to the great many people without whom this work could never have been completed. I thank all the patients and study participants worldwide whose contributions go well beyond the thousands of specimens they donated to research.  To all the members of the BC Centre for Excellence in HIV/AIDS research laboratory past and present, thank-you for your tireless efforts. Together you have produced a rich repository of data with which to work and have made the lab a great place to be. Winnie and Theresa – thank-you for your years of teaching and technical support. Conan, Brian, Dennison, Don, Carolyn, and Leslie – your invaluable work behind the scenes often goes unnoticed and underappreciated. My fellow students, Luke, Guinevere, and Rachel, I hope you enjoyed these past few years as much as I have. I sincerely appreciate the guidance and encouragement of my committee members Julio Montaner, Viviane Lima, and Art Poon. My parents, Hans and Sonata, who always believed in me and pushed me to always be my best. The one constant presence in my life, Celia. You have stuck with me all these years despite distance and difficulty. Thank-you for being a companion, partner and friend. My supervisor and mentor, Richard Harrigan. Thank-you for taking a chance on an out-of-work physicist all those years ago. You have always challenged me to question everything that I see, read or hear. Finally, my sister Zabrina. I have always looked up to you – this is just another example of me trying to copy everything that you do. Thank-you for setting me on this path and for always looking out for your little brother. 1     Acquired Immune Deficiency Syndrome (AIDS) was first described in 1981 when young homosexual men began presenting to physicians with rare malignancies and opportunistic infections [1,2]. Human Immunodeficiency Virus Type 1 (HIV-1), the cause of AIDS, has since become a major global health issue [3–5]. At the end of 2013 an estimated 35 million persons worldwide were living with HIV [6]. Approximately 2.1 million individuals became infected that year [6] and 1.3 million died from HIV-related causes [7]. While the number of HIV-related deaths has declined from a peak of 2.3 million in 2005 [6], HIV/AIDS remains among the top six leading causes of death worldwide [8].  Sub-Saharan Africa bears a disproportionate burden of HIV infection with over 70% of new HIV infections being from the region. While 4.7% of adults (aged 15-49) living in Sub-Saharan Africa are estimated to be HIV-infected, HIV prevalence is highest in Swaziland (26.5%), Lesotho (23.1%) and Botswana (23%). With an estimated 6.1 million infected persons (17.5%), South Africa has the largest epidemic of any country [6]. In contrast, it is estimated that approximately 71,300 Canadians were living with HIV in 2011, and that approximately 3,175 new infections occur annually [9]. HIV is transmitted through direct contact with blood and certain body fluids, including semen, vaginal secretions and breast milk. Heterosexual contact continues to account for the majority (~85%) of new infections worldwide [10]. HIV prevalence is substantially higher among men who have sex with men (MSM) compared to the rest of the adult population; among MSM in the Americas, South- and South-East Asia, and Sub-Saharan Africa, HIV prevalence ranges from 14-18% [11]. Less than 0.5% of the global population is made up of people who inject drugs; however, these individuals account for 5-10% of all HIV-infected persons worldwide [12]. Injection drug use accounts for approximately one-third 2  of new HIV infections outside Sub-Saharan Africa, with the highest rates observed in South-East Asia and Eastern Europe [6,13]. Finally, while rates of perinatal HIV transmission have declined significantly in recent years, particularly in the Western world, mother-to-child transmission remains a concern; in 2012, approximately 260,000 children in low- and middle-income countries were newly infected with HIV [6].   HIV is a lentivirus of the Retroviridae family. These are RNA viruses that replicate though a DNA intermediary. HIV virions are roughly spherical in shape with a diameter of approximately 120 nm [14]. An outer lipid bilayer envelope encloses an internal protein core, which in turn contains two copies of an approximately 9700 base pair-long, single stranded positive-sense RNA genome [15] and the viral proteins required for the first stages of infection. The nine open reading frames of the RNA genome, gag, pol, vif, vpr, vpu, tat, rev, env, and nef, encode a total of fifteen viral proteins [16]. Three of these open reading frames code for the group-specific antigen (Gag), polymerase (Pol) and envelope (Env) polyproteins which are further processed into structural proteins and enzymes common to all retroviruses. The four component proteins of the Gag polyprotein precursor (Pr55Gag), matrix (MA, p17), capsid (CA, p24), nucleocapsid (NC, p7), and p6, form the major structural components of the virion. The Env polyprotein (gp160), is cleaved into the surface (SU, gp120) and transmembrane (TM, gp41) glycoproteins of the viral envelope by the host cellular protease Furin [17]. The gag-pol polyprotein (Pr160Gag-Pol) is initially generated by a ribosomal frameshift event during translation of the viral mRNA [18]. This precursor is subsequently cleaved by the HIV protease into the essential enzymes required for viral replication, integration and processing: protease (PR), reverse transcriptase (RT), RNAse H, and integrase (IN) [19]. These products are included in the mature virion. The remaining six open reading frames encode regulatory or accessory proteins each with a specialized function in viral replication. Three of these, Vif, Vpr, and Nef, are also packaged in the viral particle [20]. 3  Five stages of the HIV lifecycle, co-receptor binding, fusion, reverse transcription, integration, protease cleavage, are the targets of approved HIV antiretrovirals (Figure 1.1).   Illustrated is the HIV-1 replication cycle from viral entry to progeny release. Antiretroviral drug classes are noted adjacent to the viral life cycle step that they inhibit. This figure is adapted from Wong RW, et al., PLoS Pathog (2013) with modifications by Michael T. O’Shaughnessy. This figure is used under the terms of the Creative Commons Attribution License. © 2013 Wong et al. [21] 4  The HIV replication cycle begins with viral attachment and entry into the target cell (Figure 1.1, no. 1), a process mediated by the envelope glycoproteins [22]. The envelope surface glycoprotein gp120 trimer first binds the CD4 receptor found primarily on the surface of T lymphocytes, monocytes, macrophages and dendritic cells [23,24]. Expression of CD4 alone, however, is not sufficient to permit infection of the host cell. Interaction with one of two human 7-transmembrane chemokine co-receptors, CCR5 or CXCR4, is required for efficient entry of target cells [25–30]. Although rare, HIV-1 has been demonstrated to be capable of using other co-receptors to infect target cells in vitro [31–34].  Binding CD4 causes gp120 to undergo a series of conformational changes [35,36]. First, the co-receptor binding domains of gp120, defined primarily by the third variable (V3) loop [30,37,38], are exposed, allowing them to interact with CCR5 or CXCR4. Co-receptor binding induces further structural re-arrangement of gp120, leading to the exposure of the transmembrane glycoprotein (gp41) trimer [39]. The exposed hydrophobic gp41 fusion peptide is then inserted into the cellular membrane. Further structural changes bring the viral and target cell membranes into close proximity resulting in viral and cellular membrane fusion (Figure 1.1, no. 2) [39–41].  Following virus-cell membrane fusion, the viral contents are inserted into the cellular cytoplasm. The viral capsid, a multimer of several thousand copies of the CA protein, dissociates in a process facilitated by the host protein cyclophilin A [42,43]. This “uncoating” process (Figure 1.1, no. 3) results in the release of the viral genetic material, nucleocapsid proteins and reverse transcriptase along with integrase and Vpr, the factors required for import of viral DNA into the nucleus [44].  Prior to nuclear import, the single-stranded viral RNA genome must first be converted into double-stranded DNA by the viral reverse transcriptase enzyme (Figure 1.1, no. 4). To perform this action, RT contains two functional active sites: a RNA- and DNA-dependent DNA polymerase (to synthesize proviral DNA) and an RNase H domain (to degrade the viral RNA template as cDNA is synthesized) [45]. Importantly, the RT enzyme lacks proofreading ability and thus has extremely poor fidelity. 5  Given the high error rate (almost one mutation per replication cycle [46]) and rapid turnover rate (approximately 109 new virions produced per day [47], it is estimated that each possible point mutation in the HIV genome occurs thousands of time every day in each untreated HIV-infected individual [48]. The error-prone nature of RT, combined with host- and drug-selection pressures, contributes to the extreme genetic diversity observed worldwide as well as to the selection of drug-resistant and immune-escaped variants (to be discussed in greater detail in 1.1.3 and 1.1.4). Successful reverse transcription results in the formation of the pre-integration complex (PIC), consisting of double-stranded viral DNA, host factors and the viral proteins RT, integrase, Vpr, matrix and nucleocapsid [49]. The PIC is transported to the nucleus (Figure 1.1, no. 5) in an incompletely-understood process mediated by Vpr [50], integrase [51] and host nuclear transport proteins [52]. Having entered the nucleus, the PIC is tethered to the host genomic DNA where integrase catalyzes the insertion of the double-stranded viral DNA into the host cell genome (Figure 1.1, no. 6) [53]. Thus, the HIV genetic material is stably integrated into the host genome as a provirus which can subsequently serve as the template for the synthesis of viral RNA. Alternatively, infected cells can enter a transcriptionally inactive state, resulting in a “reservoir” of latently infected cells [54]. Transcription of the integrated proviral DNA is initiated by a promoter within the HIV 5’ LTR (Figure 1.1, no. 7). The full length (~9 kilobases) plus strand RNA transcript is spliced by host spliceosomes into singly- (~4-5 kilobases) and multiply-spliced (~2 kilobases) transcripts [55] which are exported through nuclear pores to the cytoplasm (Figure 1.1, no. 8). Once exported, viral mRNA is translated into viral proteins by host ribosomes (Figure 1.1, no. 9) [45]. The short multiply-spliced mRNA are translated into the regulatory proteins Tat and Rev which regulate the early stages of transcription, as well as the accessory protein Nef [55,56]. The presence of Tat increases the level of transcription from the LTR, while an increasing concentration of Rev inhibits the host cell splicing of transcripts. Therefore, the accumulation of Rev mediates a shift in mRNA production from short multiply-spliced to longer singly-spliced or unspliced transcripts [56].  6  The remaining accessory proteins Vif, Vpr, Vpu, along with the Env polyprotein, are synthesized from medium length singly-spliced transcripts [57]. The accessory proteins aid in evading host adaptive and innate immune responses, and are required for efficient viral replication and release in vivo [58]; Nef and Vpu downregulate cell surface CD4 [58,59]. In addition, Nef also downregulates MHC class I from the cell surface, allowing HIV-1 infected cells to evade killing by cytotoxic T-lymphocytes [59]. Vpu also enhances release of virions from the cell surface by antagonizing the host factor tetherin [60]. Vif antagonizes the host restriction factor APOBEC3G, a cytidine deaminase that induces guanine to alanine hypermutation during reverse transcription [61]. In addition to facilitating nuclear transport of the PIC, Vpr also increases the efficiency of transcription from the LTR [62]. Finally, unspliced 9-kilobase transcripts are translated into the Gag and Gag-Pol precursor polyproteins or packaged as genomic RNA into budding virions [18]. The assembly of new viral particles occurs at the host cell plasma membrane in a process that is largely regulated by the Gag structural proteins (Figure 1.1, no. 10). The Pr55Gag and Pr160Gag-Pol polyproteins are targeted to the cell membrane by MA [63]. The Env polyprotein gp160 is cleaved into gp41 and gp120 by the host protease furin during vesicular transport to the plasma membrane [17,63]. The remaining viral proteins along with the genomic RNA also co-associate with various portions of the Gag polyprotein. So-called “late domains” in the p6 portion of the Gag polyprotein engage components of the ESCRT pathway (endosomal sorting complexes required for transport) to initiate viral budding [45,64]. Immature progeny viruses, enveloped in a lipid membrane derived from the host cell, must first undergo a maturation step in which the viral protease cleaves the Gag and Gag-Pol polyproteins into their respective enzymatic and structural components (Figure 1.1, no. 11). A spontaneous structural rearrangement then occurs; around 2000 MA molecules assemble into a matrix lining the lipid membrane, preserving its three-dimensional structure. A similar number of CA molecules arrange to form a cone-shaped capsid than encloses the viral RNA and essential viral and host proteins required for replication (Figure 1.1, no. 12) [45,63]. Once mature, the infectious viral particle may begin the next round of replication in a new target cell. 7   The two Human Immunodeficiency Viruses, HIV-1 and HIV-2, are each the result of cross-species transmissions of variants of Simian Immunodeficiency Virus (SIV) from non-human primates [65]. HIV-1 infection accounts for the majority of the HIV epidemic worldwide [6], while the HIV-2 epidemic is largely confined to Western Africa where 1 to 2 million individuals are estimated to be infected [66]. HIV-1 can be divided into four phylogenetic lineages, or groups: M (main), O (outlier), N (non-M/non-O), and P (putative) [67,68]. The pandemic group M and the rare group N “strains” are thought to have been derived from multiple independent zoonotic transmissions of SIVcpz from allopatric chimpanzee (Pan troglodytes troglodytes) communities in southern Cameroon during the late 19th century [69,70]. Beginning in the 1920s, the epidemic spread from a focus in Leopoldville, Belgian Congo (now Kinshasa, Democratic Republic of the Congo), as rail and river transport increased [71]. During this time HIV-1 group M diversified into several distinct subtypes (A-D, F-H and J-K) and inter-subtype recombinants emerged [67]. The global epidemic was seeded by multiple founder viruses and thus the geographic distribution of these subtypes is not uniform (Figure 1.2). The greatest subtype diversity is observed in Central Africa, consistent with it being the origin of the first cross-species transmissions. Subtype C is the most prevalent globally, accounting for nearly 50% of infections worldwide. Subtype A (12% worldwide) is the most prevalent subtype in Eastern Europe and East Africa, while Subtype B infections (10% worldwide) dominate in Western Europe, Oceania and the Americas. Circulating recombinant forms (CRF) [67] make up significant proportions of the regional epidemics in South-East Asia (CRF01_AE) and in Western Africa (CRF02_AG) [72].  8   Countries are grouped into 15 regions (grey shading). Overlaid pie charts depict the relative prevalence of HIV-1 subtypes in each region. Pie chart diameter is proportional to the number of HIV-1 infected individuals in the region. Estimates of prevalence and subtype use available data from 2004-2007. Adapted from Hemelaar, Trends Mol Med (2012). The right to reproduce this figure has been granted by Elsevier Inc. via RightsLink (license number 3619001421970) © 2012 Elsevier Inc., Hemelaar [72] Similarly, groups O and P also originated in Central Africa. Group O, which accounts for approximately 100,000 infections in West-Central Africa, is thought to have originated from cross-species transmission of SIVgor from western lowland gorillas (Gorilla gorilla gorilla). SIVgor, in turn, likely originated through a single cross-species transmission event of SIVcpz from sympatric chimpanzee populations [73]. HIV group P, which has been identified in only two individuals from Cameroon to date, also shares ancestry with SIVgor [68,73]. The less pathogenic HIV-2, which affects 1 to 2 million individuals primarily in western Africa, is most closely related to SIVsmm endemic in sooty mangabeys (Cercocebus atys atys) native to south-western Côte d'Ivoire [74].  9  As previously discussed in Section 1.1.2, the low fidelity of HIV reverse transcriptase, the high rate of replication, and the life-long duration of infection all contribute to the virus’ substantial genetic diversity. Amino acid sequence variability between subtypes is typically between 17-35%, depending on the genome regions examined. Intra-subtype variation of sequences from different individuals ranges from 8-17% [75]. In addition, while most productive HIV infections are thought to be initiated by a single viral variant [76,77], host immune selection [78–80], neutral mutation and genetic drift [81] quickly result in substantial diversification within an infected individual [82]. The result is a diverse mixture of viral quasispecies, in which virus sequences within an individual can differ by up to 10% [75].   The risk of HIV acquisition varies widely depending on the method of exposure. Estimates of per-act risk of HIV infection via unprotected sexual exposure varies from 0.04 to 1.4% for insertive vaginal and receptive anal sex, respectively. Needle-sharing during injection drug use carries an approximate 0.6% risk of HIV infection, while an estimated 93% of blood transfusions from an HIV-positive donor result in infection. Multiple factors including condom use, male circumcision, pre- or post-exposure prophylaxis and antiretroviral treatment decrease the risk of HIV infection [83–85]. Acute HIV infection is characterized by a mononucleosis-like seroconversion illness that begins several days after exposure. Symptoms including lymphadenopathy, rash, fever and fatigue can last for several weeks [86]. In the first several weeks following transmission the virus establishes a systemic infection of CD4-expressing cells in the lymphoid organs, notably the lymph nodes, spleen and gut-associated lymphoid tissues (GALT). This leads to a severe and rapid depletion of both circulating and mucosal CD4+ T-cells in these compartments [87–89]. The decline of CD4+ cells to levels below 200 cells/µL and/or the emergence of one of several AIDS-associated illnesses defines AIDS [90].  Viral RNA becomes detectable in plasma in as little as 10 days following infection and rises to a typical maximum of several million copies per milliliter of plasma in the first 3-6 weeks post-infection [89,91]. 10  HIV RNA levels gradually decline from peak viremia concomitant with the emergence of the host adaptive immune response. These emergent humoral and cellular immune selection pressures also drive within-host HIV diversification during this time [76,78,80]. The initial immune response is primarily in the form of HIV-specific CD8+ cytotoxic T-lymphocyte (CTL) killing of infected cells [78,92,93] as demonstrated in CD8+ cell depletion studies in non-human primates [94,95]. CTL are able to identify and kill HIV-infected cells through interaction with cell surface major histocompatibility complex (MHC) molecules: MHC class I (also known as human leukocyte antigen; HLA) molecules bind intracellular protein fragments (peptides) and display them on the cell surface. Acting through the T-cell receptor, CTL recognize antigenic peptide bound to HLA and eliminate the infected cell [96]. HIV-specific antibodies begin to be detectable approximately 3-4 weeks following infection; however, these responses tend to be non-neutralizing [89,91,97]. Broadly-neutralizing antibodies are rarely generated; when they do emerge, it is typically only after 20-30 months of infection [89,98].  Ultimately, the host immune response is incapable of eliminating HIV. Instead, plasma viremia is reduced to a relatively stable “setpoint” level, after 3-6 months of infection [89]. On average, setpoint viremia is established at a level between 104 and 105 HIV RNA copies per milliliter; however, this level can vary by over 3-4 orders of magnitude between individuals [99–102]. Establishment of setpoint viremia marks the beginning of the chronic, clinically asymptomatic phase of HIV infection that typically lasts a median of 5 to 11 years with a large inter-individual variation [103,104]. However, if treatment is not initiated, CD4+ T-cell levels continue to decline leading to the emergence of opportunistic infections, AIDS, and eventually death [90].   The natural history of HIV infection varies substantially between individuals [105,106]. For example, some individuals remain HIV uninfected despite repeated high-risk exposures [107,108]. Other so-called “elite controllers” or “long-term nonprogressors” become HIV-infected, but are able to naturally 11  control viremia to undetectable levels and/or do not exhibit substantial declines in CD4 count over time [109–111]. In contrast, “rapid progressors” are those individuals who progress to a clinical AIDS definition in as little as one or two years after infection [112,113]. While demographic factors, such as age at infection, play a role in the rate of disease progression, the level at which the viremia setpoint is established accounts for a substantial proportion of variation [100,104]. Several studies have suggested that HIV viral load setpoint is largely heritable, with HIV sequence variation being the primary contributor [99,114,115]. That is, individuals infected with genetically similar viruses may establish viral load setpoint at similar levels. For example, mutations in the HIV genome can impair the virus’ ability to replicate. In the early 1980s eight individuals in the Sydney blood bank cohort received blood products from a single donor who was infected with HIV containing a large (~290-bp) deletion in nef. In the subsequent thirty years, these individuals exhibited slower than typical CD4 decline or maintained undetectable viremia, characteristics consistent with infection with a defective virus [116–118]. In contrast, amino acid changes in the HIV V3 region of gp120 that induce a co-receptor tropism switch from CCR5- to CXCR4-using are associated with more rapid progression [37,113,119]. On a broader scale, different HIV-1 group M subtypes may be associated with differential rates of disease progression. In regions where multiple HIV subtypes c0-circulate, several studies have demonstrated that individuals infected with subtype A have slower progression to AIDS and death than individuals infected with subtypes C or D [120–123].  While viral genetic factors contribute to the establishment of viral load setpoint, host genetic variation, rather than infection with a replication deficient virus, is the major determinant of spontaneous “elite” control of HIV viremia [124]. Single nucleotide polymorphisms (SNP) in the MHC region that serve as surrogate markers for decreased HLA-C expression, specific HLA alleles (namely HLA-B*57:01), or specific amino acids that define the epitope binding specificity of HLA-B are primarily responsible for the level at which viral setpoint is established and in turn the rate of disease progression [125–129]; however, the influence of host and viral variation on HIV viremia and disease progression are not independent. As previously stated, host selection drives HIV evolution towards variants that “escape” 12  immune pressures [78–80]. As HLA epitope binding and CTL recognition is highly specific, these patterns of escape mutations are highly predictable; two individuals with the same HLA alleles are likely to develop the same escape mutations [130]. At the individual level, the selection of escape variants often precedes viral breakthrough and disease progression [131,132]. However, some mutations allow HIV to evade immune selection, but result in a substantial decrease in replicative fitness [133,134]; transmission of these less fit variants have been associated with lower setpoint viral load in both HLA-matched and mismatched recipients [135–137]. Thus, the heritability of viral load setpoint is at least partially explained by the transmission of less fit viruses carrying mutations selected by the previous host’s immune response. The effect of antiretroviral treatment on the continued selection of these immune escape mutants is the subject of the study presented in Chapter 4.  An estimated 150 million persons worldwide, including 240,000 Canadians have chronic Hepatitis C Virus (HCV) infection [138,139]. HCV infection is largely asymptomatic, however long-term infection can lead to liver damage including fibrosis, cirrhosis and hepatocellular carcinoma and eventually death [140,141].  HCV is primarily transmitted via percutaneous exposure to infected blood. Injection drug use remains the principal risk factor for HCV acquisition. An estimated 60-80% of the global injection drug-using population is HCV antibody-positive, with over 80% of all new HCV infections being attributable to injection drug use [139,142]. Given the similarities in risk factors, HCV/HIV coinfection is relatively common with approximately 20% of HIV-infected persons worldwide estimated to be HCV coinfected [143]; however, this rate varies considerably with geography. In North America and Europe up to 30% of HIV-positive persons are also HCV-infected, while in Sub-Saharan Africa the coinfection rate is estimated to be around 6% [144,145]. In contrast to HIV, heterosexual transmission of HCV is much less efficient [146,147]. Despite this, increasing HCV incidence among MSM populations, in particular HIV-infected MSM, is becoming a concern [148].  13  Iatrogenic transmission has historically been a major contributor to the global burden of HCV infection. During the 1970s and 1980s, transfusions with contaminated blood products were a significant source of HCV infections in developed countries. The risk of contracting post-transfusion non-A/non-B hepatitis was estimated to be 10 to 15% per transfusion in some countries over this time [149–151]. An estimated 30,000 Canadians became infected with HCV by blood and blood products before 1992, the year in which Canada began screening blood donations for HCV [152]. This finding may partially explain the extremely high prevalence of HCV infection among the “baby boomer” population. These individuals, born between 1945 and 1965, account for nearly 60% of HCV prevalent cases in Canada [153].  The extremely high prevalence of HCV in Egypt is the result of the largest recognized iatrogenic transmission of a blood-borne pathogen in human history. Parenteral antischistosomal therapy (PAT) campaigns have been linked to large-scale transmission of HCV in Egypt [154–156]. Approximately 15 to 20% of the Egyptian population is HCV antibody-positive. The geographic distribution of HCV infection in Egypt is, however, not uniform. The majority of HCV infected persons live in rural communities in the Nile Valley and Nile Delta, regions in which schistosomiasis (infection with parasitic flatworms of the genus Schistosoma) is hyper-endemic. Between the 1950s and 1980s the Egyptian government conducted mass-treatment PAT campaigns in these regions. Treatment consisted of multiple intravenous injections of “tartar emetic” (potassium antimony tartrate) over a period of several weeks. Reusable injection equipment was used; however, given the number of patients treated – an average of two to three million injections were administered to over 250,000 patients annually – sterilization procedures were likely followed incompletely or omitted entirely [154,157]. Since all persons over the age of 5 years were eligible for treatment, PAT campaigns brought together patients of all age groups and risk categories. Furthermore, individuals who became HCV-infected early during the course of PAT would develop infectious viremia while still receiving injections, thus potentially accelerating the rate of HCV transmission. PAT campaigns continued until the mid-1980s at which time oral antischistosomal drugs replaced tartar emetic injections. As individuals born before 1981 were never eligible for PAT, the substantially lower rate of HCV infection 14  among younger Egyptians supports PAT as a major contributor to the HCV epidemic in Egypt [154,155]. HCV is a member of the genus Hepacivirus of the family Flaviviridae. It is a 55-65 nm diameter, enveloped, single-stranded, positive-sense RNA virus. Two envelope proteins are anchored in a host cell-derived lipid bilayer that surrounds the viral core containing a single copy of the viral genetic material. The approximately 10,000-base pair long HCV RNA genome consists of a single open reading frame encoding three structural (Core, E1, E2) and seven non-structural (NS) proteins (p7, NS2, NS3, NS4A, NS4B, NS5A, NS5B). HCV replicates primarily in liver hepatocytes though a process that remains incompletely understood. Briefly, the envelope glycoproteins E1 and E2 mediate entry into the host cell through a complex interaction with cell surface molecules including CD81, scavenger receptor class B member 1 (SRB1), claudin 1 (CLDN1), and occludin (OCLN) (Figure 1.3, steps a and b). Once inside the cell, the RNA template is translated into a single polyprotein. This polyprotein is cleaved by host cell proteases and the viral serine protease NS3, releasing the individual structural and non-structural viral proteins (Figure 1.3, step c). The viral RNA and the non-structural proteins assemble into a replication complex located in a “membranous web” adjacent to the cell nucleus. RNA replication proceeds through a negative strand RNA intermediate catalyzed by the action of the RNA-dependent RNA polymerase, NS5B (Figure 1.3, step d). Synthesized RNAs serve as templates for translation or are packaged into progeny viruses at the endoplasmic reticulum membrane (Figure 1.3, step e). These viruses are subsequently released through the cellular secretory pathway (Figure 1.3, step f) [158,159].  15   Simplified illustration of the HCV replication cycle from viral entry to progeny release. Classes of direct-acting antivirals are noted in red text adjacent to the viral life cycle step that they inhibit. This figure is adapted from Moradpour D, et al., Nat Rev Microbiol (2007). The right to reproduce this figure has been granted by Nature Publishing Group via RightsLink (license number 3619001421970) © 2007 Nature Publishing Group Inc., Moradpour [160] Importantly, and in contrast to HIV, HCV does not integrate into the host genome. It is therefore possible to cure HCV infection either through the host immune response, or by therapy. Approximately 25% of individuals are able to spontaneously clear the virus within 6 to 12 months following infection [161]. Clearing HCV infection, however, does not confer sterilizing immunity although some degree of protection against re-infection is conferred [162,163]. Several demographic, behavioral and genetic factors influence the probability of viral clearance, the most notable being a polymorphism in the IL28B gene that encodes interferon-λ3 [164]. Individuals of European or African descent with the favorable C/C IL28B genotype (~40% of Europeans, ~15% of Africans) are approximately three-fold more likely to spontaneously clear HCV than those without [164]. Like HIV, the CTL mediated immune response plays a large role in regulating HCV RNA levels. Two HLA class I alleles, B*27 and B*57:01, have been associated with both spontaneous clearance of HCV and “elite control” of HIV viremia 16  [125,126,165–167]. In addition, the cellular immune response selects for escape mutations and drives the intra-host diversification of HCV [167–169]. HCV exhibits extreme genetic diversity, exceeding even that of HIV (Figure 1.4). HCV is classified into 7 major genotypes (GT1-7) [170]. Nucleotide sequence divergence between genotype consensus sequences typically exceeds 30%. Each HCV genotype is further divided into several subtypes which are defined by the presence of nucleotide differences at 15% or more of positions in Core/E1 and NS5B in samples of the same genotype [170]. While HCV GT1 infections are the most common worldwide, the global distribution of HCV genotypes is not uniform. HCV GT2 infections dominate in West Africa, GT3 in South Asia, GT4 in North Africa and the Middle East, GT5 in South Africa and GT6 in East Asia [171,172].  Maximum likelihood phylogenetic trees of HIV-1 group M Pol and HCV NS5B polymerase reference sequences illustrating the substantial genetic variability of these viruses. Trees are depicted on the same scale. Trees were drawn with FastTree using a generalized time-reversible model. HIV subtype and HCV genotype references were obtained from the Los Alamos HIV (http://hiv.lanl.gov/) and HCV (http://hcv.lanl.gov/) Sequence Databases, respectively.  17  This section has outlined the nature of the worldwide HIV and HCV epidemics and has described the natural history of infection with these two viruses. It has also highlighted how their capacity to mutate has led to incredible genetic diversity both at a global and individual level. Individual host immune selection can lead to control of infection or can drive viral diversification leading to disease progression. The following sections will further explore the relevance of genetic sequencing of these diverse viral populations from a clinical standpoint. They will focus on how viral diversity and the capacity to mutate affect the ability to effectively treat these infections.   Prior to the advent of antiretroviral therapy, the median time from HIV infection to AIDS ranged from 5 to 11 years, depending on an individuals’ age at infection [104]. By the early 1990s, HIV/AIDS had become the leading cause of death among young American adults aged 25 to 44 [173]. In the subsequent 25 years, effective combination therapy regimens have transformed HIV infection into a manageable chronic condition. Today, the average life expectancy of HIV-positive persons receiving antiretroviral therapy is comparable to that of the general population [174,175].  In 1987, the US Food and Drug Administration approved the first HIV antiretroviral drug, zidovudine (AZT), a nucleoside reverse transcriptase inhibitor (NRTI) [176]. Treatment with zidovudine resulted in CD4 count recovery and improved patient survival in the short-term, but the effects were not durable [176–179]. The following years saw the approval of four additional drugs in the NRTI class including didanosine, zalcitabine, and lamivudine (3TC), leading to trials of combination therapy with two agents. Dual-combination therapy improved treatment outcomes and patient survival relative to AZT monotherapy [180,181]. Unfortunately, the rapid development of antiretroviral resistance limited the efficacy of mono- and dual-NRTI therapies [182–185]. It was not until the mid-1990s that the development of drugs with different mechanisms of action, namely protease inhibitors (PI) and non-18  nucleoside reverse transcriptase inhibitors (NNRTI), led to truly substantial and sustained improvements in patient survival at the population level. Ultimately, the universal recommendation of three-drug combination regimens, denoted Highly Active Antiretroviral Therapy (HAART), led to the dramatic improvements in HIV-related morbidly and mortality that we have today [186–190]. To date, over 25 antiretrovirals in six drug classes have been approved for the treatment of HIV infection. Typically, combination regimens consisting of a minimum of three drugs from two or more classes are prescribed [191,192]. Nucleoside reverse transcriptase inhibitors were the first class of antiretrovirals approved (Figure 1.5). NRTI are nucleoside (or nucleotide) analogues: lacking a 3’-hydroxyl group, NRTI inhibit reverse transcription by acting as chain terminators during DNA synthesis [193,194]. Protease inhibitors (PI) competitively bind the protease active site preventing the cleavage of the Gag-Pol polyprotein [193,194]; the first PI, saquinavir, was approved in 1995 (note, unless otherwise stated, all dates correspond to the years in which approval was granted by the US Food and Drug Administration). The PI ritonavir, originally approved in 1996 as a standalone drug, is currently used exclusively, at a lower dosage, as a “boosting” agent for other PIs. Ritonavir inhibits the cytochrome P450-3A4 (CYP3A4) enzyme, thus reducing metabolic degradation and increasing bioavailability of co-administered PIs [195]. Non-nucleoside reverse transcriptase inhibitors bind allosterically to the viral reverse transcriptase enzyme and inhibit its action [193,194]; in 1996, nevirapine (NVP) became the first approved drug in this class. Three-drug combinations consisting of a backbone of two NRTIs and either an NNRTI or a boosted PI formed the basis of the initial effective HAART regimens and remain as such today [194].  19   Selected, commonly-used HIV antiretrovirals and combinations as of January 2015. Drugs are placed on the timeline according to the date of approval by the US Food and Drug Administration. The list is non-exhaustive and does not depict some approved drugs that are no longer manufactured.   Nucleoside Reverse Transcriptase Inhibitors (NRTI): zidovudine (AZT), didanosine (ddI), stavudine (d4T), lamivudine (3TC), abacavir (ABC), tenofovir disproxil fumarate (TDF), emtricitabine (FTC).Non-Nucleoside Reverse Transcriptase Inhibitors (NNRTI): nevirapine (NVP), efavirenz (EFV), etravirine (ETV), rilpivirine (RPV). Protease Inhibitors: saquinavir (SQV), indinavir (IDV), nelfinavir (NFV), amprenavir (APV), lopinavir/ritonavir (LPV/RTV), atazanavir (ATV), fosamprenavir (FPV), tipranavir (TPV), darunavir (DRV). Integrase Inhibitors: raltegravir (RAL), elvitegravir (EVG), dolutegravir (DTG). Fusion Inhibitor: fuzeon (T20). Entry Inhibitor: maraviroc (MVC). Fixed-Dose Combinations (Canadian brand names): Combivir (3TC + AZT), Trizivir (ABC + 3TC + AZT), Truvada (FTC + TDF), Kivexa (ABC + 3TC), Atripla (TDF + FTC + EFV), Complera (TDF + FTC + RPV), Stribild (TDF + FTC + EVG), Triumeq (ABC + DTG + 3TC). 20  Integrase inhibitors are the final class of antiretrovirals that target a viral enzyme. First approved in 2007, raltegravir prevents integration of the viral genetic material into the host genome by inhibiting integrase activity [196]. While the majority of approved antiretrovirals prevent viral replication in infected cells, the remaining two antiretroviral drug classes, fusion and entry inhibitors, aim to prevent viral entry into host cells. Both of these classes each contain a single approved agent. The fusion inhibitor enfuvirtide, approved in 2003, is a 36-amino acid peptide that binds gp41 preventing the conformational changes necessary for viral and host membrane fusion [194]. Finally, the entry inhibitor maraviroc, approved in 2007, targets a host rather than a viral protein. Maraviroc is a small molecule antagonist of the host CCR5 cell-surface receptor. By binding to the CCR5 chemokine receptor, maraviroc blocks the attachment of virions that use CCR5 as a co-receptor. However, HIV using the CXCR4 co-receptor are unaffected by maraviroc [197].  In contrast to HIV, it is possible to cure HCV infection through antiviral therapy. A sustained virologic response (SVR), the complete elimination of detectable HCV viremia after therapy completion, is considered an accurate surrogate of cure; fewer than 1% of patients with SVR relapse with detectable HCV viremia even after several years following the end of therapy [198,199]. Historically, SVR was determined 24 weeks after therapy end (SVR24); however, recent studies support the use of SVR measured 12 weeks post-therapy (SVR12) as the primary efficacy endpoint in clinical trials of new antivirals [200,201]. The first antiviral agent used to treat HCV, recombinant interferon (IFN) alfa, was first approved for use in 1991. Administered by subcutaneous injection thrice-weekly for 6 months, IFNα resulted in SVR24 in fewer than 10% of treated patients (Figure 1.6) [202]. Increasing treatment duration to 48 weeks resulted in marginal improvements in treatment response: approximately 16% of treated patients achieved SVR24 [203]. Initial trials demonstrated that monotherapy with the nucleoside inhibitor ribavirin (RBV) was ineffective in reducing HCV viremia [204]; however, combination 21  treatment with IFNα plus RBV resulted in SVR rates of approximately 40% after a 48-week course of treatment [205]. Pegylation of IFN, the covalent attachment of polyethylene glycol to interferon, reduces proteolytic breakdown and increased the biological half-life of the interferon molecule. Pegylated inferferon (PegIFN), given by subcutaneous injection once weekly for 48 weeks, improved SVR24 rates to approximately 39% when administered as monotherapy and to approximately 55% when given in combination with RBV [206–209].   Approximate rates of sustained virologic response (SVR) are displayed for patients with HCV GT1 infection for a selection of historic and modern HCV treatment regimens. The figure summarizes findings from [202,203,205–220]. From 2001 until the release of the first direct-acting antivirals 10 years later, combination PegIFN and ribavirin for 48 weeks was the only approved treatment for chronic HCV infection. During this time treatment guidelines were modified based on several key observations of the effectiveness (or ineffectiveness) of therapy in certain populations [221]. Pegylated interferon plus ribavirin has pan-22  genotypic activity; however, it is not equally effective against each HCV genotype. HCV GT1 was shown to be more difficult to treat with PegIFN; SVR24 rates on PegIFN/RBV are substantially higher in individuals infected with HCV GT2 or GT3 (~75%) than those with HCV GT1 infection (~45%) [207,208]. Based on these observations, PegIFN/RBV treatment duration was reduced to 24 weeks for individuals with HCV GT2 or GT3 [209]. In addition to HCV genotype, several other factors including older age, African-American ancestry, high pre-therapy HCV viral load, and more advanced liver disease were found to be associated with poorer treatment outcomes [207,222]. In these populations, a complex set of rules based on the individual patient’s initial virologic response was used to guide the duration of treatment [221]. Briefly, patients with a rapid reduction in HCV viremia are most likely to achieve SVR24. In these patients, SVR24 could be achieved with treatment durations as short as 24 weeks [223]. In contrast, patients without an initial viral response, such as those who fail to reduce HCV viremia to undetectable levels by 12 weeks of treatment, are extremely unlikely to achieve SVR24 and thus treatment was discontinued early [221,224,225]. Despite the partial effectiveness of PegIFN/RBV therapy, uptake was limited. On average fewer than 5% of persons with chronic HCV infection in Europe and the United States ever accessed treatment. This trend was driven largely by the lack of proper HCV screening and diagnosis in these populations [226,227]. In addition, treatment acceptance among individuals with known HCV infection was also low. Given the complexity and duration of treatment, and the significant side effects of PegIFN/RBV (notably fatigue, depression and hemolytic anemia), many patients opted to defer treatment until their disease progressed further or new drugs were available [207,208,228]. In fact, the side effects of PegIFN/RBV therapy were often perceived to be worse than HCV disease itself, which provided little incentive to diagnose HCV infection in the first place [226–228]. In contrast, a relatively high proportion of HCV-infected IDU expressed willingness to undergo PegIFN/RBV treatment [229,230]; however, significant structural and social barriers limited treatment delivery in IDU populations [230–232]. 23  The creation of efficient cell culture models in the early-2000s enabled the development of the first drugs targeted directly at HCV itself [233]. The first direct-acting antivirals (DAA), the NS3 protease inhibitors telaprevir and boceprevir, were approved in 2011 for the treatment of HCV GT1. When administered in combination with PegIFN/RBV, these protease inhibitors resulted in marked improvements in treatment outcomes in all patient groups including historically “difficult to treat” populations. SVR24 rates exceeding 75% were observed in several of the registration clinical trials [210–213]; however, adverse events and the emergence of drug resistance in patients failing therapy limited their effectiveness over the long term [210,212,234,235]. The following years have seen the rapid development and approval of several new direct-acting antivirals in four drug classes [198]. The second generation NS3 protease inhibitor simeprevir, was approved in 2013 for use in combination with PegIFN/RBV. Treatment with as little as 12 to 24 weeks of simeprevir in combination with 24 to 48 weeks of PegIFN/RBV resulted in SVR12 rates of approximately 80% [214,215]. The first nucleoside analogue inhibitor of the HCV NS5B polymerase, sofosbuvir, was approved in late 2013 [216,217]. Sofosbuvir has pan-genotypic activity and it remains to date the only direct-acting antiviral to be approved in North America for treatment of non-GT1 HCV. In addition, the approval of sofosbuvir marked the availability of the first all-oral, interferon-free regimens which reduce the negative side effects associated with PegIFN/RBV: sofosbuvir plus ribavirin for GT2 or GT3 infection, sofosbuvir plus simeprevir for GT1 infection, and sofosbuvir co-formulated with the NS5A inhibitor ledipasvir for GT1 infection. SVR12 rates for sofosbuvir combination therapy often exceed 95% although SVR12 rates vary by patient population, HCV genotype, HIV serostatus, and duration of treatment [216–220]. For example, an 8 week course of sofosbuvir co-formulated with ledipasvir is sufficient to achieve SVR12 in 95% of patients with pre-treatment HCV viral loads under 6.0 log10 IU/mL [219]. Finally, the triple-class DAA combination of the NS5A inhibitor ombitasvir, co-formulated with the NS3/4A protease inhibitor paritaprevir (boosted with ritionavir) plus the non-nucleoside NS5B polymerase inhibitor dasabuvir, received FDA approval in late 2014. Approved for use in GT1 infection only, SVR12 rates exceeding 95% were 24  observed in most populations, including patients who had previously failed PegIFN/RBV therapy [236,237].  The emergence of opportunistic infection, AIDS diagnosis, and death were the primary endpoints used to evaluate treatment efficacy in the initial clinical trials of novel antiretroviral compounds. It was recognized early on that meaningful surrogate markers of disease progression, rather than long-term outcomes, needed to be identified in order to advance the development of new therapeutics and to monitor individual patients’ disease progression; clinical trials using AIDS-related mortality as an endpoint took several years to complete [238]. Absolute CD4 lymphocyte count, an indirect measure of immune function that correlates well with progression to AIDS [239,240], was the first widely used prognostic marker of HIV disease progression. For example, early clinical trials demonstrated that didanosine monotherapy resulted in significant increases in CD4 counts over pre-therapy levels. These findings, rather than long-term follow-up of disease progression or mortality, helped expedite the approval of the nucleoside analogue didanosine [241,242]. However, it was soon demonstrated that short-term increase in CD4 count was an incomplete correlate of long-term survival [243–245].  It was not until the mid-1990s that a direct virologic marker of HIV disease progression became available: the plasma viral load (pVL) assay. Commercial kits using reverse transcription polymerase chain reaction (RT-PCR) amplification [246,247], nucleic acid sequence-based amplification (NASBA) [248], or branched-chain DNA (bDNA) signal amplification [249,250], were developed to quantify the amount of HIV RNA circulating in peripheral blood. Shortly thereafter, it was demonstrated that CD4 count and the magnitude of plasma viremia were independent predictors of HIV disease progression [251]. In untreated infection, patients with elevated setpoint viral loads were at greatly increased risk of progression to AIDS or death regardless of their CD4 count [100,252,253]. Furthermore, the degree of reduction in plasma viral load upon administration of combination antiretroviral therapy was demonstrated to be highly predictive of improved clinical prognosis, AIDS-free survival and decreased 25  mortality [254–257]. Specifically, the ability of HAART to limit viral replication, thus reducing and maintaining plasma viremia at low (<10,000 HIV RNA copies/mL) or ideally undetectable levels, is essential to ensuring long-term survival [258,259]. This key finding was illustrated in several pivotal clinical trials of structured treatment interruptions [260–264]. In the early 2000s treatment interruptions were becoming an increasingly popular strategy. Briefly, patients were removed from therapy once CD4 counts were restored to a certain threshold with a goal of reducing adverse events, drug toxicity and limiting the selection of drug resistance [260]. In theory, this strategy would preserve the limited number of drug options available at the time. Treatment was re-initiated either after a predetermined length of time, or once CD4 counts declined below specified thresholds. Several large clinical trials evaluated the effect of treatment interruption versus continued therapy on the development of opportunistic infection, progression to AIDS or mortality [260–264]. These studies clearly demonstrated that the increased HIV viremia associated with interrupting HAART was associated with increased disease progression, opportunistic infection and all-cause mortality. These effects were so pronounced that the SMART, DART and Trivacan studies were all stopped early [260–264].  Importantly, it was noted that the majority of deaths in the SMART trial were not attributed to opportunistic disease, rather non-AIDS-defining cancers and cardiovascular disease. In addition, higher rates of major renal, or hepatic disease were observed in the treatment interruption group. Thus, the SMART study was one of the first to suggest that persistent HIV viremia is associated with chronic immune activation and inflammation which in turn drives increased risk of cardiovascular disease [260,261,265,266]. Historically, treatment guidelines have recommended that HIV treatment be initiated when CD4 counts decline below a given threshold [191,267,268]. Over time these thresholds have increased to the point where, in most developed countries, treatment is recommended for all HIV-infected persons regardless of CD4 count [191,268]. The goal of treatment is the rapid (within 24 weeks) suppression of plasma HIV RNA to below 50 copies/mL, the lower limit of detection of first-generation 26  ultrasensitive pVL assays, and the long-term maintenance of HIV viremia at this level. Routine monitoring of CD4 counts and pVL is essential to the success of therapy; confirmed pVL rebound to detectable levels on two consecutive measurements is suggestive of treatment failure and should prompt immediate investigation of adherence, regimen tolerability or the development of drug resistance [191,268]. The recent development of commercial viral load assays with lower quantification thresholds (20 to 40 HIV RNA copies/mL) have prompted investigations of the clinical relevance of low-level viremia and have led to revisions of the definition of treatment failure. The implication of the approval of these new tests is discussed further in Chapter 3.  A major barrier to successful treatment with HAART is the development of antiretroviral resistance. The lack of proofreading capacity of the HIV reverse transcriptase enzyme leads to a high error rate during reverse transcription. These errors, coupled with the high rate of HIV replication, lead to the rapid diversification of the viral population within the first months following infection [46,48]. While many of these mutations can result in viruses with impaired replicative capacity, other mutations or combination of mutations can confer decreased susceptibility to antitretrovirals. Thus, when viral replication is allowed to continue under drug selection pressure, resistant HIV variants can emerge. Given the high genetic diversity of HIV within a single untreated individual, minority resistant quasispecies exist even before treatment is initiated [48]. In this case, resistant variants can be selected in a matter of weeks. This was demonstrated following the first clinical trials of NRTI mono- and dual-therapy; these early regimens could only partially inhibit viral replication which led to the selection of drug-resistant variants after only a few months of treatment [182,184]. In some cases, a single exposure to antiretrovirals, such as in the perinatal administration of single-dose nevirapine to prevent mother-to-child transmission, is sufficient to select for low-but-detectable levels of drug-resistant variants [269,270]. 27  In contrast, modern HAART regimens are capable of completely inhibiting HIV replication [187,188]. Furthermore, as HAART involves treatment with multiple drug classes, emergent variants that are resistant to one drug class in the combination may still be susceptible to inhibition by the other components. This is clearly demonstrated by examining the history of antiretroviral treatment in British Columbia; the incidence of new cases of drug resistance has declined steadily since the advent of HAART, concomitant with a marked increase in rates of viral suppression [271,272].  Under HAART, incomplete therapy adherence is the major driving force behind the development of drug resistance [273,274]. Patients who are able to reliably take their prescribed medications are significantly more likely to suppress HIV viremia to undetectable levels. In these individuals, low levels of HIV viremia can be detected by ultrasensitive assays [275]; however, this viremia is thought to be the result of virus release from the reservoir of latently infected cells rather than new cycles of active infection and replication [275–278]. In the absence of active HIV replication, the development of drug-resistant variants is unlikely [279,280]. Conversely, individuals who are entirely non-adherent maintain high levels of viral replication; however, in the absence of any drug selective pressures, development of resistance is also unlikely. The problematic scenarios arise when individuals exhibit intermediate levels of adherence, or when impaired metabolism limits drug bioavailablity – these patients are at highest risk to develop drug resistance. In such individuals, ongoing viral replication in the presence of sub-therapeutic drug levels can quickly select for drug-resistant variants, leading to treatment failure [273,274]. The relationship between therapy, plasma viral load and ongoing viral evolution will be discussed further in Chapter 4. Once selected, drug-resistant HIV variants can be transmitted to new recipients, thereby potentially compromising treatment options in these individuals. Primary infection with NRTI-resistant HIV was first reported in 1993, six years after the approval of zidovudine [281]. Since the advent of HAART, multiple studies have described the transmission of antiretroviral resistant HIV. The population-level prevalence of transmitted drug resistance varies by country and population, depending on the history of HIV treatment in those settings. On average 8 to 12% of new infections in North America and Europe 28  are with HIV variants harboring at least one drug resistance mutation [282–285]. Of note, there is a trend towards increasing rates of transmission of drug-resistant HIV variants in some jurisdictions. Of particular concern is the increasing incidence of transmitted NNRTI resistance in developing countries, as the WHO-recommended first-line HAART regimens are based on this class [286–288].  Both transmission of drug-resistant HIV and the emergence of drug-resistant variants during the course of infection is associated with poorer treatment outcomes, including earlier mortality [289]; the presence of resistant variants prior to therapy initiation is associated with poorer virologic response to that regimen [290]. In addition, patients with emergent drug resistance who remain on a failing regimen are at risk of developing additional resistance mutations to the remaining drug classes in the combination. Physicians therefore use the results of HIV drug resistance testing to help guide treatment decisions by modifying regimens if drug resistance is detected. Several prospective studies have demonstrated that patients whose physicians have access to HIV drug resistance testing have improved treatment outcomes over those who do not [291–294], and that routine drug resistance testing is cost-effective [295,296]. As such, most treatment guidelines in resource-rich settings advocate the routine use of drug resistance monitoring both prior to HAART initiation as well when treatment failure is suspected [191,267,268]. Assays testing for HIV drug resistance can be classified into two broad categories: phenotypic assays that measure changes in virus susceptibility to drug in culture, and genotypic assays that use sequence data to infer drug susceptibility.  In a modern phenotypic HIV drug resistance assay, the HIV gene encoding the drug target is RT-PCR amplified from extracted viral RNA and subsequently inserted into a HIV vector by homologous recombination. A cell line is then transfected with the recombinant HIV and cultured in the presence of serial dilutions of drug. Viral replication is typically quantified through the expression of a reporter gene, for example the luciferase gene which replaces HIV env in the Monogram Phenosense assay. Resistance results are reported as the fold change of drug concentration required to reduce HIV 29  replication by 50% (IC50) relative to the wild-type virus [297,298]. However, phenotypic drug resistance assays are both time- and labor-intensive: generating a complete resistance profile requires culturing the recombinant HIV in the presence of multiple concentrations of all licensed drugs in the class targeted by the inserted gene – a process that may take up to a month to complete.  In contrast, genotypic HIV drug resistance assays can be performed for lower cost and at higher throughput. Antiretroviral resistance is associated with sequence changes in the viral genes targeted by drugs [299]. In a genotypic resistance test, the viral genes targeted by the drug of interest are amplified from extracted HIV RNA by RT-PCR and sequenced, typically on an automated Sanger sequencing instrument [300]. Multiple sequencing reactions are used to span the region of interest; bases are called from individual sequence chromatograms, which are then trimmed of poor-quality regions and assembled into a single contiguous sequence (the details of sequence analysis are discussed further in Chapters 1.3.1 and 2). The resulting sequence is then analyzed by one of several drug resistance interpretation algorithms. In a “rules-based” algorithm such as the Sierra web service implemented by the Stanford HIV Drug Resistance Database [301], sequences are examined for the presence of known or suspected resistance mutations. These lists of mutations are determined by observing mutations that frequently emerge in patients failing therapy, or in long-term in vitro culture [299,302]. Each mutation in the list is assigned a score that reflects the degree of resistance that it confers to each drug in the class. Summing the scores of each detected mutation results in a total resistance score for the sample, the magnitude of which is used to classify the sample as being susceptible, having decreased susceptibility or being resistant to each individual drug. Other more sophisticated algorithms, such as the modified vircoTYPE HIV-1 resistance test algorithm used in British Columbia, compare the HIV sequences to databases of linked sequence, phenotype and patient outcome data and estimate a “virtual phenotype” IC50 score using mathematical models [303]. Drug susceptibility is reported based on clinically-relevant IC50 cutoff values inferred from the linked treatment outcome data [304]. Concordance between genotypic resistance interpretation methods is generally high, though drug class specific variation is observed between methods, particularly for NRTI [305–307]. Despite these scoring differences, commonly-used resistance interpretation 30  algorithms appear to perform equally well at predicting virologic outcomes in patients initiating or switching therapy [308,309]. The main limitation of the current generation of licensed genotypic drug resistance assays employing traditional Sanger sequencing is their inability to reproducibly detect low-frequency drug-resistant variants, particularly in samples with low viral load [269,310–312]. Increasing evidence suggests that the presence of minority drug-resistant variants negatively impacts treatment success [313–317], although some studies have failed to observe a statistically significant association [270,318]. The effect may be limited to treatment with NNRTI-based regimens, as several studies have failed to demonstrate that boosted protease inhibitor- or integrase inhibitor-based regimens are negatively affected by the presence of low-frequency resistant variants [318–321]. The critical threshold at which these rare variants become clinically relevant remains to be defined; however, it is not unreasonable to assume that resistant variants comprising as little as 1 - 2% of the viral population could have a substantial impact on treatment outcome. For example, retrospective analysis of the maraviroc registration trials demonstrated that viral tropism assays (which are not, strictly-speaking, drug resistance tests) must be able to accurately detect minority populations – patients harboring virus populations consisting of as little as 2% CXCR4-using HIV have poorer virologic outcomes when treated with co-receptor antagonists [322,323]. Consequently, a current research priority is the development of validated, sensitive minority variant detection methods, such as “deep” sequencing on next-generation instruments (see Section 1.3.2).   Methods for DNA sequencing, the determination of the order of nucleotides in a DNA molecule, were first developed in the early 1970s. The chain-termination method developed by Frederick Sanger and colleagues in 1977 represented a significant improvement over previous methods; it both simplified the labor-intensive procedure and allowed longer sequences to be determined [324]. However, it was 31  not until the mid-1980s when automated versions of the Sanger method became available that DNA sequencing became a widely accessible molecular biology tool [325,326].  The Sanger sequencing method (also known as chain-termination or dideoxy sequencing) uses a DNA polymerase enzyme and primer to synthesize copies of a single-stranded DNA template in a reaction similar to PCR amplification. Sanger’s innovation was the inclusion of chain-terminating dideoxyribonucleoside triphosphates (ddNTP) in the reaction. As ddNTP lack the 3’ hydroxyl group required to form the phosphodiester bond between neighboring nucleotides, the incorporation of a ddNTP during synthesis causes strand extension to stop. Random incorporation of low concentration (relative to regular dNTP) ddNTP into the growing DNA chain results in a mixture of synthesized DNA molecules of varying length. If only a single type of ddNTP (i.e. ddATP, ddGTP, ddCTP, ddTTP) is included in the reaction, then the terminal base in the chain is known (more precisely, the complementary base to the original DNA template). Four separate sequencing reactions are thus performed each with a different ddNTP. In the original method the resulting mixture of products are separated by length electrophoretically in four lanes of a dense polyacrylamide gel which is then visualized radiographically or by ultraviolet light. The sequence is then “read” off the resulting image from the bottom to top by comparing the relative vertical positions of the dark bands across the four lanes [324]. The original Sanger method used radioactively-labeled primers to expose X-ray film after an hours- to days-long gel electrophoresis step [324]. In the mid-1980s, the sequencing procedure was further simplified through the use of fluorescently-labeled primers [327] or ddNTP [326]. By tagging each type of ddNTP with a different fluorescent dye, only a single sequencing reaction and single lane on the gel, rather than four, was required. Furthermore, the ability to read bases by “color” rather than position across four lanes allowed computer-assisted reading to replace manual interpretation. In the mid-1990s, polymer-based electrophoresis in hollow glass capillaries replaced gel slabs [328], resulting in significant time savings. Multiple sequencing reactions could be run and read in parallel. In dye-terminated capillary sequencing, the identity of the terminal base in a DNA fragment is 32  determined as the fragment migrates down the capillary and passes through a fluorescence detector. In the detector, laser excitation of the fluorophore attached to the ddNTP causes it to emit light of a specific wavelength. Emitted light is detected by a photomultiplier tube after passing through one of four bandpass filters mounted on a rotating wheel. The identity of the filter positioned in front of the detector at the time light is detected determines the identity of the base sequenced. The captured data are represented in a four color chromatogram trace – essentially a record of luminous signal intensity over time in each of the four color channels [325,326]. Software is used to convert the colored peaks of the chromatograms into nucleotide base calls of the original template sequence. By examining the peak shapes and spatial distribution as well as the background “noise” in the trace files, individual base calls can also be assigned quality scores that represent the probability of the base call being correct [329,330]. When a heterogeneous population is sequenced, such as a portion of HIV amplified from a plasma sample from a chronically-infected individual, resulting fragments of identical length may have different terminal bases. When sequenced, these nucleotide “mixtures” are depicted on the chromatogram as two (or more) differently-colored, overlapping peaks at the same position. In such cases, the relative height or area of the overlapping peaks can give an approximate measure of the relative abundance of the two nucleotides at that position [331,332]. However, given the inherent variability in the RNA extraction, RT-PCR, and sequencing reactions, only bases representing a minimum of 20 to 25% of the population are reliably and reproducibly detected [332]. On average, approximately 1% of nucleotide positions in protease and reverse transcriptase in isolates from antiretroviral-treated individuals have detectable mixtures by population sequencing [333]; however, these mixed-base positions account for the overwhelming majority (>90%) of sequence discordances in studies of intra- and inter-lab variability of genotypic drug resistance testing [312,333,334]. Furthermore, most commercial sequencing software does not automatically call mixtures. Manual human review of chromatograms is typically required to call mixtures and the subjective nature of this procedure results in a substantial amount of inter-operator variability [335]. Previous attempts to automate the sequence interpretation step for HIV drug resistance testing have had poor results. When 33  compared to manually-curated sequences, substantially fewer mixtures were called by automated methods [336]. Chapter 2 will present a validated solution to these problems.  The past decade has seen the rapid development of a “next-generation” of technologies that have revolutionized how DNA sequencing is performed. In contrast to the Sanger method which generates a single “population” sequence representing the most prevalent variants, these novel methods perform parallel clonal sequencing of individual DNA molecules in a library. These instruments generate incredible amounts of data per run, enabling large genomes to be sequenced economically, or allowing high-resolution detection of rare variants. The latter is often termed “deep sequencing”: small numbers of samples are sequenced with great multiplicity (depth) of coverage, allowing the detection of low-frequency members of the viral population. In deep sequencing applications, individual samples are tagged with unique “barcode” sequences (typically incorporated into the amplification primer) to allow individual sequence reads to be assigned to a specific sample during post-run data processing. While each instrument differs in its chemistry and detection methods, they share a similar principle of operation: After library construction, short DNA fragments are anchored to a solid medium so that their physical positions are fixed and known. The fragments are then clonally amplified in place. The immobilized, amplified fragments serve as templates for the sequencing reaction in which the incorporation of individual bases is detected as the complementary strand is synthesized. Since the DNA fragments are immobilized, the detection of base incorporation can be linked to individual templates, allowing clonal sequences to be determined [337]. A third generation of instruments forgo the clonal amplification step, allowing individual DNA strands to be probed directly [337]. These instruments, however, are not yet in widespread use. What follows is a brief description of some of the most commonly-used platforms for viral sequencing. 34   Researchers at 454 Life Sciences (later acquired by Roche Diagnostics) developed the 454 GS20 instrument by adapting the pyrosequencing technique first described in the late 1980s [338]. Pyrosequencing is a type of “sequencing-by-synthesis”; rather than detecting fluorescently-labeled chain terminators, pyrosequencing monitors the activity of DNA polymerase in real time [338,339]. The term “pyrosequencing” is derived from the detection of pyrophosphate that is released upon nucleotide incorporation. Briefly, the four nucleotides are sequentially washed (T, A, C, G) over the immobilized templates along with DNA polymerase, adenosine triphosphate (ATP) sulfurylase, luciferase and luciferin. The incorporation of a nucleotide complementary to the template strand results in the release of pyrophosphate. The ATP sulfurylase converts pyrophosphate to ATP, which is in turn used by luciferase to oxidize luciferin resulting in the release of visible light. Emitted light is detected by digital camera imaging after each nucleotide flow. As chain-terminating nucleotides are not used in the reaction, multiple nucleotides can be incorporated in a single cycle. In such a case, the intensity of the emitted light is proportional to the number of bases incorporated. After imaging, unincorporated nucleotides are degraded by apyrase prior to the next cycle [338,339].  The innovation brought by 454 Life Sciences was to convert pyrosequencing from a “bulk” sequencing technique run in 24- or 96-well microplates, to an array-based technique that allows imaging and detection at the single-template level for hundreds of thousands of templates at once [340]. First, biotinylated adaptors are added to the ends of short (300 to 800-bp long) DNA fragments either by ligation or incorporated into PCR amplification primers. Biotinylated DNA fragments are captured on streptavidin-coated microbeads. The bead to template ratio in the reaction is controlled such that, in the majority of cases, only a single template is immobilized on a single bead. The immobilized templates are then subjected to emulsion PCR [340,341]. Beads, DNA and PCR reagents are mixed with oil and shaken to form an emulsion. Each bead-template complex is thus isolated in an aqueous “microreaction” droplet which is subjected to thermal cycling. The immobilized templates are clonally amplified to cover the surface of the beads, which serves to amplify the luminous signal in the 35  subsequent pyrosequencing reaction. The emulsion is broken and the beads are deposited into the miniature (~44 μm diameter, 75 pL volume) wells of a PicoTiterPlate (a transverse slice of a large fiber optic bundle that is subjected to a surface etching treatment to form individual wells) along with DNA polymerase and additional “enzyme beads” containing sulfurylase and luciferase. Centrifuging the PicoTiterPlate forces a single DNA bead into a well and packs it into place with enzyme beads. The pyrosequencing reaction proceeds as described above [340]. Being the first commercially-available next-generation sequencer, 454 instruments became the most widely used platforms in pivotal studies of HIV deep sequencing [322,323,342–346]. At the time, 454 technology was particularly suited for HIV-1 resequencing applications as the platforms offered the longest read lengths; for example, the “titanium” chemistry of the 454 GS FLX instrument resulted in read lengths of over 400 base pairs, in contrast to other contemporaneous instruments which were limited to short reads of approximately 35 to 100 base pairs [337,347]. Despite their widespread use, 454 instruments suffer from elevated sequencing error rates, primarily in the form of insertion/deletion errors in homopolymer regions [337]. Since multiple bases can be incorporated in a single flow, pyrosequencing must infer the number of bases in the homopolymer from the brightness of the spots on the captured images. Unfortunately, the relationship between light intensity and number of nucleotides incorporated is non-linear. Homopolymeric stretches exceeding approximately six bases show increasingly poor base calling accuracy [337]. This results in frequent under-or over-calling of homopolymeric bases. Overall, these insertion/deletion errors, rather than base misincorporation errors, contribute to an average error rate on the order of 1% per base [347]. This error profile is of particular concern to HIV-1 clinical sequencing applications as several important resistance mutations, including the reverse transcriptase K103N mutation conferring high-level NNRTI resistance, exists in highly homopolymeric regions. The popularity of 454 instruments has waned as alternative less expensive, higher capacity platforms with increasingly longer read lengths have become available [347]. As a result, Roche will discontinue support of the 454 platform in mid-2016.  36   In 2006, the Solexa Corporation (later acquired by Illumina, Inc.) released their first short read sequencer, the Solexa Genome Analyzer. Using reversibly-terminated, fluorescently-labeled nucleotides, Illumina instruments detect base incorporation into template strands bound to the surface of a “flow cell” [337,348]. The flow cell is an optically-transparent glass slide upon which two types of oligonucleotide anchors are covalently attached. Single-stranded DNA libraries prepared with complementary adaptor sequences are annealed to the flow cell anchors; the concentration of the DNA library determines the density of templates bound to the flow cell. The complementary strand is synthesized using the anchor oligo as a primer. The original template is removed by denaturation resulting in a single-stranded DNA template library affixed at one end to the flow cell surface. A “bridge amplification” PCR procedure is then used to generate clonal clusters of DNA templates. Briefly, the free end of the anchored template bridges to a complementary surface oligo from which the complementary strand is extended. Denaturation results in two surface-anchored DNA templates. Repeated cycles of annealing, extension and denaturing produce millions of localized clusters of clonal templates across the flow cell surface [348]. The Illumina sequencing-by-synthesis reaction proceeds after a chemical cleavage reaction that leaves only the positive strand templates attached to the flow cell. The sequencing reaction is initiated by annealing a primer complementary to the adaptor sequences, followed by the cyclic addition of polymerase and dNTP. Reversibly-terminated nucleotides, each labeled with a different fluorescent dye, are used. Thus a single nucleotide complementary to the template is incorporated in each cycle. After incorporation, excess nucleotides are washed away, and the identity of the incorporated base is determined optically following laser excitation of the fluorophore; four digital images per flow cell “tile”, each using a different bandpass filter, are captured per sequencing reaction cycle. The fluorescent dye terminators are subsequently removed chemically and the next cycle of base incorporation and imaging proceeds [337,348]. Enzymatic cleavage of the terminators has subsequently replaced chemical cleavage, allowing for shorter run times. Depending on the instrument 37  and reagents chosen, up to 300 sequencing cycles are performed. If “paired-end” sequencing is used, the sequencing step is repeated following a paired-end turnaround reaction that resembles one cycle of the cluster generation step. Since the Illumina sequencing chemistry uses reversibly-terminated nucleotides, it does not suffer from the same limitations as 454 sequencing in homopolymeric regions. Instead, the Illumina error profile is dominated by nucleotide substitutions. On average, an approximate error rate of 0.1% per base is observed [347]; however, this rate is not uniform across the length of a read. Error rates of 1% or more are typically seen towards the end of a 250-bp read, and random cycles with error rates exceeding 10% are occasionally observed. Despite slightly higher error rates and shorter read lengths compared to 454 sequencing, the substantially higher capacity (up to 13 gigabases for the MiSeq compared to 50 megabases for the 454 GS Junior), simplified workflow (no emulsion PCR) and greatly reduced reagent costs (approximately two orders of magnitude cheaper per base) of Illumina instruments have contributed to their emergence as the current platform of choice for clinical sequencing applications [347].  Over the past ten years, several other next-generation sequencing platforms have been released with varying levels of commercial success.  The Applied Biosystems (now Life Technologies) SOLiD platform uses a complicated sequencing-by-ligation technique [349] with a total of 16 specific 8-mer oligonucleotide probes. The first two bases of the probes are complementary to the template strand being sequenced while the remaining six bases are degenerate. One of four fluorescent tags is attached to the 5’ end. Cycles of ligation are performed in which a sequencing primer is annealed, 8-mers are ligated to the template strand, excess oligos are washed off, and the identity of the fluorescent dye is determined by imaging. The three terminal bases and the fluorescent tag are chemically cleaved between ligation cycles leaving behind a 5-bp probe. At this point, the identity of the first two bases is partially known (there are four possible nucleotide 38  combinations for each fluorescent tag) while the third through fifth bases remain undetermined. Following five to seven cycles of ligation and imaging the entire process is repeated four additional times with primers that are offset by one base. Thus each base in the sequence is interrogated at least twice and its identity is deconvoluted from the two observed fluorescent tags [337,350]. Two-base encoding results in a relatively low sequencing error rate; however, limitations of the SOLiD platform include short read lengths, long runtimes and difficulties sequencing palindromic motifs [337,347]. The Ion Torrent Systems (now Life Technologies) Ion Personal Genome Machine was the first “semiconductor sequencing” instrument to use electronic rather than optical detection of nucleotide incorporation. Following emulsion PCR amplification, bead-bound DNA templates are added to microwells on a manufactured semiconductor chip. Similar to 454 sequencing, individual nucleotides are cyclically flowed across the chip. Incorporation of the correct complementary base (or bases) releases pyrophosphate and a hydrogen ion. The resulting pH change of the solution is detected by the semiconductor circuitry, the magnitude of which is proportional to the number of bases incorporated in that cycle. No complicated optics or modified nucleotides are used in the sequencing procedure and instrument and reagent costs are low; however, similar to pyrosequencing, high error rates are observed in homopolymeric regions [347,351]. Third-generation sequencing technologies are able to observe DNA synthesis in real time, often without requiring a PCR amplification step [351]. For example, in Pacific Biosciences Single Molecule Real-Time (SMRT) sequencing, single DNA polymerase enzymes are immobilized in microscopic wells termed “zero mode waveguides” along with a single single-stranded DNA template. The incorporation of nucleotides is optically detected as a phospholinked fluorophore is cleaved off as part of the extension reaction. The size of the zero mode waveguide is such that fluorescence can only be detected near the bottom of the well where the DNA polymerase is anchored [352]. The SMRT technology is thus able to generate multiple kilobase-long reads in short runtimes at the expense of high error rates (up to 15%) [347,351,352]. Circularizing the DNA template allows it to be sequenced multiple times in succession and error correction to be performed on the replicate sequences. Improvements in error 39  rate, however, are at the expense of read lengths: three to five successive sequencing passes are required to achieve error rates under 1% [353]. Each of the current next-generation sequencing platforms has its relative benefits and disadvantages. Multiple platforms have already been used to sequence HIV and other diverse populations, including studies presented in this thesis. The Roche 454 GS Junior instrument was used in the studies presented in Chapter 5. Given the previously described technological limitations of this platform, alternate technologies were subsequently investigated: The suitability of the IonTorrent Personal Genome Machine for HIV sequencing was explored in experiments not presented here. Sequencing-by-synthesis on the Illumina MiSeq was used for the studies presented in Chapters 6 and 7. Ultimately, cost, capacity, accuracy and ease-of use may determine which instruments come to the forefront in clinical applications.   This thesis is divided into eight chapters. Chapter 1 provides an overview of the HIV epidemic as a whole. It contains a review of viral structure, replication, diversity and pathogenesis. The effect on the host is also discussed in a summary of the natural history of infection, and the history and current availability of HIV treatment options. A similar, concise review of Hepatitis C Virus follows. Finally, an overview of clinical monitoring tools including sequencing technologies is presented. A particular focus is placed on the clinical utility of viral sequencing in infectious disease management and the history of the development of these technologies.  Chapters 2 through 7 address the primary objectives of this thesis – to develop novel assays and analysis tools for use in a clinically-accredited laboratory setting and to provide evidence for the ongoing role of viral sequencing in infectious disease management. Four main hypotheses are investigated: 40  1. Bioinformatic software can play an important role in standardizing subjective laboratory processes, and improving throughput, consistency and quality. 2. Viral sequencing can be used as independent evidence when investigating suspected anomalous assay results. 3. Longitudinal monitoring of HIV sequences can give insight into evolutionary processes and viral fitness. 4. The future of personalized approaches to infectious disease treatment lies in next-generation sequencing platforms, the adaptation of their current use to new applications, and the extension of existing techniques to new viral targets. Specifically, Chapter 2 presents the performance and validation of novel sequence analysis software that is used extensively in the subsequent studies. Chapter 3 investigates seemingly anomalous results from a Health Canada and FDA-approved laboratory test and uses HIV sequencing to distinguish assay false positives from clinically significant values. Chapter 4 uses Sanger sequencing of longitudinal samples to study whether HIV continues to evolve in response to host immune selective pressures after antiretroviral treatment is initiated. Similarly, Chapter 5 uses next-generation sequencing of short-term viral cultures to measure the relative fitness of several antiretroviral-resistant HIV variants. Finally, new applications of existing technologies and concepts are explored. Chapter 6 examines how the capacity of next-generation sequencing instruments can be leveraged to maximize the breadth of samples tested rather than the depth of sequence coverage and proposes an application of this method to HIV drug resistance testing in resource-limited settings. Chapter 7 extends genotypic drug resistance testing to a different virus, HCV, and presents the development and validation of two independent assays aimed at screening for resistance to a new class of antiviral agents. The secondary goal of several of these studies is the direct translation of a test or tool into clinical practice.  To conclude, Chapter 8 summarizes the research findings, discusses their relevance and comments on their implications, impact and contributions to the field. Tables and Figures are placed after their first 41  mention in the text. All references are presented in a single Bibliography after the final chapter. Additional supporting materials are presented in the Appendix. 42    Human immunodeficiency virus (HIV) drug resistance genotyping has been used for over ten years in clinical practice to help guide and tailor highly active antiretroviral therapy (HAART) regimens [292]. By identifying resistance mutations in the areas of the viral genome targeted by antiretroviral drugs genotypic drug resistance testing allow physicians to optimize antiretroviral therapy regimens for each patient, increasing the chance of successful virological suppression [290,292] and subsequently reducing the overall cost of treatment by minimizing the use of ineffective drugs and avoiding treatment failure-related inpatient care [296]. The predominant methodology used for HIV genotypic drug resistance testing involves RT-PCR of extracted viral RNA from plasma followed by population-based (bulk) sequencing [300]. As multiple sequencing primers are required to provide bidirectional coverage over the entire length of the amplicon, individual DNA sequence reads are then assembled into a contiguous consensus sequence using analysis software. Commercially-available HIV drug resistance genotyping kits such as TRUGENE (Siemens, Deerfield, IL) [354,355] and ViroSeq (Abbott Laboratories, Des Plaines, IL) [356] are distributed with custom analysis software, however simple software solutions do not exist for in-house developed genotyping methods. The available “generic” sequence analysis programs require considerable hands-on time; highly trained technicians must first inspect each trace file and trim out regions of problematic or low-quality sequence before manually specifying the sequence reads to assemble. The sequence assembly is subsequently verified by slow and labor-intensive human visual examination. A final consensus sequence is then exported and processed with a drug resistance interpretation algorithm.  43  Drug resistance mutation reporting often varies between laboratories even when identical samples are tested [333,334]. While many inter-laboratory discrepancies can result from differences in sample preparation (e.g. primer choice, stochastic variation) variation may be introduced by technicians as they subjectively review the assembled sequences [335]. As drug-resistant HIV variants may be present at low frequency in clinical isolates, accurate identification of nucleotide “mixtures” (positions where two or more nucleotides are observed) is required. Differences in individual technicians’ propensity to identify low-level nucleotide mixtures could result in clinically-relevant drug resistance mutations being missed [335,354]. In order to minimize erroneous HIV drug resistance reporting and optimize genotyping protocols, clinical and research laboratories often participate in external quality assessment programs (EQA), where identical samples are sent to multiple laboratories for independent analysis [333,334]. Unfortunately, due to the complications of subjective sequence interpretation, sources of any aberrant results can be difficult to ascertain. The implementation of an automated sequence analysis tool would enable objective and consistent interpretation of HIV genotype data and would provide considerable practical advantages; most notably improvements in processing speed, and significantly decreased labor and software costs. At the British Columbia Centre for Excellence in HIV/AIDS (BCCfE), we have developed a bioinformatics tool, RECall, to address these challenges. RECall is a pipeline for assembling, aligning, analyzing, and finishing sequence chromatogram files and has been tailored specifically for HIV genotyping. While these steps are performed by most sequence assembly programs, RECall has been specifically designed to reproducibly call nucleotide mixtures. RECall is available for free as a web application (http://pssm.cfenet.ubc.ca/).  Here, we present the results of an external validation of automated RECall analysis of sequence data generated by an independent laboratory. 44    HIV genotyping was performed at the Stanford University Hospital Diagnostic Virology Laboratory (Stanford, CA). Clinical genotypic resistance testing was performed on 981 sequentially-collected plasma samples using a previously described approach [333]. Briefly, plasma virus extraction and purification was performed on Qiagen BioRobot M48 or QIAsymphony SP automated nucleic acid extraction instruments (QIAGEN Inc, Valencia, CA) followed by one-step RT-PCR and nested second round PCR. Direct bidirectional sequencing encompassing HIV-1 protease (PR) and the first 296 codons of reverse transcriptase (RT) was performed on an ABI 3730 sequencer (Life Technologies, Carlsbad, CA, USA). Chromatograms were created using Sequencing Analysis v 5.2 (ABI). Nucleotide mixtures (positions containing two or more nucleotides, with the minor peak ≥20% of the major peak height) were marked with 3730 Data Collection Software v 3.30 (ABI). A laboratory technologist assembled the sequence trace files for each sample and generated a consensus sequence using Lasergene SeqMan v 8.0 (DNAStar, Madison, WI). SeqMan uses user-defined parameters to identify positions in the assembly having potential "conflicts". Conflicts are any nucleotide positions called a mixture (20% threshold), positions where overlapping sequences do not have the same base call, and any “N” calls that the sequencer could not distinguish. The analyzing technologist visually inspected each sequence, stopping at each conflict, making manual edits where necessary. The edited sequence was then inspected by a second technologist who verified the conflicts and any manual edits that were made.  Raw ABI chromatograms were re-analyzed using RECall. The software requires a consistent file naming convention to automatically group multiple sequence reads (primers) belonging to the same sample into a single consensus sequence (contig). 45  The sequencing trace files (.ab1) are first processed with the software package phred [329,330] which calls bases and assigns quality scores to each nucleotide. In a trace file, a “mixed” or “ambiguous” base is represented by overlapping peaks. When calling bases, phred determines the location and area of the primary peak (“called base”) and the largest secondary peak (“uncalled base”) in the trace file. As the primary and secondary peaks are often offset, RECall attempts to align the peak positions to their corresponding locations in the .ab1 files. The quality scores that phred assigns are a measure of the accuracy of the base call. Regions of poor sequence quality at the beginning and end of fragments are identified and trimmed automatically by RECall. Phred quality scores are also used to identify low-quality regions (phred score <20) within a fragment, which are also flagged and excluded from the final contig assembly. Grouped fragments are assembled and aligned to a user-supplied reference sequence (e.g. HIV-1 HXB2, GenBank Accession K03455) using a modified Smith-Waterman algorithm [357]. For this study, all chromatograms (.ab1) were submitted to RECall in a single batch and were processed without any human intervention using a standard desktop PC (Intel Core-i5 660 3.33 GHz CPU, 3GB RAM, Windows XP).  The most important feature is the process by which RECall calls ambiguous nucleotides (mixtures). Following the assembly and alignment step, RECall identifies mixtures based on the quality and area under the curve of the called and uncalled bases as determined by phred. The RECall configuration variables for mixture calling for clinical drug resistance testing at the BCCfE are listed in Table 2.1. Each position in the sequence alignment is examined sequentially. At each position, a list is first generated by counting the frequency of each nucleotide that appears as either a called base or an uncalled base meeting the Mixture Area criterion. This list is then reduced to include only nucleotides that are observed in at least half the sequence reads. The list is then ordered by frequency and the most common (majority) and second most common (secondary) bases are retained. If the secondary base is called with greater than half the frequency of the majority base, a mixture is called. If two bases tie for 46  most common, but no majority is achieved, then a mixture is called from those two bases. Finally, if none of these conditions is achieved then the most commonly called base is used. As phred is limited to calling a maximum of two nucleotides per chromatogram peak (“called” and “uncalled” bases), RECall does not call mixtures of three nucleotides, instead defaulting to the predominant two-base mixture.  Parameter  Value  Criteria Quality Censoring Cutoff phred <10 Bases with phred quality scores below the cutoff are excluded from the assembly. Mixture Area ≥20% The area of the uncalled peak must be at least 20% of the called peak area. If >50% of the reads pass this threshold then a mixture is called. Mark Area ≥17.5% The area of the uncalled peak must have at least 17.5% of the called peak area. If ≥50% of the reads pass this threshold then a mark is made. Mark Average Quality Cutoff phred <20 If the average quality of the base across all reads is below this value, a mark is made. Additional Marks  Insertions, deletions, and single primer coverage are also marked.  During the mixture calling step, RECall also “marks” potentially problematic sequences according to the parameters listed in Table 2.1. Insertions, deletions, low-quality and problematic positions are flagged for optional confirmation by a human user. In addition, positions meeting the Mark Area criterion are flagged for review in a manner similar to the mixture calling procedure (Table 2.1). In this study, mixtures and marks were not reviewed by human interpretation.  Sequences were passed or failed based on criteria established in the BCCfE laboratory which form the default parameters in RECall. Multiple quality checks are performed on every sample to ensure that 47  the sequence is acceptable. Problems leading to sequence rejection by RECall are listed in Table 2.2. If desired, these parameters can be modified by the user. Sequences that pass internal quality control checks are exported automatically as plain text or FASTA formatted files. Because RECall by default requires double primer coverage over the entire sequence length, some samples that Stanford deemed acceptable by human interpretation were rejected by RECall. In the following analyses, we included only those sequences that passed RECall’s default quality control criteria. Failure Reason Criteria Stop Codon Any unambiguous stop codons (TGA, TAA, TAG) Bad Inserts An insertion relative to the reference sequence that is not a multiple of three bases, resulting in a frameshift Bad Deletion A deletion relative to the reference sequence that is not a multiple of three bases, resulting in a frameshift Too many mixtures > 3.5% of nucleotides sequenced are called as mixtures N count ≥ 5 Ns (aNy base) in the sequence Mark count ≥ 100 positions “marked” as being potentially problematic Single coverage > 3 consecutive bases of single-read coverage with phred score < 40 Low quality Any section where the quality of all coverage is too low to make a call   Personal password-protected user accounts allow sequencing jobs to be saved and re-analyzed in the future without the need to upload files again. Two types of accounts are available; for traceability, operators with “User” level access are not given access to the program parameters, but may only process data using parameters provided by the local “SuperUser”. Processed sequences are retained on the RECall server for a user-chosen length of time after which they are automatically deleted. No submitted data is re-processed, collected, analyzed or used for any purpose, nor shared with anyone. 48   Nucleotide concordance between sequences generated by RECall and those determined by the Stanford laboratory using conventional manual human review (hereafter referred to as “human”) was calculated. A partial nucleotide discordance was considered to be present when one methodology reported a nucleotide mixture and the other reported one of the mixture’s components (e.g., human reported Y, RECall reported C). A complete nucleotide discordance was considered to be present if each method reported a different unambiguous nucleotide at the same position for a sample (e.g., human reported T, RECall reported C), or if an unambiguous nucleotide called by one method was not contained in a mixture called by the other (e.g. human reported G, RECall reported Y).  In addition to an analysis over the entire protease-RT sequence length, a specific analysis was performed comparing only antiretroviral drug-resistance mutation positions, which were defined as key resistance mutations recognized by the International AIDS Society (USA table) [358]. A drug resistance mutation was considered present if it was observed either alone or as part of an amino acid mixture. The Stanford HIV drug resistance genotyping web service Sierra [301] (http://hivdb.stanford.edu/pages/algs/sierra_sequence.html; Stanford University, Stanford, CA; Algorithm version 6.0.1), was used to infer antiretroviral drug susceptibilities from both human and RECall analyzed PRRT nucleotide sequences. Concordance of drug resistance interpretations was evaluated using Cohen’s Kappa statistic [359].   During software development at the BCCfE, RECall showed >99.5% agreement between human-reviewed and automated base calls when tested on in-house sequences (data not shown). We therefore wished to perform an external validation of the applicability of RECall to independently generated sequence data. 49  HIV protease-RT sequences and raw .ab1 sequence trace files were shared for 981 samples sequenced by the Stanford laboratory (with manual technician review) and re-analyzed by RECall. Of these, 875 (89.2%) met the default RECall acceptability criteria after automated processing. The primary reason for failure was lack of double primer coverage over the entire sequence length. Using a standard desktop PC (Intel Core-i5 660 3.33 GHz CPU, 3GB RAM, Windows XP) RECall completed base calling, assembly, and alignment in less than 6 hours with no hands-on analysis. In contrast, manual analysis required an estimated 150 hours of technician time.  There was 99.7% overall agreement in base calling between human and RECall over 1,036,875 analyzed bases. The rates of complete sequence concordance were 99.6% at 259,875 protease (PR) nucleotide positions, and 99.7% at 777,000 reverse transcriptase (RT) nucleotide positions (Figure 2.1). Of the 944 discordant PR nucleotides, 940 (99.6%) were “partially discordant” (mixtures called by one method, but not the other), and 4 (0.4%) were completely discordant. Of the 2535 discordant RT nucleotides, 2517 (99.3%) were partially discordant, and 18 (0.7%) were completely discordant. Most of the partially discordant bases (2530 of 3457, 73.2%) comprised nucleotide pairs resulting from transitions (R=A/G, Y=C/T) rather than transversions (K=G/T, M=A/C, S=C/G, W=A/T). The completely discordant positions were relatively equally distributed among transitions, transversions, and a combination of both (N=11, 6, and 5, respectively; Figure 2.1). Nucleotide mixtures were detected at approximately 1.1% of all bases, corresponding to 12.5 mixtures per 1185-bp PR-RT fragment. Overall, the human operator called a marginally higher number of mixtures (human: 10,996 mixtures, 1.06%; RECall: 10,921 mixtures, 1.05%; p=0.8). Positions with three-nucleotide mixtures (i.e., B, D, H, V) were automatically discordant because phred (and therefore RECall) is not programmed to recognize these. A representative sample of nucleotide positions with discordant calls by RECall and the human operator is shown in Figure 2.2.   50   Matrices depicting the frequency of nucleotides called by human operators (vertical axis) and by RECall (horizontal axis) in A) protease and B) reverse transcriptase. Concordant base calls are highlighted in green. Partially discordant base calls (mixtures called by one method, but not the other) are highlighted in yellow. Entirely discordant base calls are highlighted in red. Blank cells represent zero. International Union of Biochemistry and Molecular Biology ambiguity codes are as follows: R = A/G; Y = C/T; W = A/T; M = A/C; K = G/T; S = G/C; B = C/G/T; D = A/G/T; H = A/C/T; V = A/C/G. Columns for B, D, H, and V are not shown for RECall, as the software does not call three-base mixtures.51   The majority of differences between the two analysis methods were due to partial discordances in which one method called a nucleotide mixture while the other method called only one nucleotide component of a mixture. Depicted here are representative chromatogram traces of discordant mixture base calls. Panels A-C were positions called mixtures by human visual inspection, but not RECall. Panels D-F were positions called as mixtures by RECall, but not human interpretation. In each panel, the top line of text contains the consensus human base calls, while the lower line of text above the chromatograms are the consensus RECall base calls. The discordant mixtures are circled in orange. 52   The 944 discordant PR nucleotide positions resulted in 904 discordant PR codons. Of these, 380 (42.3%) resulted in non-synonymous discordances between the human and RECall interpretations when translated to amino acids: 378 (99.5%) were partial amino acid discordances (where at least one amino acid was shared between the two interpretations), while only 2 (0.5%) were complete amino acid differences. In RT, the 2535 discordant nucleotide positions occurred in 2469 unique codons. When translated to amino acids, 729 (29.5%) discordant substitutions were observed between the human and RECall interpretations: 724 (99.3%) were partial differences and 5 (0.7%) were completely discordant.  Overall, human and RECall sequence review identified 1096 (266 in PR, 830 in RT) and 1098 (269 in PR, 829 in RT) “key” antiretroviral drug resistance mutations (12), respectively either as complete amino acid substitutions, or as part of mixtures. In PR, the two methods were in agreement for 265 of cases (264; 99.6% in complete agreement). Human identified 1 PR resistance mutation that RECall did not while RECall identified 4 that human did not. Similarly, in RT the two methods both identified resistance mutations in 824 of cases (809; 98.2% in complete agreement). Human identified 6 RT resistance mutations that RECall did not while RECall identified 5 that human did not. In general it was not obvious which method was “correct”.  All 875 PR-RT sequences interpreted by both methods were submitted to Sierra, the Stanford HIV Drug Resistance Database genotyping tool (Algorithm version 6.0.1), and were scored for susceptibility to all currently available protease inhibitors (PI), nucleoside/nucleotide reverse transcriptase inhibitors (NRTI) and non-nucleoside reverse transcriptase inhibitors (NNRTI) . Briefly, Sierra identifies documented resistance mutations in each sequence and uses a rules-based algorithm to generate a resistance score for 19 PI, NRTI, and NNRTI [301]. A higher score indicates a greater probability of resistance. We calculated the susceptibility score differences between human- and 53  RECall-interpreted sequences. In total, 34 samples (3.9%) had discordant scores for one or more antiretrovirals (median 5 drugs). Of these, 17 samples had a score difference ≥10, with the maximum difference being 72. However, small differences in susceptibility scores may not translate into clinically-relevant differences in resistance. In addition to providing raw susceptibility scores, Sierra categorizes each sequence interpretation as susceptible (susceptibility score <15 = “S”), intermediate (15-59 = “I”), and resistant (≥60 = “R”). For simplicity, “I” and “R” interpretations were grouped together into a single “resistant” category. Only 13 samples (1.5%) had a discordant drug resistance interpretation for ≥1 drugs (median 2 drugs). Out of 16,625 scored drugs, only 35 (0.2%) had discordant resistance interpretations between human- and RECall-interpreted sequences (κ>0.98 for each drug) (Table 2.3). Of these discordances, 25 (71.4%) were cases where human calls resulted in a “susceptible” interpretation, while RECall did not. However, there was no statistically significant difference in the frequency of “resistant” interpretations between human- and RECall-analyzed sequences (13.2% vs. 13.3%; Chi-square p=0.82).   54  Class Drug Sierra Resistance Interpretation of Human/RECall Analyzed Samples (Number of Samples) S/S R/R S/R R/S NNRTI delavirdine (DLV) 710 164 1 0 efavirenz (EFV) 706 168 1 0 etravirine (ETV) 768 105 0 2 nevirapine (NVP) 706 168 1 0 NRTI lamivudine (3TC) 709 163 2 1 abacavir (ABC) 735 138 1 1 zidovudine (AZT) 734 139 0 2 stavudine (d4T) 726 148 1 0 didanosine (ddI) 737 136 1 1 emtricitabine (FTC) 709 163 2 1 tenofovir (TDF) 762 111 1 1 PI atazanavir/r (ATV/r) 783 89 2 1 darunavir/r (DRV/r) 839 35 1 0 fosamprenavir/r (FPV/r) 792 81 2 0 indinavir/r (IDV/r) 792 82 1 0 lopinavir/r (LPV/r) 807 66 2 0 nelfinavir (NFV) 777 95 3 0 saquinavir/r (SQV/r) 794 78 3 0 tipranavir/r (TPV/r) 816 59 0 0 Total 14402 2188 25 10  Resistance interpretations: S/S = scored susceptible by both methods; R/R scored resistant by both methods; S/R scored susceptible by human interpretation, resistant by RECall; R/S scored resistant by human, but susceptible by RECall. “Resistant” interpretations included both intermediate (“I”) and resistant (“R”) Sierra calls. Resistance interpretations from both methods were in near-perfect agreement (κ>0.98 for all drugs).  55   This study evaluated the performance of RECall, an automated sequence analysis tool developed by the BC Centre for Excellence in HIV/AIDS laboratory, to quickly and accurately interpret HIV genotypic data for drug resistance testing. RECall is available free of charge as a web application (http://pssm.cfenet.ubc.ca). We compared the results generated by RECall to human-verified sequences from the Stanford University Hospital Diagnostic Virology Laboratory, a well-recognized institution that has conducted routine HIV genotypic drug resistance testing for over ten years. Using a set of 875 HIV-1 protease and reverse transcriptase sequences, we analyzed the concordance of detection of ambiguous nucleotides, amino acid changes, and presence of drug resistance mutations between sequences interpreted manually by lab technicians or automatically by RECall. RECall showed excellent agreement with subjective human interpretation of HIV sequence data, with 99.7% concordance over more than one million bases compared. Similar degrees of agreement (>99.5%) were noted in previous analyses of smaller datasets from other independent laboratories [360,361]. Of the limited number of differences in base calling, the vast majority were due to partial nucleotide discordance, where one method detected a mixture and the other detected one component of the mixture. As a result, the majority of amino acid differences detected by human versus RECall were also due to partial discordances. Human and RECall review agreed on 98.1% of PI and 98.7% of NRTI/NNRTI resistance mutations identified by either method. In comparison, when identical sequence trace files are inspected and edited by multiple human operators the identification rate of resistance mutations can be <90% [335] depending on the samples tested. Although a very small number of key resistance mutations were identified by a single method only, these were all a result of partial mismatches due to differential detection of nucleotide mixtures.  Overall, 89.2% of all samples re-analyzed were successfully processed by RECall using the default parameters. The majority of sample failures were due to the strict requirement for double primer coverage across the entire sequence length. In some cases, single coverage was limited to only short 56  stretches of several bases. In resource-limited settings, this relatively high failure rate could be considered unacceptable, especially if the remaining sequence is of high quality and the areas of single coverage are limited to positions that do not inform drug resistance interpretations. In such cases, a human technician can overrule RECall’s decision and manually export the assembled sequenced for analysis. Alternatively, the RECall parameters could be reconfigured to allow single coverage at sequence positions that are not associated with drug resistance.  Despite the extremely high concordance between methods, there may be several reasons for the observed discrepancies. First, RECall relies on phred peak areas to call mixtures and is therefore unable to call mixtures of three nucleotides. The impact of this shortcoming, however, is negligible as 3-base mixtures were called exceedingly rarely by human interpretation (0.007% of bases called) and could simply represent technical artifacts [333,335]. Second, technicians, especially lesser-experienced ones, can arguably be biased during sequence interpretation: for example at a single position, visibly “cleaner” chromatograms with taller peaks may be assigned more weight, mixtures within lower quality areas of sequence may not be considered “true” mixtures, and frequently observed “patterns” of nucleotide mixtures may influence a person’s decision to call mixtures with a borderline secondary peak area. In contrast, RECall is not programmed to weigh sequence reads based on peak height; determination of mixtures is based solely on peak area, rather than any perception of shape and quality of data is strictly dependent upon phred scores. While the inflexibility of a fully automated system for sequence analysis and interpretation may appear to be a drawback, the results of this study show otherwise. RECall is configured to mark unusual sequence positions, including mixtures, which a technician could visually check. In this study, RECall was run without human intervention and still rapidly produced unbiased, consistent results on a dataset generated by different methods in an external laboratory.  RECall did call marginally fewer mixtures overall than the human operator, however, this difference was not statistically significant. Subjectively, these discordant bases could be considered “hard to call” by human operators and mixture calling frequency would be directly related to the individual’s 57  personal biases (Figure 2.2). In a related experiment, eight BCCfE lab technicians were presented with a panel of chromatograms for which the two sequence interpretation methods produced discordant results: one method calling a mixture, the other not. In general, the majority of operators preferentially called a mixture; 75% of those surveyed chose a mixed-base call over half the time. However, mixture calling frequency varied widely between technicians, ranging from 25-75% (data not shown), illustrating the extremely subjective nature of calling mixed bases. The results of our human vs. RECall analysis falls well within the range of inter-operator mixture-calling variability [335]; the small difference in mixed base frequency may be as likely due to over-calling by the technicians, than under-calling by RECall. If this discrepancy is of concern it can be easily lessened by modifying the mixture calling threshold (% uncalled base area) to more closely mimic a favored laboratory technician’s tendencies. Regardless of the mixture calling parameters chosen, RECall provides standardization of base calling frequencies – an extremely important feature of a clinical reporting tool, and one that is clearly not achievable solely with human interpretation [335]. Furthermore, Good Laboratory Practice standards call for traceability of data; if manual edits are made to data generated by automated instruments, both the change and its justification should be robustly documented [362]. In HIV drug resistance testing, the large number of manual changes required during the assembly and editing of a consensus sequence precludes this. RECall, however, provides a system for minimizing and tracking these manual edits. Most importantly, RECall significantly improves the processing efficiency of HIV drug resistance genotyping sequence data. Specifically, RECall removes the need to perform several time-consuming and potentially error-prone manual analysis tasks, including identifying and grouping chromatograms from a single sample, trimming regions of low-quality data, aligning primer sequences to a reference standard, manual review of mixed bases, and exporting a finished FASTA file. Once the RECall program is initiated (a process requiring only a few mouse clicks), automated analysis requires no subsequent human intervention. Furthermore, additional efficiency gains are achieved by fully integrating RECall into the data processing pipeline; ideally, RECall is set to run immediately as soon as chromatogram data are released from the sequencing instrument.  58  While the results presented here are limited to HIV drug resistance genotyping of protease and RT, RECall can easily be extended to analyze other regions of HIV or any protein coding regions that can be sequenced by population-based methods. At the BCCfE, RECall is currently the primary software used for drug resistance genotyping of protease-RT, integrase and gp41, as well as genotypic tropism testing of the V3 loop. RECall was used (without human review) to process sequences from several randomized clinical trials of HIV tropism and the results were found to be predictive of virological outcome [363–365]. The results of our inter-laboratory comparisons show that RECall can provide an objective, standardized protocol for HIV sequence interpretation in clinical and research laboratories. The speed and cost-effectiveness of using an automated tool for sequence analysis are the primary advantages. Standardizing sequence interpretation enables changes in laboratory procedures to be evaluated independently of the sequence interpretation steps. Furthermore, RECall enables unbiased sequence interpretation and its internal parameters provide additional quality control mechanisms, both of which ensure that only consistent, high-quality data are reported. 59    Highly Active Antiretroviral Therapy (HAART) has resulted in dramatic improvements in HIV outcomes [189,366]. A direct measure of HAART success is an undetectable HIV plasma viral load (pVL). The Roche COBAS AmpliPrep/COBAS AMPLICOR HIV-1 MONITOR UltraSensitive Test, version 1.5 ("Amplicor v1.5"; Roche Molecular Diagnostics, Laval, Quebec, Canada) [367] was introduced for clinical use in British Columbia, Canada in 1999 and has been commonly used worldwide. International treatment guidelines have recommended that the goal of therapy is to achieve pVL below 50 copies/mL, the lower limit of detection of the Amplicor v1.5 assay [267,268]. For over 10 years, an undetectable pVL by the Amplicor v1.5 assay was the de facto gold standard definition of therapy success in antiretroviral clinical trials and HIV treatment programs. However, the manufacture of the Amplicor v1.5 assay has been discontinued. The Roche COBAS AmpliPrep/COBAS TaqMan v1 HIV-1 Test ("TaqMan v1"; Roche Molecular Diagnostics, Laval, Quebec, Canada) [368] has replaced Amplicor v1.5. While the assays are highly correlated over their dynamic ranges (50-1,000,000 and 40-10,000,000 HIV RNA copies/mL for Amplicor and Taqman v1, respectively) [369], a sudden increase in the number of patients with detectable pVL has been reported upon switching to TaqMan v1 [370–374]. These unexpected detectable pVL (typically 40-250 copies/mL) occurred among patients who had consistently suppressed pVL (<50 copies/mL) by Amplicor v1.5.  Sudden unexpected HIV detectability leads to additional pVL testing, and medical visits for re-evaluation of adherence, concomitant medications, drug resistance and comorbidities, resulting in significant additional costs and stress for patients, their support network, and health care providers [375]. 60  We undertook the present study to compare the results of the previous Amplicor v1.5 and the replacement TaqMan v1 assays in cases of low-level viremia. First, we evaluate the concordance of the two assays when low-level pVL (40-250 copies/mL) are reported by TaqMan v1. Second, we evaluate short-term virological and antiretroviral resistance outcomes in patients with low-level detectable TaqMan v1 pVL receiving stable HAART. Third, we show that the TaqMan v1 assay may also systematically underestimate pVL in a minority of cases due to imperfect binding of the primers and/or probes.  The TaqMan v1 assay was introduced in British Columbia (BC), Canada in February 2008. After observing a high number of patients with newly detectable viral loads by the new assay and poor concordance of pVL results <100 copies/mL from both assays [370] a new testing protocol was implemented. In brief, between October 2009 and April 2010 samples having a TaqMan v1 pVL result ≥40 and <250 HIV RNA copies/mL were re-tested using the Roche Ultrasensitive Amplicor v1.5 assay. In these patients, only the results of the re-testing with Amplicor v1.5 were reported to the prescribing physicians, as this was the standard of care in BC prior to the introduction of the TaqMan v1 assay (Figure 3.1). HIV viral load measurements were undertaken as per the specific test kit instructions [367,368]. The TaqMan v1 assay was performed either on fresh (for locally-collected samples) or previously frozen (for samples requiring shipment) plasma aliquots. Prior to re-testing by Amplicor v1.5, samples were stored at -20°C until TaqMan v1 pVL results were available. The concordance between TaqMan v1 and Amplicor v1.5 values was assessed in samples with low-level pVL (40-250 copies/mL). The analysis presented is therefore limited to a one-way comparison between low-but-detectable TaqMan v1 pVL results and their corresponding Amplicor v1.5 results. Due to the limited availability of Amplicor v1.5 pVL test kits following its discontinuation, a two-way comparison could not be performed.  61   All plasma samples were initially tested with the Roche COBAS AmpliPrep/COBAS TaqMan v1 HIV-1 Test ("TaqMan v1”) assay. TaqMan v1 pVL results <40 or ≥250 copies/mL were reported to physicians. Samples with TaqMan v1 pVL ≥40 and <250 copies/mL were re-tested by the Roche COBAS AmpliPrep/COBAS AMPLICOR V1.5 HIV-1 MONITOR UltraSensitive Test, version 1.5 (“Amplicor”). The Amplicor v1.5 test results were reported to physicians. Subsequent short-term virological and resistance outcomes were followed in patients who started HAART at least 6 months prior and who did not change therapy regimens during follow-up (N=279 patients). HIV RNA detectability by TaqMan v1 and Amplicor v1.5 were compared at the latest follow-up timepoint up until April 2010. Antiretroviral drug resistance genotyping was performed for all follow-up samples with viral loads >250 copies/mL by TaqMan v1, as described elsewhere [376]. Re-testing samples with Amplicor v1.5 revealed a subset of samples (N=29) with systematically underestimated viral loads by TaqMan v1 compared to Amplicor v1.5. For these patients, archived samples were re-tested with the Roche TaqMan version 2 (“TaqMan v2”) [377] and/or the Abbott m2000 RealTime HIV-1 assay (Abbott Molecular Diagnostics, Wiesbaden, Germany) to evaluate 62  whether these alternative real-time PCR assays returned similar results. Where sufficient plasma was available, HIV gag was also sequenced to assess potential TaqMan v1 primer incompatibility. Ethical approval for this study was granted by the Providence Health Care/University of British Columbia Research Ethics Board (H10-01778). The requirement for individual consent was waived by the Research Ethics Board as the study involved no more than minimal risk to subjects.   Between October 2009 and April 2010 a total of 1198 samples (from 950 patients) with pVL 40-250 copies/mL by TaqMan v1 were re-tested with the Amplicor v1.5 assay. This comprised approximately 10% of all tests performed in British Columbia during this period. In BC ~97% of antiretroviral-treated patients are infected with subtype-B HIV-1 [378].  The concordance between the two viral load assays was poor at low viral load strata (Figure 3.2). The median (Q1-Q3) pVL was 79 (54-124) and <50 (<50-74) HIV RNA copies/mL by the TaqMan v1 and Amplicor v1.5 assays, respectively. This difference was statistically significant (Wilcoxon signed-rank test; p<0.0001). Visual inspection of the Bland-Altman plot (Figure 3.3) suggests a bias towards higher pVL results obtained by TaqMan v1 in this range. Of all the samples with detectable pVL by TaqMan v1, 66% were undetectable by Amplicor v1.5 (789/1198 samples). If only the 965 samples with pVL ≥50 copies/mL by TaqMan v1 were considered, only 385 (40%) had detectable pVL by Amplicor v1.5. Overall, the majority of samples tested (82%, 984/1198) fall below the line of identity in Figure 3.2, suggesting that TaqMan v1 reports pVL that are consistently higher than results obtained by Amplicor v1.5 in this range.  63   Between October 2009 and April 2010 plasma samples from British Columbia with low-level viremia (40-250 copies/mL) by the Roche TaqMan v1 assay (N=1198) were systematically re-tested by the Amplicor v1.5 assay. Poor concordance was observed between TaqMan v1 and Amplicor v1.5 results, with 82% of values falling below the line of identity (red solid line). Points above the blue dashed line are samples in which TaqMan v1 underestimated pVL by ≥0.5 log10 copies/mL relative to the Amplicor v1.5 assay.  64   Visual inspection of the plot suggests a bias towards higher pVL results obtained by TaqMan v1 at low pVL (40–250 copies/mL). When attempting to interpret these results it is important to note that 1) Amplicor v1.5 pVL values below the limit of quantification (<50 copies/mL) were coded as 49 copies/mL, and 2) the plotted results are restricted to samples with TaqMan v1 results 40–250 copies/mL. Despite the low concordance between the two assays, there was a stepwise increase in detectability by Amplicor v1.5 with increasing pVL by TaqMan v1 (Figure 3.4). Samples with TaqMan v1 values 50-99 copies/mL were detectable (≥50 copies/mL) by Amplicor v1.5 in 22% of cases (116/535 samples), while samples with TaqMan v1 pVL 200-250 copies/mL were detectable by Amplicor v1.5 in 85% of cases (72/85 samples). 65    Of 1198 samples with pVL results 40-250 copies/mL by TaqMan v1, only 34% were detectable when re-tested by Amplicor v1.5 (>50 copies/mL). When TaqMan v1 results were grouped into 50 copies/mL strata we observed a stepwise increase in detectability by Amplicor v1.5 with increasing viral load by TaqMan v1.  To determine whether low-level TaqMan v1 pVL is a predictor of emergent virologic failure, we followed a subset of patients treated with stable HAART. A total of 279 of 950 (29%) eligible patients with low-level viremia by TaqMan v1 and remaining on an unchanged regimen were followed for a median of 3.2 months (Q1-Q3: 2.0-4.2 months). A median of one (Q1-Q3: 1-2) follow-up pVL test per patient was performed. Note that during this period, the Amplicor v1.5 pVL results were used to guide therapy decisions. Baseline patient characteristics are shown in Table 3.1. At baseline, 31% of these 66  patients had a detectable pVL when re-tested by Amplicor v1.5. Overall, 59 (21%) of patients had a least one pVL ‘blip’ ≥ 50 copies/mL by Amplicor v1.5 during follow-up, compared to 126 (45%) by TaqMan v1. At the latest follow-up timepoint, 17% of patients had detectable pVL by Amplicor v1.5, whereas 38% of patients had detectable pVL by TaqMan v1. Furthermore, as baseline TaqMan v1 pVL increased, the proportion of patients with detectable pVL at follow-up by both assays also increased (Figure 3.5). Overall, 20% of patients had an increase in pVL by TaqMan v1 (median pVL increase [Q1-Q3]: 80 [36-283] HIV RNA copies/mL) during the follow-up period. This suggests that the higher pVL results obtained by TaqMan v1 relative to Amplicor v1.5 are systematic and not merely the result of increased viral ‘blips’. Overall, the median viral load of all patients in this cohort decreased over the follow-up period from 75 (Q1-Q3: 54-121) copies/mL at baseline to <40 (Q1-Q3: <40-74) copies/mL at the last follow-up timepoint. Variable N (%) or Median (Q1-Q3) Male sex 226 (81%) Age at HAART initiation (years) 40 (33 − 47) History of injection drug use 126 (45%) Baseline CD4 cell count (cells/µL) 390 (230 − 530) Baseline plasma viral load (HIV RNA copies/mL)  TaqMan v1 75 (53 – 121) Amplicor v1.5 <50 (<50 – 68) Previous therapy experience (months) 54.2 (21.1 − 94.6) Follow-up from baseline to last visit (months) 3.2 (2.0 − 4.2)   67   A subset of patients (N=279) with low-level viremia (40-250 copies/mL) by TaqMan v1 were followed longitudinally for a median of 3.2 months (Q1-Q3: 2.0-4.2 months). Patients initiated HAART at least 6 months prior, and treatment regimens remained unchanged over the course of follow-up. Samples from patients’ latest follow-up visit were re-tested with the TaqMan v1 and Amplicor v1.5 assays. Overall 17% of patients had a detectable viral load by Amplicor v1.5 at their latest follow-up visit, while 38% were detectable by TaqMan v1. When patients were grouped according to their baseline TaqMan v1 into 50 copies/mL strata we observed a stepwise increase in the proportion of patients with detectable pVL at their latest follow-up visit by both the Amplicor v1.5 (red bars) and TaqMan v1 (blue bars). Consistent with previous results, more patients had detectable pVL by TaqMan v1 than by Amplicor v1.5 at follow-up in all strata.  68  Follow-up samples collected prior to September 2010 with viral load >250 copies/mL by TaqMan v1 were also tested for antiretroviral resistance using standard genotypic methods (N=66 successfully tested of a total of 69 patients). A median of one sequence per patient (Q1-Q3: 1-2) was obtained over a median of 4.4 (Q1-Q3: 0.9-6.9) months. New resistance mutations in protease/reverse-transcriptase (PR-RT) were observed in eight patients (12%) compared to pre-therapy sequences, suggesting that, in general, low-level pVL by TaqMan v1 may not be indicative of emerging resistant variants. It is notable that only three patients with newly-detected resistance mutations had undetectable viral loads by Amplicor v1.5 at their baseline visit.  In addition to the frequent pVL overestimation reported by TaqMan v1, a small subset of samples tested (~2.4%, 29/1198) had pVL levels by TaqMan v1 >0.5 log10 lower than by Amplicor v1.5 (Figure 3.2). For these patients, archived plasma samples with low TaqMan v1 pVL (range: <40-458 copies/mL) were re-tested using Roche COBAS AmpliPrep/COBAS TaqMan HIV-1 Test, v2.0 ("TaqMan v2"; Roche Molecular Diagnostics, Laval, Quebec, Canada) (N=31) and/or the Abbott m2000 RealTime HIV-1 assay (“Abbott”; Abbott Molecular Canada, Toronto, Ontario, Canada) (N=15) if sample volume permitted. Of 18 samples with undetectable pVL by TaqMan v1, five (28%) were detectable by TaqMan v2 (range 81-382 copies/mL). In 13 samples with low but detectable pVL by TaqMan v1, TaqMan v2 results were a median of 1.1 log10 copies/mL higher (Q1-Q3: 0.87-1.34 log10 copies/mL), consistent with a systematic underestimation of the pVL levels by TaqMan v1 in a specific subset of patients (Figure 3.6 depicts the pVL history of one representative patient). This observation was not restricted to samples from BC. One sample collected in Saskatchewan with a reported “undetectable” pVL by TaqMan v1 had a confirmed a pVL level >200,000 copies/mL by other tests (data not shown). All patients with underestimated pVL were infected with HIV-1 subtype B. 69   A minority (2.4%) of samples had viral loads underestimated by >0.5 log10 copies/mL by TaqMan v1 after re-testing with Amplicor v1.5. Depicted here is the viral load history of one representative patient showing systematic underestimation of viral load by TaqMan v1 when compared to results obtained by re-testing samples with the Amplicor v1.5, TaqMan v2 and/or Abbott assays. TaqMan version 1 viral load results are shown as solid triangles (▲) joined by a solid line. Overlaid are the corresponding results from the Amplicor v1.5 (solid squares ■), TaqMan version 2 (unshaded triangles △) and Abbott (unshaded squares □) assays where available. For this patient TaqMan v1 systematically under-reported pVL by an average of 1.3 log10 copies/mL.  70  We hypothesized that the underestimation of pVL results by the TaqMan v1 assay was due to inefficient binding of the HIV gag assay primers. To explore this issue further, we sequenced HIV-1 gag in samples exhibiting low TaqMan v1, but high Amplicor v1.5 pVL. As the sequences of the TaqMan v1 assay primers are unpublished, gag sequences were sent to Roche (Laval, Quebec) for interpretation. Of the nine gag sequences shared with Roche, seven (78%) apparently had mutations potentially incompatible with the TaqMan v1 primers: two (22%) and five (56%) showed incompatibilities with the “upstream” and “downstream” primers, respectively. Since the TaqMan v1 primer and probe sequences remain proprietary, we cannot report the specific mutations involved. New resistance mutations were detected in one patient sample out of seven with longitudinal PR-RT sequences available. In this sample, viral load was underestimated by 2.5 log10 copies/mL as compared to TaqMan v2.   The introduction of a new HIV viral load assay (“TaqMan v1”) in February 2008 resulted in an unexpected increase in the prevalence of detectable HIV loads in BC, Canada [370]. When samples with low but detectable pVL (40-250 copies/mL by TaqMan v1) were re-tested using the Amplicor v1.5 assay, we found a poor concordance between the values, with close to two-thirds of samples giving undetectable pVL by Amplicor v1.5. It is highly unlikely that this sudden change in pVL detectability is due to differences in assay performance in non-B HIV subtypes, as the HIV-1 epidemic in BC consists primarily of subtype B infections [378]. Our data clearly demonstrate that a pVL <50 copies/mL by Amplicor v1.5 is not equivalent to <50 copies/mL by TaqMan v1. Therefore, an unintended consequence of replacing the Amplicor v1.5 assay with the TaqMan v1 assay was that it altered the nearly universally accepted definition of virological failure – approximately 6% of all pVL tests performed in BC between October 2009 and April 2010 had false-positive results using the previous criteria. 71  The clinical consequences of low-level viremia were unclear even before the TaqMan v1 assay was instituted. Some studies have indicated that intermittent low-level viremia is associated with a higher risk of virological failure or drug resistance [379,380] and have found associations between the risk of failure and the magnitude of viral ‘blips’ [381,382]. However, other studies have failed to find such links [383]. To evaluate the management of patients on HAART monitored with the TaqMan v1 assay, we prospectively followed a subset of these patients for a period of approximately 3 months while on stable therapy. Physicians were informed of only Amplicor v1.5 pVL values, so patients who had undetectable pVL by Amplicor v1.5 but detectable pVL by TaqMan v1 did not receive any additional adherence counseling or pVL monitoring. A full 83% of these patients had undetectable pVL by the Amplicor v1.5 assay at follow-up, and had little evidence of drug resistance evolution. The median TaqMan v1 pVL in this group decreased over the follow-up period with 62% having a pVL <40 copies/mL at their latest follow-up visit. Unfortunately, the limited availability of Roche Amplicor v1.5 assay kits following their discontinuation restricted parallel testing by TaqMan v1 and Amplicor V1.5 to a relatively narrow range of pVL (40-250 copies/ml) and a short follow-up time.  New resistance mutations were observed in only 12% of tested patients with pVL >250 copies/mL by TaqMan v1 over the course of follow-up. However, it should be noted that uncertainty and stochastic variation in testing samples with low viremia could result in resistance mutations being missed or detected by chance in some cases [332]. Nevertheless, these results suggest that low but detectable pVL <250 copies/mL by TaqMan v1 do not correlate with detectability by Amplicor v1.5, and are not indicative of impending short-term virological failure or drug resistance. Rather, we estimate that on average a pVL value of 150 copies/mL by TaqMan v1 is approximately equivalent to a pVL value of 50 copies/mL by Amplicor v1.5. However, it is important to note that the reverse may not be true - an Amplicor v1.5 pVL value of 50 copies/mL may not be equivalent to a TaqMan v1 value of 150 copies/mL as this scenario was not evaluated. While this is not sufficient to entirely repudiate the validity of the TaqMan v1 assay, there exists no positive evidence to support that TaqMan v1 pVL levels of 40-250 copies/mL are clinically relevant. Given the change in assays, the definition of “virological failure” should be re-validated against clinically meaningful parameters; the BC HIV treatment guidelines 72  have revised the definition of treatment failure as two consecutive pVL measurements >250 copies/mL by TaqMan [384]. Similarly, the treatment guidelines from the US Department of Health and Human Services now define virological failure as a pVL rebound to >200 copies/mL following previous suppression [192]. In contrast, other studies have suggested that detectable pVL below the limit of quantification may be predictive of viral rebound, and therefore the definition of virological failure be adjusted downwards [385]. The definition of virological failure in the treatment guidelines of the European AIDS Clinical Society, and the International AIDS Society-USA remains unchanged. This study also confirmed an unrelated but potentially more serious shortcoming of the TaqMan v1 assay. The TaqMan v1 primers were unable to efficiently amplify their targets in a subset of ~2.4% of samples tested leading to the systematic underestimation of viral loads by up to 2.5 log10 copies/mL as confirmed by re-testing by Amplicor v1.5, TaqMan v2, and/or the Abbott RealTime assays (see also [371,377,386]). When HIV-1 gag sequences from these samples were sent to Roche for analysis, 78% were identified as having mutations incompatible with the TaqMan v1 primers. Of concern, as the TaqMan v1 primer and probe sequences remain unpublished, there is no mechanism available to predict or to retrospectively identify which patients belong to this subset. However, if the locations and sequences of the primers were made available, HIV-1 gag could be sequenced for primer incompatibilities in patients suspected of having systematically underestimated pVL. The consequences of incorrectly diagnosing an undetectable pVL in a pregnant woman or the HIV-positive partner in a serodiscordant couple, for example, could be disastrous. Unfortunately, the study design did not permit the accurate quantification of the proportion of “false-negative” results (i.e. an “undetectable” pVL by TaqMan v1, but pVL >50 copies/mL by Amplicor v1.5). Due to the limited availability of Amplicor v1.5 kits in Canada following its discontinuation only a one-way analysis could be performed.  The discrepancies between Amplicor v1.5 and TaqMan v1 pVL at low-level pVL may have arisen due to the methods used to validate the TaqMan v1 assay. Validation was performed by testing TaqMan v1 results against the Amplicor v1.5 assay over a very wide range of pVL. No special attention was placed 73  on pVL in the clinically-relevant range (<1000 copies/mL), where the test results are directly used to guide therapeutic decisions [267,368]. Validating TaqMan v1 with pVL nearing 107 copies/mL may have obfuscated clinically meaningful differences between the two assays [387]. Therefore, we recommend that in the future new or “upgraded” pVL assays be validated with a particular focus on the clinically-relevant range before commercial approval.  An updated version of the TaqMan v1 assay (version 2.0) is now available. While the assay upgrade appears to have corrected the systematic underestimation of pVL occasionally seen in TaqMan v1 [388], version 2.0 may still show higher pVL quantitation than Amplicor v1.5 in the lower range of detection; nearly 12% of samples with an undetectable pVL by Amplicor v1.5 still have a pVL result >50 copies/mL by TaqMan v2 [374]. Other higher-sensitivity HIV pVL assays with decreased lower limits of quantification, such as the Abbott m2000 RealTime assay, are in routine clinical use elsewhere. Although these assays have not been evaluated here, given further investigation they may represent alternatives to the Roche TaqMan system. However, regardless of the pVL assay used, the uncertainty surrounding the clinical relevance of low-level viremia requires that a set of technical and regulatory standards to evaluate the clinical utility of new diagnostic assays be rigorously defined. 74    Human Leukocyte Antigen class I (HLA)-restricted Cytotoxic T-lymphocytes (CTL) and Highly Active Antiretroviral Therapy (HAART) shape HIV-1 evolution by selecting mutations that compromise immune [78,132] and therapy-induced [389,390] virologic control. The extent and rate at which immune pressures drive HIV-1 evolution during HAART remains incompletely characterized.  HAART suppresses HIV replication. As a consequence of antigen withdrawal, HIV-specific CTL responses decline in frequency [391] (although they remain present in the CTL population) [392]. Ongoing immune escape occurs in specific HLA-restricted epitopes during HAART [393–395]. However, no cohort studies have systematically investigated the incidence and determinants of HLA-associated viral evolution in large populations starting HAART, nor investigated whether HAART reverses immune escape.  Systematic classification of HIV substitutions as immune escape, reversion or other events is possible with the availability of HIV proteome-wide reference lists of HLA-associated polymorphisms [79,396–401]. Here, we characterize HLA-associated evolution in protease and reverse transcriptase (RT) before and after HAART initiation in a large population-based cohort [190,274,397] as a function of HAART efficacy. 75    Studied patients (N=619) represent a subset (52.1%) of the British Columbia HAART Observational, Medical Evaluation and Research (HOMER) cohort, which comprises all clinically-characterized antiretroviral-naïve adults initiating HAART in BC between August 1, 1996 and September 30, 1999 (N=1188) [190,274,397]. Of these, 201 were excluded based on incomplete HLA data and a further 258 for lack of baseline (pre-therapy) protease/RT sequence data. A further 110 were excluded for lacking follow-up sequences, yielding the final N=619.  At HAART initiation, the median pVL was 5.1 (Q1-Q3: 4.7-5.5) log10 copies/mL; the median CD4 count was 290 (Q1-Q3: 140-420) cells/μL (Table 4.1). Included subjects had significantly higher pVL, higher prevalence of injection drug use, and lower adherence in the first year of HAART compared to those excluded (Table 4.1); factors that reflect the likelihood that post-HAART samples would have pVL sufficient for HIV genotyping, thus enhancing the likelihood of patient inclusion. Post-HAART samples had significantly lower pVL than baseline (median 4.34 [Q1-Q3: 3.45-5.00] vs. 5.11 [Q1-Q3: 4.72-5.51] log10 copies/ml; p<0.0001). Patients were followed for a median 5.2 (Q1-Q3:2.2-9.0) years following HAART initiation. This study was approved by the Providence Health Care/University of British Columbia Research Ethics Board. 76  Variable Whole Cohort (N=1188) Included Subset (N=619) Excluded Subset (N=569) P value pVL (log10 copies/mL) 5.08 (4.62-5.49) 5.11 (4.73-5.51) 5.00 (4.52-5.46) 0.002 CD4 cell count (cells/μL) 280 (130-420) 290 (140-420) 260 (100-420) 0.06 Age (years) 37.1 (31.9-43.5) 37.0 (32.2-43.5) 37.1 (31.8-43.5) 0.94 NNRTI in first HAART (yes vs. no) 306 (25.8%) 165 (26.7%) 141 (24.8%) 0.47 Sex (female vs. male) 185 (15.6%) 89 (14.4%) 96 (16.9%) 0.26 History of injection drug use (yes vs. no) 351 (29.5%) 207 (33.4%) 144 (25.3%) 0.002 AIDS diagnosis prior to HAART initiation (yes vs. no) 157 (13.2%) 71 (11.5%) 86 (15.1%) 0.07 First year prescription refill adherence (%)  93 (55-100) 90 (50-100) 100 (59-100) 0.02  pVL, plasma viral load; NNRTI, non-nucleoside reverse transcriptase inhibitor; PI, protease inhibitor. For continuous variables (pVL, CD4 count, Age, Adherence) median values are presented with Q1-Q3 ranges in parentheses. Binary variables are presented as count (percentage). P values were calculated by Mann-Whitney U test (continuous variables) or Chi-squared test (binary variables). 77   HIV Genotyping was attempted on all samples with plasma viral loads (pVL) ≥1000 copies/mL as previously described [274]. Protease and the first 240 codons of RT were amplified from extracted plasma HIV RNA by nested RT–PCR using gene-specific primers, and sequenced bidirectionally on an ABI automated DNA sequencer (Applied Biosystems, Foster, California, USA). Chromatograms were analyzed using Sequencher (Genecodes, http://www.genecodes.com) or the custom program RECall [402]. Sequences were aligned to HIV-1 subtype B reference HXB2 (GenBank accession K03455) using a modified NAP algorithm [403] and nucleotide mixtures were called if the secondary peak exceeded 20% of the predominant peak area. A median of 5 (Q1-Q3: 2-8) post-HAART samples were sequenced per patient. Of these, 61% (median 2 [Q1-Q3: 1-5] sequences/patient) were collected during a time period covered by a HAART prescription. HLA Class I sequence-based typing was performed as described elsewhere [397,398].  Previous studies of immune-driven HIV evolution lacked standardized definitions of immune escape. We therefore defined escape (and reversion) according to a published list of HLA-associated HIV polymorphisms identified in a cohort of N~1500 chronically infected individuals which included the present study participants at the baseline timepoint [398]. This list was derived using phylogenetically-informed methods incorporating corrections for evolutionary relatedness of HIV sequences, HIV codon covariation and HLA linkage disequilibrium [398,404], and featured a q-value correction for multiple tests [405]. A threshold of q<0.2 (corresponding to a 20% false-discovery rate) was used in the present analysis, yielding 120 unique HLA-associated polymorphisms occurring at ~19% (63/339) of codons in the protease/RT region studied (Appendix I). These are dichotomized into “adapted” and “non-adapted” forms (those most likely to emerge in the presence or absence of selection pressure by the relevant HLA allele, respectively) [398], and must always be interpreted in context of their restricting allele. For example, RT codon 135 features “I” and “T” as B*51-associated 78  “non-adapted” and “adapted” forms respectively, “T” as the “adapted” form associated with A*02 (with no specific “non-adapted” form identified for this allele), and numerous other polymorphisms restricted by other alleles (Appendix I). Therefore, in B*51-expressing persons, “escape” at codon 135 is defined as evolution away from isoleucine (either to threonine or to another amino acid) while in A*02-expressing persons, “escape” at codon 135 is defined as selection of threonine only. Patients without HLA alleles associated with escape at codon 135 were not included in analyses at that codon. Mutations in protease/RT associated with drug resistance were defined as the key resistance mutations from the IAS-USA list [358]. There were four cases where immune and drug pressures acted on the same codon (protease 82, RT 103, 106, 184), however there was no overlap in their respective mutational patterns. Therefore, changes at these positions were not counted twice as both drug resistance and immune escape mutations. Since bulk sequencing was employed, antiretroviral and immune-associated polymorphisms were classified as “present” if they were observed alone or as part of an amino acid mixture detected at a >20% threshold.   The baseline protease-RT timepoint was defined as the sequence collected on the closest date prior to HAART initiation (median 38 [Q1-Q3: 23-64] days prior). Cox proportional hazards regression was used to investigate the relationship between baseline sociodemographic and genetic parameters, as well as time-dependent clinical parameters on the risk of developing immune escape or drug resistance mutations (defined as the years elapsed between HAART initiation and the first detection of a full or partial amino acid change consistent with the predefined list of HLA-associated polymorphisms [398] or drug resistance mutations [358], respectively). Individuals not reaching an endpoint were censored at the date of their last sequenced sample. Binary variables investigated were history of self and/or physician reported injection drug use [IDU] (yes [comparison group] vs. no [reference group]), pre-therapy AIDS diagnosis (yes vs. no), initial regimen type (non-nucleoside reverse transcriptase inhibitor [NNRTI] vs. protease inhibitor [PI]-79  containing) and prescription-refill adherence in the first year of therapy (≥95% vs. <95%). Baseline age was modeled as a continuous variable (per 1 year increment). Plasma viral load (pVL; copies/mL, per log10 increment) and CD4 count (cells/μL, 100 cell increment) were modeled as time-dependent variables, meaning that the model accounted for fluctuations in these covariates over the follow-up period. Patients had a median 19 (Q1-Q3: 8-32) viral load measurements and 20 (Q1-Q3: 8-33) CD4 measurements. Baseline and time-dependent variables were incorporated as individual predictors in the Cox proportional hazards model using the R package ‘survival’. Multivariate models were constructed using a stepwise Akaike Information Criterion (AIC) procedure. Briefly, an optimized model is identified by creating multiple models, initially including all variables. At each step, the variable with the highest p-value is eliminated and an AIC value calculated for the new model, repeating until no variables remain. The final model is that with the lowest overall AIC. Hazards were proportional in the final model with the exception of adherence. P-values <0.05 were defined as statistically significant.    HLA-associated immune escape [398] and drug resistance [358] mutations were defined according to published lists. As expected, baseline HLA escape prevalence was relatively high in this chronically infected cohort (median 6 [Q1-Q3: 4-8] escape mutations/patient). Some codons approached saturation of escape in individuals expressing the relevant HLA, including specific polymorphisms at protease codon 35 (75% escape prevalence in persons expressing B*44 and/or C*18), 36 (99% in C*07-expressing persons), 93 (74% in B*15-expressing persons) and RT codon 203 (99% in A*02-expressing persons). RT codon 135 [406] exhibited 59% overall escape prevalence among individuals expressing the relevant HLA alleles, including 92% of N=80 B*51-expressing persons (Figure 4.1).   80   Pre-therapy prevalence of mutations associated with immune escape at a given codon in protease (panel A) and reverse transcriptase (panel B) are indicated by blue bars. The cumulative prevalence of immune escape each codon observed at any point during follow-up are indicated by red bars. Note that the immune escape analysis was restricted to patients expressing the relevant HLA alleles associated with escape at that particular codon. Numbers above bars represent the number of eligible patients at each codon. Only codons with 20 eligible patients are shown.    81  Over the median 5.2 year follow-up, 269 (43%) patients developed ≥1 new HLA-associated polymorphisms (median 0 [Q1-Q3: 0-1]/patient). New escape events occurred in a median of 4.5% (Q1-Q3: 0.4-8.8%) of patients expressing the relevant HLA at each studied codon. B*35-associated codon 177 ranked among the most frequently escaping residues during HAART (Figure 4.1), with 20% of B*35-expressing patients developing a new escape mutation at this position. Among patients developing new escape mutations, the median time to their detection was 2.1 (Q1-Q3: 0.7-5.0) years. To provide context, corresponding rates of escape during untreated chronic infection were estimated using available pre-therapy sequence data from a subset of 210 patients (median 1 [range 1-4] additional pre-therapy samples/patient spanning a median 0.7 [Q1-Q3: 0.3-1.4] year period). A total of 55 immune escape events (median 0 [range 0-3]/patient) were observed, yielding a mean estimated escape rate of 0.84 (standard deviation [SD] 5.29) substitutions per person per year during untreated chronic infection. After HAART initiation, the mean estimated escape rate in these same patients decreased approximately sevenfold to 0.13 (SD 0.43; p=0.02), or fourfold to 0.22 (SD 1.18; p=0.03) substitutions per person per year, based on total time elapsed from treatment initiation or when restricted to periods of detectable viremia, respectively. Baseline drug resistance prevalence was low (median 0; range 0-6/patient). The most common polymorphism was M41L in RT (Figure 4.2). During HAART, 278 patients developed ≥1 new resistance mutations (median 0 [Q1-Q3: 0-2] per patient). Among the most frequent were M184I/V, K103N, and Y181C in RT, observed in 31%, 16%, and 13% of patients respectively (Figure 4.2). The median time to detection of the first new resistance mutation was 1.5 (Q1-Q3: 0.6-3.6) years following HAART initiation, slightly more rapid than the occurrence of immune escape (Wilcoxon rank sum test p=0.05). Escape and resistance events were significantly associated (Chi-squared test p=0.0004), indicating that persons at risk of resistance were similarly at elevated risk of immune escape.  82   Pre-therapy prevalence of key drug resistance mutations at a given codon in protease (panel A) and reverse transcriptase (panel B) are indicated by blue bars. The cumulative prevalence of drug resistance mutations at each codon observed at any point during follow-up are indicated by red bars.   83   We wished to identify the parameters most strongly associated with escape at HLA associated positions after HAART initiation. In univariate Cox proportional hazards regression analyses, risk of immune escape was significantly positively associated with time dependent pVL and female sex, and significantly negatively correlated with time-dependent CD4 count and ≥95% prescription refill adherence (Table 4.2). In multivariate analyses adjusting for demographic, clinical and therapeutic variables, both time dependent pVL (Hazard Ratio [HR] 1.9 per log10 increment; 95% Confidence Interval [CI] 1.7-2.1) and female sex (HR 1.4; CI 1.0-1.9) remained significant, as did NNRTI-containing initial HAART (HR 1.5; CI 1.1-1.9) (Table 4.2). To provide context, risk of drug resistance was also significantly positively associated with time-dependent pVL (HR 1.4 per log10 increment; CI 1.3-1.6), NNRTI-containing initial HAART (HR 1.4; CI 1.1-1.8), injection drug use (HR 1.3; CI 1.0-1.7), and ≥95% prescription refill record (HR 1.5; CI 1.1-2.0), and significantly negatively associated with time-dependent CD4 count (HR 0.9 per 100 CD4 cell/μL increment; CI 0.85-0.95) in multivariate analyses (Table 4.2). Results were similar when restricting the analysis to samples collected prior to the first treatment interruption (not shown). Immune-driven viral evolution during therapy therefore occurs at a rate inversely related to the virologic success of HAART: the more effective the viral suppression, the less likely that immune escape will continue to occur. 84   Drug Resistance HLA Escape Univariate  Analysis Multivariate Analysis Univariate Analysis Multivariate Analysis Variable Hazard Ratio (95% CI) P value Hazard Ratio (95% CI) P value Hazard Ratio (95% CI) P value Hazard Ratio (95% CI) P value pVL per log10 increment* 1.4 (1.3-1.6) <0.0001 1.4 (1.3-1.6) <0.0001 1.9 (1.7-2.1) <0.0001 1.9 (1.7-2.1) <0.0001 CD4 cell count per 100 CD4 cell/μL increment* 0.85 (0.81-0.90) <0.0001 0.90 (0.85-0.95) 0.0001 0.87 (0.82-0.91) <0.0001   Age (per year increment) 1.0 (0.99-1.0) 0.25   0.99 (0.98-1.0) 0.40   NNRTI in first HAART (yes vs. no) 1.2 (0.91-1.5) 0.22 1.4 (1.1-1.8) 0.01 1.3 (0.97-1.7) 0.09 1.5 (1.1-1.9) 0.005 Sex (female vs. male) 0.99 (0.71-1.4) 0.95 …  1.5 (1.1-2.1) 0.007 1.4 (1.0-1.9) 0.03 History of injection drug use (yes vs. no) 1.4 (1.1-1.8) 0.007 1.3 (1.0-1.7) 0.02 1.1 (0.87-1.4) 0.39   AIDS diagnosis prior to HAART initiation (yes vs. no) 1.3 (0.88-1.9) 0.20   0.89 (0.58-1.4) 0.58   First year prescription refill adherence (≥95% vs. <95%) 0.81 (0.64-1.0) 0.080 1.5 (1.1-2.0) 0.003 0.56 (0.44-0.72) <0.0001    CI, confidence interval; pVL, plasma viral load (log10 copies/mL); NNRTI, non-nucleoside reverse transcriptase inhibitor; PI, protease inhibitor; *time dependent variable  85  To further quantify the effect of HAART on HIV evolution, a time-weighted average pVL and CD4 was calculated for each patient between baseline and the first escape or resistance event (or the censoring date). Subjects were then dichotomized into those exhibiting pVL/CD4 averages above/below the population medians, and the relationship between these events and average pVL and CD4 assessed using Kaplan-Meier analyses. Exhibiting an average pVL below the population median was associated with significantly decreased rate of immune escape (p<0.0001) and drug resistance (p=0.01) (Figure 4.3A, B). Half of the high pVL group remained at risk of escape after ~4 years, compared to >10 years in the low pVL group. Exhibiting an average CD4 above the population median was significantly associated with a decreased rate of drug resistance (p<0.0001); a similar trend for immune escape did not reach statistical significance (p=0.22) (Figure 4.3C, D).  86   Kaplan-Meier curves showing the relationship between time-weighted average pVL on detection of immune escape and drug resistance (panels A and B), as well as the relationship between time-weighted average CD4 count on detection of immune escape and drug resistance (panels C and D). In panels A and B, blue and red curves differentiate patients with weighted average pVL < or ≥ the population median, respectively. In panels C and D, green and orange curves differentiate patients with weighted average CD4 > or ≤ the population median, respectively. Dotted lines represent 95% confidence intervals. The number of patients remaining at risk are shown at baseline, 5 years and 10 years.  87   Selection of drug resistance often precedes viral breakthrough during HAART, but the relationship between CTL escape and viral breakthrough is less well characterized. We therefore compared the last sequences prior to pVL suppression (median 211 [Q1-Q3: 118-525] days prior to suppression) to the first sequences following pVL breakthrough in a subset of N=214 patients who experienced viral suppression for median period of 1.0 (Q1-Q3: 0.5-2.4) years. No overall net change in the number of HLA-associated escape mutations was observed during this period (p=0.85), however a trend towards increasing number of drug resistance mutations was observed between pre- (median 0 [Q1-Q3: 0-0]) and post- (median 0 [Q1-Q3: 0-1]) suppression time-points (Wilcoxon rank sum test p=0.09) (Figure 4.4). Together, these data support selection of drug resistance, but not immune escape, as early sequelae of pVL breakthrough. Note, however, that only a minority of all immune escape (58/269 [22%]) and drug resistance (69/278 [25%]) events were observed in the first sequence following a period of suppression (median 165 [Q1-Q3: 73-381] and 162 [Q1-Q3: 98-260] days following the last undetectable pVL measurement, respectively).   The number and distribution of immune escape (panel A) and drug resistance mutations (panel B) are shown prior to, and following, a period of virologic suppression for 214 individuals. The median, interquartile range and 1.5-times the interquartile range are depicted by the thick bar, box, and whiskers, respectively. Points represent outliers. P values were calculated using Wilcoxon rank sum test.  88   Immune and antiviral selection pressures converging on the same HIV protein could yield synergistic (or antagonistic) effects [407]. If so, differential immune escape could be observed in protease based on whether the patient’s regimen contained a PI or not. However, no significant difference in relative risk of immune escape in protease was observed in patients receiving initial PI-based (N=427) compared with NNRTI-based HAART (N=158) (HR 1.3; CI 0.84-2.0), suggesting that effects of HAART on immune escape are not associated with the specific viral protein being targeted by therapy.   “Relaxation” of CTL pressure as a result of immune deficiency may facilitate reversion of certain escape mutations in vivo [408,409], raising the possibility that a similar phenomenon could occur under incompletely suppressive HAART. A previous study reported replacement of escape variants with wild-type after 14 weeks of HAART, followed by selection of novel polymorphisms [393]. We therefore investigated seven common HLA-associated polymorphisms with high baseline prevalence: codons 14, 35 and 93 in protease (in B*51, B*44 and B*15-expressing persons, respectively) and codons 123, 135, 162 and 177 in RT (in B*35, B*51, B*07 and B*35-expressing persons, respectively). “Reversion”, defined as harboring the escape variant at baseline but returning to the “non-adapted” form during HAART with no further detection of an escape variant thereafter, was rare in our study. In total, only 21 cases of reversion were observed at seven codons in protease and RT (Table 4.3). Similar reversion frequencies have been observed in acute/early HIV infection for transmitted escape mutations at these codons [410].  89  Protein Codon HLA Individuals with HLA-Associated Mutations Individuals Reverting (%) Protease 14 B*51 26 5 (19.2%) Protease 35 B*44 101 3 (3.0%) Protease 93 B*15 81 1 (1.2%) RT 123 B*35 56 6 (10.7%) RT 135 B*51 74 1 (1.4%) RT 162 B*07 82 2 (2.4%) RT 177 B*35 43 3 (7.0%)    We quantified rates of immune-mediated evolution in HIV-1 Pol after starting HAART in a large, initially antiretroviral-naïve cohort over a median 5.2 year period. Despite relatively high baseline escape prevalence, immune-driven HIV evolution continued at 78% of the known HLA-associated sites after starting HAART, at an approximately four- to seven-fold decreased rate compared to untreated chronic infection. This rate was slightly slower than that of the selection of drug resistance mutations during therapy. Although immune escape and drug resistance were significantly associated, important differences were noted: drug resistance, but not immune escape, likely drives pVL breakthrough to a greater extent during HAART. The rate of immune escape during HAART was inversely proportional to the virologic (but not immunologic) success of therapy. Importantly, although effective HAART reduced the rate of immune escape, it was not capable of reversing HLA-associated evolution.  The observation that starting HAART significantly reduces the incidence of immune escape has potential implications for timing HAART initiation. Given that the majority of immune escape mutations occur in acute/early infection and that escape can lead to profound loss of viral control [78,131,132], earlier initiation of HAART could help preserve efficacy of CTL responses for as long as possible. More importantly, it appears that once these epitopes are lost to mutational escape, they are 90  lost forever. However, the actual clinical relevance of preserving these CTL responses is not completely understood. Some limitations merit mention. First, the reference list of immune escape mutations was derived via statistical analysis of population-level HLA/HIV datasets which included the current cohort [398], possibly leading to overestimates in baseline immune escape rates. Rare or patient-specific substitutions are not captured in such population-level analyses [398]; moreover, not all these polymorphisms have been experimentally validated as CTL escape mutations. Secondly, the fact that HIV genotyping can only be performed on samples with detectable pVL may result in inaccurate estimates of rates of HIV evolution. Nevertheless, the observation that only a minority of new immune escape and drug resistance events occurred following a period of virologic suppression supports the reported estimates. Thirdly, inability to reliably detect minority species below ~25% prevalence is a recognized limitation of bulk genotyping methods [411,412]; however, failure to detect minority species would presumably affect both pre- and post-HAART samples equally. Higher sensitivity methods would be required to study minority populations in greater detail, however application of such methods to population-based studies is currently precluded by technological and cost limitations. Furthermore, in our observational cohort study, rates and types of drug resistance mutations would have been affected by HAART regimen changes that have not been captured in our intent-to-treat analysis [274]. Finally, though we have attempted to control for various sociodemographic and therapeutic variables (including female gender, initial HAART regimen, history of IDU and adherence), these factors are tightly linked in the study population and therefore should not be interpreted as general risk factors for immune escape during HAART.  In summary, HAART is unable to reverse the process of immune escape, but decreases its rate of occurrence to an extent inversely related to the virologic success of therapy. Minimizing HIV immune escape could represent a secondary benefit of effective HAART. 91    Replication fitness defines the ability of a virus to replicate under the selective pressures present in its environment. For HIV-1, replication fitness impacts the viral variants that predominate in the quasi-species and therefore influences treatment response and disease progression [413]. The replication fitness of drug-resistant HIV-1 variants is typically assessed by measuring the growth kinetics (in the absence or presence of drug) of two viral variants that are mixed at defined ratios and grown in competition in a single culture. Alternatively, individual variants can be cultured separately and their growth kinetics compared to that of a wild-type reference virus [413]. A major limitation of these approaches is that they are not conducive to high-throughput analysis and thus considerable effort is required to evaluate the fitness of multiple drug-resistant HIV-1 variants. The development of high-throughput screening methods would facilitate studies of replicative fitness of drug-resistant HIV. In particular, the replicative fitness of variants associated with resistance to second-generation non-nucleoside reverse transcriptase (RT) inhibitors (NNRTI) remains incompletely characterized. Etravirine (ETV) is second-generation NNRTI with activity against several HIV variants containing mutations that confer resistance to the first generation inhibitors nevirapine (NVP) and efavirenz [414]. However, the DUET-1 and DUET-2 clinical trials identified 17 mutations associated with ETV resistance: V90I, A98G, L100I, K101E, K101H, K101P, V106I, E138A, V179D, V179F, V179T, Y181C, Y181I, Y181V, G190A, G190S and M230L [415,416]. A weighted genotypic score that optimizes resistance interpretation has been assigned to each of these mutations and is used to guide treatment involving ETV usage in treatment-experienced HIV-infected individuals [417]. However, very little is known about the in vitro fitness profiles of HIV-1 variants containing these mutations grown in the presence of ETV.  92  Complera® is a fixed dose antiviral drug combination used as a first-line antiretroviral therapy regimen for the treatment of HIV-1 infection. Complera® is composed of the nucleoside reverse transcriptase (RT) inhibitors (NRTIs) emtricitabine (FTC) and tenofovir disoproxil fumarate, and the second-generation NNRTI rilpivirine (RPV). In the phase III ECHO and THRIVE clinical trials, the most frequent mutation combination that emerged in the RT gene of Complera® virologic failures was E138K and M184I [418,419]. In this regard, Hu and Kuritzkes [420], and Xu et al [421] reported that the E138K mutation compensated for the poor replicative capacity of M184I. However, Kulkarni et al found that HIV-1 containing E138K and M184I was less fit than viruses containing either E138K or M184I [422]. In addition to E138K, the L100I, K101E/P, E138A/G/Q, Y181C/I/V, Y188L, G190A/S/E, and M230L substitutions are associated with RPV resistance. Here, we describe an efficient method to characterize the relative fitness of multiple resistance mutants using simultaneous co-culture and longitudinal next-generation “deep” sequencing. Using these methods, we characterize the replicative fitness of multiple variants associated with resistance to second generation NNRTIs.   We generated 15 infectious viruses containing the single NNRTI resistance mutations V90I, K101P, K103N, V108I, E138A, E138K, V179D, Y181C, Y181I, Y181V, Y188C, G190A, G190S, M230L and P236L by site-directed mutagenesis of a HIV-1LAI molecular clone [423]. A wild-type (WT) virus and the 15 mutant viruses were then normalized for relative infectivity (infectious units/ng p24) in TZM-bl cells and pooled. The pooled virus was used to infect 1.6 × 106 HUT-78 cells at a multiplicity of infection of 0.004. After a 2 hour incubation period, cell free virus was removed and ETV (20nM, 200nM, or 93  500nM), NVP (5μM) or DMSO (no drug control) was added to the cultures. Culture supernatants were collected at 3 day intervals up to day 26, or until the observed peak of p24 production (Figure 5.1).   Concentrations of p24 in culture supernatant from cells infected with pooled virus and exposed to different concentrations of etravirine (ETV) or nevirapine (NVP), or no drug. Data are from a single determination. Viral RNA was extracted and deep sequencing (454 GS Junior) of an amplicon spanning RT codons 96-194 was used to quantify the distribution of mutants at each time point. A median of 2971 (Q1-Q3: 2726-3128) sequence reads per sample were collected. Sequencing errors (indels) in homopolymer-rich regions (e.g. RT codon 103) were corrected after alignment by deleting extraneous bases or by inserting the HIV-1LAI reference base where necessary. A mean PCR/deep-sequencing error rate of 0.24 ± 0.18% per amino acid was determined by calculating the amino acid substitution rate at unmodified codons in viruses grown in the absence of drug (N=3). The mean PCR/sequencing error rate at the nucleotide level was calculated to be 0.16 ± 0.28%. 94  The initial infectious pool of virus was found to contain similar concentrations of all mutant viruses as assessed by 454 deep sequencing (median: 6.7%; Q1-Q3: 5.5-8.4%; all N≥3). The effective concentration of ETV required to inhibit 50% of the pooled virus (i.e. EC50) in TZM-bl cells was found to be 62.6 ± 3.4 nM. This value was ~11-fold higher than the EC50 value (5.7 ± 1.3 nM) determined for the WT virus. In contrast, the pooled virus showed >50-fold resistance to NVP (EC50 ≥ 20 μM) compared to the WT virus (280 ± 33 nM).   Longitudinal deep sequencing revealed that in the absence of drug, the frequencies of most NNRTI resistant variants either remained the same or decreased as a function of time (Figure 5.2A). As expected, there was also a concomitant increase in the frequency of the wild-type (WT) sequence at most resistance codons (Figure 5.2B). By contrast, the Y181V virus emerged as the most common variant by days 3, 6, and 9 in cultures grown in 20 nM, 200 nM and 500 nM ETV, respectively (Figure 5.2C, D, E). The frequency of the Y181I virus also increased up to day 6 in the 20 nM ETV culture, but declined thereafter (Figure 5.2C). Transient increases in viruses containing other mutations were also observed at day 3 in the cultures grown in 200 nM ETV (K103N, E138A, Y188C and G190S) and 500 nM ETV (K101P, K103N, E138A, V179D, Y188C, and G190A/S), before Y181V-containing viruses outgrew to dominate the viral pool (Figure 5.2D, E). Importantly, 3 independent fitness experiments revealed that the spectrum of mutant viruses that grew out in cultures containing 200 nM ETV did not vary; Y181V mutants made up 90% of the sequenced reads at day 6 in all 3 experiments. In contrast to ETV, K101P and Y181C were the most common variants in the NVP containing cultures at day 6 (Figure 5.2F). However, the Y181I and Y181V variants were also selected. The deep sequencing results were confirmed by population-based sequencing (Table 5.1). Of note, there was no evidence of virus recombination in any of the experiments performed during culture, PCR or sequencing. By fitting a deterministic haploid selection model of allele frequency evolution to the observed variant frequencies 95  over time, Y181V was estimated to have a 1.8, 2.0, and 1.5-fold selective advantage over all other variants when cultured in the presence of 20nM, 200nM, and 500nM ETV, respectively.  96   The relative prevalence of wildtype viruses (Panel A) and 12 viruses each containing a single resistance mutation (Panels B-F) as determined by longitudinal deep sequencing of viral RNA from culture supernatant from cells infected with pooled virus and exposed to A+B) no drug (control), C) 20 nM etravirine (ETV), D) 200 nM ETV, E) 500 nM ETV or F) 5 μM nevirapine (NVP). Data are from a single determination.  97    Amino Acid(s) at Codon NNRTI Day 90 101 103 108 138 179 181 188 190 230 236 None 0 V K K V E V Y Y G M P 3 V K K V E V Y Y G M P 6 V K K V E V Y Y G M P 5 μM NVP 3 V K/P K V E V Y/C Y G M P 6 V K/P K V E V Y/C Y G M P 20 nM ETV 3 V K K V E V I/V Y G M P 6 V K K V E V I/V Y G M P 12 V K K V E V V Y G M P 15 V K K V E V I/V Y G M P 200 nM ETV 3 V K K V E V I/V Y G M P 6 V K K V E V V Y G M P 9 V K K V E V V Y G M P 12 V K K V E V V Y G M P 15 V K K V E V V Y G M P 21 V K K V E V V Y G M P 500 nM ETV 3 V K K V E V Y Y G M P 6 V K/P K V E V Y/C/I/V Y G M P 9 V K/P K V E V Y/C/I/V Y G M P 12 V K K V E V V Y G M P 15 V K K V E V V Y G M P 18 V K K V E V V Y G M P 21 V K K V E V V Y G M P 24 V K K V E V V Y G M P  NVP: nevirapine; ETV: etravirine. Mutant amino acids are printed in red text  98  Drug susceptibility assays revealed that the Y181V mutation in HIV-1 RT conferred ~60-fold resistance to ETV (Table 5.2), thus explaining - in part - its fitness advantage in our assay system. However, Y181I HIV-1 also exhibited ~60-fold decreased susceptibility to ETV, suggesting that the outgrowth of viruses containing Y181V and not Y181I was likely due to a fitness disadvantage conferred by the Y181I mutation. Similarly, the G190S and K103N mutations confer significant resistance to NVP (Table 5.2), yet the growth of both these variants was significantly outpaced by the K101P and Y181C/I/V mutants. In this regard, it should be noted that the K101P, Y181I and Y181V mutations in HIV-1 RT require a double nucleotide change. As such, despite their increased replication fitness in the presence of NVP observed in this study, they are less frequently observed in HIV-infected individuals in comparison to mutations (e.g. Y181C and K103N) that require a single nucleotide change.   Mutant Amino Acid NNRTI Parameter WT K101P K103N E138A Y181C Y181I Y181V G190S ETV EC50 (nM) 5.7 27.8 3.8 15.9 44.3 353 328 2.2 Fold-Resistance  4.9 0.7 2.8 7.8 61.9 57.9 0.4 NVP EC50 (μM) 0.28 > 20 > 20 0.23 > 20 > 20 > 20 > 20 Fold-Resistance  > 50 > 50 0.8 > 50 > 50 > 50 > 50  EC50 values represent mean values from at least 3 independent experiments. Fold-Resistance values represent mean fold change in EC50 of mutant versus wild-type (WT) virus. NVP: nevirapine; ETV: etravirine   In the previous section an efficient method to characterize the relative fitness of multiple HIV-1 mutants simultaneously was described [424]. Briefly, we generated 15 infectious viruses (HIV-1LAI) containing the single NNRTI resistance mutations V90I, K101P, K103N, V108I, E138A, E138K, V179D, 99  Y181C, Y181I, Y181V, Y188C, G190A, G190S, M230L and P236L. A wild-type (WT) virus and the 15 mutant viruses were then normalized for relative infectivity (infectious units/ng p24), pooled, and used to infect 1.6 × 106 HUT-78 cells at a multiplicity of infection of 0.004. Culture supernatants were collected at 3 day intervals until the peak of p24 production, viral RNA was extracted and deep sequencing (454 GS Junior) of an amplicon spanning RT codons 96-194 was used to quantify the distribution of mutants at each time point. Using this method, we previously showed that the Y181V mutation in the HIV-1 RT confers a clear selective advantage to the virus over the 14 other NNRTI resistance mutations in the presence of the NNRTI etravirine (ETV, Figure 5.2) [424].  When cultured in the absence of drug, the proportions of each virus in the mixture remained relatively constant (Figure 5.2A). In the presence of 200 nM ETV and 10 μM FTC, longitudinal deep sequencing revealed that the Y181V virus again emerged as the most common variant by day 12 (Figure 5.3A). Unexpectedly, we found that experiments carried out in the presence of 10 μM FTC (5× to 10× the EC50), the E138A HIV-1 exhibited a clear fitness advantage (Figure 5.3B). The frequency of the E138A variant increased from 9.6 % at day 0 to 28 % by day 6. The frequencies of HIV-1 containing the V179D or E138K substitutions were also increased, whereas all of the other NNRTI-resistant variants declined over time (Figure 5.3B).  100   The relative prevalence of 12 viruses, each containing a single resistance substitution, as determined by longitudinal deep sequencing of viral RNA from culture supernatant from cells infected with pooled virus and exposed to A) 200 nM etravirine (ETV) and 10 μM emtricitabine (FTC), or B) 10 μM FTC. Data are from a single determination. In light of the resistance data from the ECHO and THRIVE clinical trials [418,419], we asked whether the fitness landscape was altered if FTC was combined with RPV. The K101P, Y181I and Y181V mutations in HIV-1 RT are seen relatively infrequently in clinical isolates (they all require 2 nucleotide changes) and confer high-level ETV and RPV resistance [424,425]. As such, we generated a new pool of virus that included the K101E, E138G, and E138Q substitutions associated with ETV and RPV resistance (and only require 1 nucleotide change) in lieu of K101P, Y181I, and Y181V. Importantly, the effective concentration of RPV required to inhibit 50% of the pooled virus (i.e. EC50) in TZM-bl cells was found to be 2.6 ± 0.7 nM, which was identical to the concentration required to inhibit the WT virus (EC50 = 2.5 ± 0.9 nM). Interestingly, the pooled virus was ~4-fold hypersusceptible to FTC (EC50s for FTC for the WT and pooled virus were 1.3 ± 0.3 μM and 0.3 ± 0.1 μM, respectively). This new pool of virus was then used to infect HUT-78 cells in the presence of RPV, FTC or a combination of both drugs (Figure 5.4). In cultures grown in the presence of 20 nM RPV, the Y181C virus emerged as the most common variant by days 3, 6, and 9 (Figure 5.4A). In the presence of 20 μM FTC, the E138A variant again increased in frequency from 2.2 % at day 0 to 32.9 % at day 9 (Figure 101  5.4B). The frequency of the V179D also marginally increased over time (12.4 % at day 0 to 17.3 % at day 9). Initially, we tried to culture the pooled virus in the presence of 20 nM RPV and 20 μM FTC. However, we observed no virus outbreak after 24 days (data not shown). Therefore, we reduced the RPV and FTC concentrations to 10 nM and 10 μM, respectively. In cultures grown in the presence of 10 nM RPV, the Y181C virus again emerged as the most common variant (data not shown). In cultures grown in the presence of 10 μM FTC, the E138A variant increased in frequency from 2.2 % at day 0 to 15.8 % at day 12 (Figure 5.4C). However, its frequency began to decline by day 18 and was replaced with virus containing V108I and/or M184I. The M184I mutation in HIV-1 was selected de novo by FTC during the fitness experiment and pair-wise analysis of the sequencing data revealed that it was selected on both WT and V108I backbones (Figure 5.4D). In cultures grown in the presence of 10 nM RPV and 10 μM FTC, the E138A variant again increased in frequency from 2.2% at day 0 to 18% by day 18, before it was replaced with a virus population that contained Y181C and M184I (Figure 5.4E). Pairwise analysis of the sequencing data showed that the M184I substitution was selected de novo on the Y181C backbone (Figure 5.4F). Virus pools cultured in the absence of drug exhibited limited changes in the proportion of variants detected (not shown). In light of the finding that HIV-1 containing E138A exhibited a clear fitness advantage in the presence of FTC, we performed drug susceptibility assays (Table 5.3). We found that E138A decreased FTC and lamivudine susceptibility 4.7- and 6.0-fold, respectively (Student’s t-test p≤0.01), but had no effect on tenofovir susceptibility. Interestingly, we found that the E138G, E138K, E138Q and E138R substitutions did not impact FTC susceptibility (Table 5.3).  102   The relative prevalence of 12 viruses, each containing a single resistance substitution, as determined by longitudinal deep sequencing of viral RNA from culture supernatant from cells infected with pooled virus and exposed to, A) 20 nM RPV, B) 20 μM FTC, C) 10 μM FTC, or E) 10 nM RPV and 10 μM FTC. Pairwise analyses of the sequencing data show that the M184I substitution was selected on a WT and V108I backbone in experiments carried out in the presence of D) 10 μM FTC, and on a Y181C backbone in experiments carried out in the presence of F) 10 nM RPV and 10 μM FTC. Data are from a single determination. 103   FTC 3TC TDF RPV ETV Virus EC50 (μM) Fold-R (P value) EC50 (μM) Fold-R  (P value) EC50 (μM) Fold-R (P value) EC50 (nM) Fold-R (P value) EC50 (nM) Fold-R (P value) WT 1.2 ± 0.4  5.8 ± 2.5  16.1 ± 6.3  0.3 ± 0.1  1.3 ± 0.2  E138A mutant 6.0 ± 2.3 4.7  (0.006) 35.1 ± 10.0 6.0  (0.01) 8.7 ± 2.4 0.5 (>0.05) 1.5 ± 0.2 5.6  (0.03) 2.9 ± 0.5 2.2  (0.05) E138G mutant 0.5 ± 0.1 0.4  (>0.05)     0.7 ± 0.4 2.5  (0.05) 3.2 ± 0.2 2.5  (0.05) E138K mutant 2.5 ± 1.7 1.9  (>0.05) 10.7 ± 6.7 1.8  (>0.05) 6.6 ± 3.8 0.4 (0.05) 0.8 ± 0.2 3.0  (0.04) 2.8 ± 0.3 2.2  (0.05) E138Q mutant 0.4 ± 0.1 0.3  (>0.05)   5.5 ± 2.4 0.3 (0.05) 1.3 ± 0.1 4.8  (0.04) 4.0 ± 0.1 3.1  (0.01) E138R mutant 1.3 ± 0.3 1.0  (>0.05)     1.6 ± 0.4 6.0  (0.05) 5.3 ± 1.6 4.1  (0.04)  EC50 are the concentrations of drug required to inhibit viral replication by 50% from 3 independent experiments that were log10 transformed and compared for statistically significant differences (P < 0.05) by using the two-sample Student paired t test. Data are reported as means ± standard deviations from 3 independent experiments. Fold-R is the mean fold change in EC50 of mutant versus WT virus.  FTC: emtricitabine; 3TC: lamivudine; TDF: tenofovir; RPV: rilpivirine; ETV: etravirine 104   We have developed an efficient method to characterize the relative fitness of multiple NNRTI-resistant HIV-1 variants simultaneously. Using this method, we show that the Y181I and Y181V mutations confer a clear fitness advantage over other NNRTI resistance mutations in the presence of ETV in vitro. Our data is consistent with the finding that Y181I and Y181V were assigned the highest weights in a genotypic scoring system that was developed for ETV resistance-associated mutations based on their impact on treatment response [417].  In addition, we report that the E138A substitution in HIV-1 RT confers an initial clear selective fitness advantage to HIV-1 in the presence of FTC, before being outcompeted by HIV-1 containing M184I. Of note, it has been previously reported that an I132M substitution in the β7-β8 loop (that contains residue E138) of the p51 subunit of HIV-1 RT conferred nevirapine resistance but 3TC hypersusceptibility [426]). Taken together, these studies suggest that residues in the β7-β8 loop of HIV-1 RT influence nucleotide selectivity. At present, the clinical significance of this finding is unknown. However, the biological cutoffs for FTC and 3TC in the vircoTYPE HIV-1 phenotype are 3.1- and 2.1-fold changes in the calculated EC50, respectively (Janssen Diagnostics), which suggests that E138A might impact the clinical response to these nucleoside analogs. Of note, E138A is a relatively rare substitution in HIV-1 subtype B (∼2% prevalence in both therapy-naive and NNRTI-experienced HIV-infected individuals) but is quite common in both therapy-naive and NNRTI-experienced individuals infected with other HIV-1 subtypes, with a prevalence of 5 to 8% [427].  A primary limitation of these studies is that the mutations were assessed in the context of a defined genetic backbone. Furthermore, the fitness assays were carried out in a T cell line. In this regard, the results reported in this study may differ if a different strain of HIV-1 or cell line was used. More generally, this growth competition assay relies on accurate quantitation of HIV variants by deep sequencing following reverse transcription and PCR. It is therefore possible that observed sequence frequencies may not correspond directly to variant abundance as a result of unequal template 105  sampling: stochastic variation, PCR bias, and template resampling may lead to disproportionate amplification of certain template subsets. While not frequently observed in these studies, in vitro or PCR recombination cam also alter the observed frequency (and the apparent fitness) of individual resistant variants [413]. Methods to tag and track individual RNA templates may be required to control for these potential problems [428]. However, the primary advantage of the fitness assay presented here is its higher throughput relative to existing competitive or parallel culture methods. Simultaneous co-culture followed by deep sequencing allows replication fitness to be measured in a single experiment, rather than the multiple head-to-head competition assays required by traditional competitive culture methods. Simultaneous co-culture also retains the benefits of competitive culture over parallel measurements – namely, fewer assay artefacts due to differences in culture conditions between experiments, less sensitivity to variation in quantification of virus inoculum concentration, and greater sensitivity to detect small differences in fitness between variants [413]. 106    Advances and expansion in recent decades of highly-active antiretroviral therapy (HAART) have resulted in a sustained decrease in HIV-related morbidity and mortality. HIV-infected individuals receiving treatment now have “near-normal” life expectancy, to the point where HIV is now considered a manageable “chronic” disease [174,429]; Antiretroviral therapy not only provides benefits at the individual patient-level, but also results in a population-level advantage through HAART-induced suppression of HIV replication and the inherent prevention of onward transmission of the virus, coined “Treatment as Prevention” [430–433]. Drug resistance testing is an essential complement to HAART, enabling clinicians to identify patients infected with drug-resistant HIV and prescribe the appropriate antiretroviral regimens [292]. The lack of access to routine HIV drug resistance testing acts as a major barrier for long-term treatment success, either through the prescription of ineffective regimens in the case of transmitted resistance, or by limiting the ability of physicians to identify causes of treatment failure [290,292]. Persistence of the virus allows continued transmission, and can compromise management of HIV on both the patient- and population-levels [433]. These challenges are particularly acute in low- and middle-income countries (LMIC), where access to drug resistance testing is difficult due to limited resources and infrastructure. Currently, the majority of incident cases of clinically relevant HIV drug resistance involve non-nucleoside reverse transcriptase inhibitor (NNRTI) resistance, thus novel resistance testing methods for LMIC should, at minimum, target the HIV reverse transcriptase region [271,286,287].  Sanger sequencing is the current standard methodology used for HIV drug resistance testing. Amplification and sequencing is performed on the genomic regions targeted by antiretrovirals, such 107  as protease, reverse transcriptase (RT) and integrase. Sanger sequences can be analyzed for drug resistance mutations by various means, including but not limited to custom interpretive bioinformatic algorithms, linked sequence/phenotype information or proprietary software included in commercial genotyping assays (Sierra Stanford HIVdb Program, vircoTYPE HIV-1, or ViroSeq HIV-1 Genotyping System) [301,303,356]. However, Sanger sequencing is unable to reliably detect clinically relevant low-frequency drug-resistant variants [269,310,311]. In contrast, next-generation sequencing (NGS) platforms can be used to sequence diverse HIV quasispecies in order to detect rare resistant variants. In this “deep” sequencing approach, HIV RNA is PCR amplified and thousands of templates per sample are clonally sequenced. The high depth of coverage obtained, typically several thousand to tens of thousands of reads per sample [315,434], can be used to detect low-frequency variants and to study within-host HIV evolution [435]. Several studies have clearly demonstrated that the presence of low-frequency drug-resistant variants, in particular NNRTI resistance, can negatively impact treatment outcomes [313–315,436–438].  As the cost-per-base of NGS is substantially lower than Sanger, low-cost resistance testing could potentially be performed on these instruments. Furthermore, the greatly increased sequencing capacity of the newest NGS instruments offers additional benefits: Samples could be sequenced to greater depth, allowing more sensitive detection of rare variants [434]. Alternatively, a correspondingly greater number of samples could be sequenced on a single run – a strategy that we refer to as “wide” sequencing. However, to our knowledge, NGS resistance testing methods have primarily been demonstrated using low sample numbers per run [439–441,434,442]. In this proof-of-principle study we describe a feasible, high-throughput sequencing method that uses the Illumina MiSeq to produce high quality sequences for hundreds of samples in parallel. This “wide” sequencing technique spreads the read coverage of “deep” sequencing, typically concentrated on a few samples, over a larger pool of amplicons. Some samples may be sequenced to a lesser depth than “deep” sequencing approaches, but the coverage obtained for most successfully sequenced samples 108  may still be sufficient to detect lower-frequency variants. Here we evaluate “wide” sequencing in terms of its overall accuracy relative to Sanger sequencing in addition to its sensitivity and specificity in detecting antiretroviral resistance mutations.    Plasma samples were collected from Canadian patients (N=759), or from participants of the Uganda AIDS Rural Treatment Outcomes cohort (UARTO, N=349) [443]. Samples from Canadian patients were collected in 2013-2014 as part of routine physician-ordered genotypic HIV drug resistance testing at the BC Centre for Excellence in HIV/AIDS (BCCfE). Samples from the UARTO cohort comprised baseline samples collected at time of enrollment and samples collected for longitudinal virological monitoring following antiretroviral therapy initiation, up to 7.5 years post-initiation [443]. Plasma samples were stored at -20°C prior to extraction. Ethical approval was granted by the University of British Columbia Providence Health Care Research Ethics Board (H13-00395, H11-01642), the University of California Human Research Subjects Committee, the Mbarara University of Science and Technology Human Subjects Committee, the Partners Healthcare Human Subjects Committee, and the Uganda Council of Science and Technology.  Plasma viral load (pVL) data were available for a majority subset of samples (N=1068; 96.4%). Of these, the median HIV pVL was 4.2 (Q1-Q3: 3.1-4.9), and 5.2 (Q1-Q3: 4.6-5.5) log10 RNA copies/mL in the Canadian and UARTO samples, respectively, reflecting the differences in treatment experience in the two cohorts.  HIV RNA extraction was performed using the NucliSENS® easyMAG® (bioMérieux, St. Laurent, Québec, Canada) or Abbott m2000sp (Abbott Molecular, Mississauga, Ontario, Canada) instruments according to manufacturer’s instructions. A laboratory clone, pNL4-3, re-suspended in normal human plasma (N=9), or a clinical isolate, “POS08” (N=5), was included in extraction runs as internal 109  controls. Aliquots of nuclease-free water (N=21) were included as negative controls in order to identify potential PCR contamination. A total of 1108 patient samples were extracted.  A two-step RT-PCR generated HIV DNA fragments spanning the HIV protease/RT region using Expand Reverse Transcriptase and Expand High Fidelity PCR System (Roche Diagnostics, Laval, Canada) as described previously [376]. Depending on the sample source, one of three PCR products was made by nested PCR. Briefly, an amplicon spanning complete protease and RT codons 1-400 was generated for Canadian samples [444]. In cases where amplification of Canadian samples failed, a secondary amplification of a product covering protease and RT codons 1-240 was attempted [444]. For UARTO samples, a smaller amplicon covering RT codons 35-234 was generated using PCR primers designed to target conserved regions between HIV-1 subtypes A, B, C and D. PCR primers are listed in Appendix II. Amplified products were visualized on a 0.5% agarose gel and sequenced bi-directionally on an Applied Biosystems (ABI) 3730xl DNA Analyzer (Life Technologies, Carlsbad, CA, USA) using the BigDye Terminator v3.1 Cycle Sequencing kit. A median of >8 and >2 sequencing primers were used for the Canadian and UARTO samples, respectively, in order to obtain a minimum of 2-fold coverage of the amplicon.  ABI chromatograms were processed by in-house software (RECall) that automatically calls bases, trims primer sequences and constructs consensus contigs [402]. Briefly, chromatogram files were pre-processed by phred to quantify the major and minor peaks at each position and to assign quality scores. Individual sequences were then aligned to the HIV HXB2 reference (GenBank accession number K03455.1) and a single contig was generated. Mixtures were called when the peak area of the minor base exceeds 20% that of the major base across the majority of reads covering that position. RECall’s 110  default quality control criteria identified potential problematic bases and excluded any Sanger sequences failing to meet these criteria [402]. Sanger sequences containing unknown bases (N, aNy base) were excluded from the analysis to allow complete comparisons of drug resistance profiles. All Sanger consensus sequences were then trimmed to a 435-bp region containing RT codons 90-234 to correspond to the region covered by the MiSeq sequences (described below).  To minimize the potential variability introduced by the RNA extraction and reverse transcription steps, MiSeq library preparation began from the same first-round PCR products generated during the Sanger sequencing procedure. As Canadian samples arrived at the BCCfE over a period of months for routine clinical drug resistance testing, PCR products were stored at room temperature for up to 12 months prior to the initiation of this study, with >75% of samples stored for no longer than 6 months. A fragment spanning RT 90-234 was amplified in a nested second-round PCR with primers incorporating Illumina indexing adaptors (Appendix II). A dual-index sequencing strategy was used in order to minimize the number of primers needed for such a large multiplex run. Briefly, 8bp-long indices (or “barcodes”) are added to both the 5’ and 3’ ends of each amplified fragment. A unique pair of indices is used for each sample. Subsequent sequencing of each index pair unambiguously identifies a sequence read as belonging to a specific sample [445]. A total of 24 i5 “forward” and 48 i7 “reverse” indices were used, allowing up to 1152 samples to be sequenced simultaneously (Appendix III). Index tags were added in a third low-cycle PCR as used in the Illumina Nextera XT indexing procedure. Indexed MiSeq amplicons were purified and normalized using Agencourt AMPure XP magnetic capture beads and pooled into 12 DNA libraries (96 amplicons/library). Library concentrations were determined using the Invitrogen Quant-iT PicoGreen dsDNA Assay. The 12 libraries were pooled at an equimolar concentration (1.0 ng/µL) prior to MiSeq sequencing with a 2x250-bp v2 kit. Illumina currently recommends a minimum 5% PhiX spike-in to avoid sequencing problems associated with low-diversity libraries. A more conservative 10% PhiX spike-in was used in this experiment. Thus, all 1143 samples were sequenced on a single MiSeq run. 111   MiSeq short read data were processed by an in-house pipeline using bowtie2 and samtools [446,447]. Reads were initially mapped to the HXB2 reference; the mapped reads were used to generate sample-specific consensus sequences to which the entire set of reads was subsequently re-mapped. Consensus sequence generation and re-mapping proceeded iteratively until >95% of all reads mapped successfully or no improvement in the number of mapped reads was observed. Paired-end reads were then merged, with any differences in base calls in the overlapping portion resolved using a quality score-informed algorithm. In brief, bases with a quality score below 15 were censored as "N"s. Any discordant bases between paired-end reads were assigned the base with the higher quality score if the difference was 5 or greater; otherwise, the merged base was censored.  MiSeq consensus sequences spanning RT codons 90-234 were produced from the empirical nucleotide frequency distributions in the aligned and merged read data. An ambiguity threshold was used to allow minority bases present above the threshold to be included in consensus sequences as mixtures. A threshold of 20% was chosen in order to mimic the parameters of the RECall software used in Sanger sequence analysis. A depth of coverage parameter was used for quality control; sequences with low coverage (defined a priori as <100 reads at any position) were rejected.  Samples that were successfully amplified and sequenced by both Sanger and MiSeq methods were assessed for nucleotide concordance, calculated as the proportion of sequence agreement observed across all nucleotides sequenced. Any differences in nucleotide base calls, including differences in mixture calling, were considered mismatches. For technical replicates (pNL4-3 clone, POS08 clinical sample), a consensus sequence constructed from all available replicate sequences was used to assess inter-assay sequencing variability. MiSeq controls were compared to their corresponding Sanger counterpart to determine variability between sequencing methods. Nucleotides appearing in ≥20% of 112  replicate sequences were included as mixtures in the consensus. HIV subtyping of MiSeq consensus sequences was performed by RIP using a 90% confidence threshold and a 200-bp window size [448]. Agreement in resistance interpretations between methods was also evaluated using Cohen’s Kappa statistic [359]. Consensus sequences were translated into amino acids and resistance mutations were identified according to the 2013 IAS-USA list [299]. Codons containing ambiguous nucleotides (mixtures) were translated to include all possible amino acids in drug resistance analyses.    A total of 881 (80%) patient samples were successfully sequenced by the Sanger protocol. Of these, 576 (76%) of the Canadian samples and 305 (85%) from the UARTO cohort were successfully sequenced. In addition, 12 (86%) technical replicate samples were sequenced by the Sanger protocol. Successfully sequenced pNL4-3 controls (N=7; 78%) were clonal, with 100% concordance between replicates, although two pNL4-3 failures were observed. Overall, 98.0% nucleotide concordance was observed between successful POS08 replicates (N=5, 100%), with three nucleotide mismatches found in key resistance-associated positions. All differences between POS08 replicates were due to differences in mixture calls, all of which were compatible. Note that both Canadian and UARTO groups (primarily the former) contained samples from patients with low-level viremia (pVL <1000 HIV RNA copies/mL) or emergent virological failure, representing a reasonable cross-section of samples sent for HIV drug resistance testing. For example, 15% had viral loads between 50 and 1000 copies/mL, 18% between 1000 and 10,000 copies/mL, 31% between 10,000 and 100,000 copies/mL and 35% above 100,000 copies/mL. The MiSeq-reported quality metrics, taken from the average across all sequence and index reads, were consistent with typical values observed in previous runs: 1086 K/mm² cluster density, 80.9% of bases ≥Q30, 77.4% clusters passing filters, 1.6% PhiX sequencing mean error rate. Median MiSeq coverage 113  in successfully sequenced samples was ~9,900 paired-end reads/sample (Q1-Q3: 5,300-13,400). Variation in sequence coverage across the amplicon was minimal, with the exception of a small drop in coverage between RT codons 167-171, corresponding to decreased sequencing quality at the end of the “reverse” R2 Illumina read. (Figure 6.1). The distribution of MiSeq read coverage was comparable across all pVL strata.  Depth of MiSeq coverage at each position across the sequenced amplicon (RT codons 90-234) after quality control filters were applied (N=892 clinical samples). The red line represents the median coverage at each position, with the interquartile range displayed by the lighter red shading. Paired-end 2x250-bp sequencing kits were used. The arrows indicate positions covered by each read in the pair. The small drop in coverage observed around RT codon 170 reflects lower-quality bases at the end of Read 2 that were discarded during data processing. In total, 892 (81%) clinical samples were successfully sequenced by MiSeq, with 579 (76%) Canadian and 313 (87%) UARTO passing the arbitrarily predefined 100-fold coverage requirement. In addition, all pNL4-3 replicates and 4 (80%) POS08 samples were successfully sequenced by MiSeq, with 100% and 98.6% nucleotide concordance observed among replicates, respectively. A single nucleotide mismatch was found in a key resistance-associated position in one POS08 sample. Again, differences in POS08 replicate sequences were the result of mixture calls; no incompatible differences were identified.  114  The success rates of both Sanger and MiSeq sequencing were largely driven by pVL, with neither method demonstrating a bias towards preferential amplification across pVL strata (Figure 6.2). Furthermore, the sequencing success rate between cohorts was similar within each pVL stratum suggesting that neither method preferentially amplified a particular subtype (Appendix IV). The median pVL of samples failing sequencing was 2.9 (Q1-Q3: 2.4-3.9), and 2.7 (Q1-Q3: 2.4-3.7) log10 copies/mL by the Sanger and MiSeq methods, respectively. The median pVL of samples failing sequencing by both methods was 2.6 (Q1-Q3: 2.2-3.4) log10 copies/mL. Overall, sequencing by either Sanger or MiSeq was successful in >88% of samples with pVL >3.0 log10 copies/mL, with ~85% being successfully sequenced by both methods. Duration of room-temperature storage of the first-round PCR products may have influenced MiSeq sequencing success rates. Success rates were marginally higher for samples stored for less than 6 months than those stored for longer periods (94% vs. 82% for Canadian samples with pVL>1000 copies/mL). 115   Overall, 881 (80%) and 892 (81%) clinical samples were successfully sequenced by the Sanger and MiSeq methods, respectively with 832 (75%) having sequences from both methods. Sequencing failure rate was driven largely by sample pVL. Overall, 793 (88%) and 810 (90%) of samples with pVL >3.0 log10 copies/mL were successfully sequenced by Sanger and MiSeq, respectively. Numbers above bars represent the total number of samples tested in each pVL category.  Overall, 832 (75%) clinical and 11 (79%) technical replicate samples were successfully sequenced by both Sanger and MiSeq. A total of 2,466 nucleotide mismatches were identified from 366,705 called bases in these samples (99.3% concordance; Figure 6.3). Sample mix-up or cross-contamination was not identified in samples with low (<98%) sequence concordance. The overwhelming majority (>95%) of observed discordances were differences in mixture calls, with neither method over- or under-calling mixtures relative to the other. Substantial agreement between methods on the identity of bases (mixed 116  base vs. not) was observed (Cohen’s κ =0.72) [359]. Note that the use of the same first-round PCR product by both methods eliminates a source of potential random variation inherent in the RNA extraction and reverse transcription steps. Thus, the observed nucleotide differences between methods are effectively limited to sequencing errors. Comparisons of the laboratory clone pNL4-3 showed clonality between all sequences both within and between methods. As previously stated, the clinical isolate POS08 showed some variability between replicates among methods (~98% concordance); however, high concordance between methods was observed when paired samples from the same first-round PCR product were compared (99.1% nucleotide concordance). The MiSeq protocol was able to successfully amplify HIV RT from a variety of HIV subtypes. Phylogenetic analysis of MiSeq sample sequences and HIV subtype reference sequences (downloaded from the Los Alamos HIV Sequence Database) displayed clustering by cohort (Figure 6.4) consistent with the expected subtype prevalence given our cohort demographics. Overall, Canadian samples were primarily subtype B (84%), with subtype C (12%) being the next most common. UARTO samples were primarily subtypes A (43%) and D (49%), with small numbers of recombinant sequences also observed (Appendix V). 117   Change matrix displaying the frequency of nucleotides detected by Sanger (rows) and MiSeq (columns) sequencing. Overall, 832 clinical and 11 technical replicate samples were successfully sequenced by both methods. A total of 2,466 mismatches were identified in 366,705 called bases in these samples. Concordant base calls are highlighted in green. Partially discordant base calls (mixed bases detected by one method, but not the other) are highlighted in yellow. Entirely discordant base calls are highlighted in red. Differences in mixture detection/calling account for >95% of all discordant bases. Columns for B, D, H, and V are not shown for Sanger base calls, as the RECall software does not call three-base mixtures. 118   Neighbor-Joining tree constructed from MiSeq consensus sequences (N=892 clinical samples) depicted as a cladogram (unscaled branch lengths). Mixed bases were called when minority bases exceeded 20% prevalence. Samples with <100-fold coverage were excluded. Sequences cluster by cohort and HIV subtype in agreement with expected prevalence. HIV-1 subtype consensus sequences (N=16; black tip labels) spanning RT codons 90-234 were included and represent subtypes A1, A2, B, C, D, F1, F2, G and H as well as recombinant viruses AE, AG, AB, BC, CD, BF and BG (retrieved from http://www.hiv.lanl.gov/). Canadian HIV RT sequences (blue tip labels) were primarily subtypes B and C, and UARTO HIV RT sequences (pink tip labels) were primarily subtypes D and A.  119   Drug resistance analysis of the 832 mutually-successful clinical samples revealed that 155 Sanger and 156 MiSeq samples possessed one or more resistance mutations found in the 2013 IAS-USA list (κ=0.96) (Appendix VI). MiSeq had 97.4% sensitivity and 99.3% specificity in detecting resistance mutations identified by Sanger. Assuming any one major drug resistance mutation confers a resistant phenotype, analysis of resistance by drug class revealed near-identical interpretations between methods. NRTI resistance mutations were detected in 62 (7.5%) and 63 (7.6%) samples sequenced by Sanger and MiSeq, respectively (κ=0.96). Similarly, NNRTI resistance mutations were observed in 119 (14.3%) and 120 (14.4%) of Sanger and MiSeq samples, respectively (κ=0.98). Sanger and MiSeq were 99.4% concordant in resistance interpretations for both NRTI and NNRTI. Given the relatively high depth of coverage obtained it may be possible to identify lower prevalence resistance mutations in the MiSeq data by lowering the threshold at which nucleotide mixtures are called. For example, 73 (8.8%) and 145 (17.4%) samples with NRTI and NNRTI mutations at >5% frequency were identified (Figure 6.5); however, the appropriateness of the mixture calling threshold and the clinical relevance of any low-frequency variants detected in this manner requires further evaluation.   120   Drug resistance interpretations of clinical samples (N=832) successfully sequenced by Sanger and MiSeq methods. Orange and blue bars represent the proportion of samples with observed NNRTI and NRTI resistance (≥ 1 mutation) from MiSeq analysis. The dashed and dotted lines represent the results from Sanger analysis for NNRTI and NRTI resistance, respectively. The effect of varying the MiSeq nucleotide mixture calling threshold suggests that the sequence coverage obtained may be sufficient to identify lower-frequency resistance mutations.  The suitability of the chosen minimum sequence coverage, and mixture calling thresholds was evaluated by recalculating the overall nucleotide concordance as these parameters were adjusted. Analysis of coverage and mixture cutoff parameters indicated that the a priori chosen 100-fold minimum coverage and 20% mixture thresholds are suitable to achieve reliable nucleotide concordance data (Figure 6.6). Nucleotide concordance begins to decrease when minority bases are called at 15% or lower, or if samples below 50-fold coverage are included. Sequencing failure rates 121  increase as the minimum coverage threshold is increased; however, limited improvement in sequence concordance is observed. Although failure rates were substantially higher in samples with low pVL, nucleotide concordance rates were comparable across pVL strata (Figure 6.7, Appendix VII). ROC analysis suggests a minimum coverage cutoff as low as 50 reads still yields acceptable sensitivity and specificity measurements for detecting resistance (Appendix VIII).  Raw nucleotide concordance between MiSeq and Sanger sequences derived from clinical samples (Panel A) increases rapidly as the minimum coverage threshold is increased from 1 to 75-fold for all of mixture-calling thresholds examined. Sequencing accuracy is highest at the 20% mixture cutoff for all levels of coverage >10 reads/sample. The number of successfully sequenced clinical samples decreases as the minimum coverage threshold is increased (Panel B). Nucleotide concordance at the 20% mixture-calling threshold (Panel A) and the number of samples successfully sequenced (Panel B) are displayed for coverage cutoffs of 10, 100, 1000 and 10,000 reads/sample.  122   Sequence concordance was high across all pVL strata. Outliers beyond 1.5 IQR of the box hinge, represented by dots, are due largely to high numbers of mixed base calls in selected MiSeq sequences. Clinical samples without viral load data (“Unknown”) were also successfully sequenced by both methods and yielded generally concordant results. Numbers above boxes represent the total number of samples successfully sequenced by both methods in each pVL category.  The introduction of the large data-generating capacity of “next-generation” sequencing platforms has markedly decreased the cost-per-base of DNA sequencing. At present, laboratories performing clinical HIV drug resistance genotyping do not generally benefit from these cost savings. Limited sample numbers, the requirement for rapid turnaround times and a comparatively small target genome mean that the cost-per-sample of HIV resistance testing remains relatively unchanged.  123  The most common current application of next-generation platforms in clinical HIV setting has lain in “deep” sequencing of diverse virus populations; the goal being to identify low-frequency variants that may influence treatment outcomes [313–315,436–438]. However, in a setting where large numbers of samples may be routinely processed, or where requirements for turnaround times may be relaxed, it may be possible to perform cost-effective drug resistance genotyping by sequencing hundreds of samples in parallel. This high-multiplex “wide” sequencing would effectively spread the coverage conferred from deep sequencing across each sample. This concept has been proposed previously and partially evaluated on other platforms; however, neither the evaluation of such a large sample pool nor a systematic comparison of sequence concordance the current standard methodology (Sanger) has been performed [440,441]. In this proof-of-principle study we demonstrated that high-multiplex sequencing on the Illumina MiSeq could accurately replicate Sanger sequencing of HIV RT over a wide range of pVL inputs. In samples with >100-fold MiSeq coverage, consensus sequences (where minority bases >20% frequency were called as mixtures) were compared to Sanger sequences and >99.3% nucleotide concordance was observed – a value comparable to that seen in studies of inter-laboratory variability of Sanger sequencing of HIV [333,334]. The majority of discordances was due to differences in mixture calling and did not substantially impact drug resistance interpretations. Amplification and sequencing success rates were comparable (~80%) for both methods, though it should be noted that these likely represent underestimates of the true success rates, as PCR was not reattempted on failed samples. For example, the Sanger assay was only able to successfully amplify and sequence ~40% of samples with pVL <1000 copies/mL in this experiment. In contrast, our laboratory typically achieves a >85% success rate in low pVL samples if samples that initially fail to produce a PCR product are retested with a backup protocol [444]. Furthermore, the sequencing success rates of the internal controls (pNL4-3, POS08) were comparable to those of the clinical samples for both the Sanger (86%) and MiSeq (93%) assays. Importantly, MiSeq and Sanger methods were both able to amplify HIV RT from multiple subtypes. Taking into account the differences in amplicon size generated (~600-bp for UARTO, up to 124  ~1700-bp for Canadian) and the pVL distribution between cohorts, no obvious differences in success rates were observed between cohorts or HIV subtype.  A previous study evaluated a pooled sequencing strategy for HIV drug resistance surveillance using the Roche 454 GS-FLX [441]. The authors demonstrated that protease inhibitor resistance mutations and sequence polymorphisms relative to the HXB2 reference observed in Sanger sequences were also identified in the pooled sequencing approach. In addition, pooled pyrosequencing represented an estimated ~35% cost savings relative to Sanger making it an attractive option for population resistance surveillance. However, as the individual samples in the pool were not uniquely barcoded, sequences from individual samples could not be distinguished and thus no direct comparisons to Sanger sequencing could be performed on a per-sample basis. A similar study demonstrated that up to 48 samples could be multiplexed on a single Roche 454 GS Junior run at a cost of ~$20/sample; however, the accuracy relative to Sanger sequencing was not evaluated [440].  In contrast to the soon-to-be discontinued 454 system, this “wide” sequencing approach offers several advantages. First, the higher multiplexing density and the use of MiSeq rather than 454 results in a further reduction in the cost of sequencing, down to ~$1/sample. More importantly, unique barcoding of samples allows sequences from individual samples to be recovered, allowing individual-level clinical decisions to be made from “pooled” sequencing data. The relatively high depth of coverage (>9,900-fold) obtained on the majority of samples tested suggests that minority variant detection using the “wide” sequencing approach is possible as the level of coverage achieved is comparable to that presented in earlier studies of HIV resistance testing by “deep” sequencing [315,434]. However, the clinical relevance of any minority variants identified by this method requires further investigation. Finally, the increasing availability of viral load testing in low- and middle-income countries (LMIC) may offer the potential for even further cost-savings. In this study, RNA extraction was performed with two automated nucleic acid extractors that are used in the preparation of samples for pVL testing (NucliSENS easyMAG, Abbott m2000sp). For both of these instruments, more RNA is eluted than is required for the pVL assay and this excess material is recoverable. We therefore envision a strategy in 125  which laboratories performing HIV pVL testing could use this excess RNA to perform the initial RT-PCR reaction described here. The resulting, heat-stable, non-biohazardous material could be shipped by regular mail to a centralized sequencing facility where the nested PCR reaction and MiSeq sequencing would be performed. Such a strategy would eliminate many of the logistical barriers to performing genotypic resistance testing in these settings. Therefore, “wide” sequencing also presents an opportunity to pursue a complementary paradigm to expanded testing from dried blood spots [449–451]. Furthermore, the marginal cost increase of the RT-PCR reaction to the pVL testing laboratory (estimated ~$5/sample) is offset by the elimination of several costly sample processing steps that are currently duplicated in pVL and drug resistance testing procedures, (e.g. sample collection, accessioning, hazardous goods shipping, RNA extraction). Thus, such an integrated strategy using centralized sequencing facilities (rather than individual local/regional laboratories) could represent a net cost-savings to the healthcare system. Several limitations to this “wide” sequencing approach should be addressed to validate this strategy. First, the requirement for a third “indexing” PCR reaction and a bead-based normalization step introduced an incremental cost increase to MiSeq sample preparation (estimated ~$2/sample). Note, however, that most of this cost can be eliminated by using barcoded primers in the second PCR reaction, thus avoiding the “indexing” PCR step entirely. Most importantly, this proof-of-principle study evaluated only a small portion of HIV reverse transcriptase. All of the major NNRTI-associated resistance mutations in the IAS-USA 2013 list [299] are covered by the sequenced amplicon; however, important NRTI resistance mutations, notably K65R, and all of protease were not. It would be possible to construct a longer amplicon covering RT codon 65 which could be sequenced using 2x300-bp Illumina MiSeq kits. A second amplicon covering protease could also be prepared and sequenced on the same run. The sequence coverage of each sample would effectively be halved if the same multiplexing density was maintained, though we have demonstrated that this would have minimal impact on sequencing accuracy. Sequencing additional amplicons, however, would marginally increase the preparation costs (only one additional second PCR reaction would be required) and create additional sample handling complexity.  126  Second, managing sample preparation of hundreds of samples in parallel is complicated compared to conventional Sanger sequencing and could increase the likelihood of sample mix-up or cross-contamination. For this reason, automated pipetting instruments are suggested, especially for the index PCR step. However, the observation that MiSeq consensus sequences are highly concordant with those generated by Sanger sequencing may allow the existing quality control tools designed for Sanger sequencing to be re-purposed for deep- or wide- sequencing approaches [452,453].  Finally, consensus sequence generation and minority variant detection by “wide” sequencing requires stringent bioinformatics pipelines for accurate analyses. Such tools might include optimization of mixture and depth of coverage cutoff parameters, or further inclusion/exclusion criteria based on the number of observed mixtures per sample. The effect of detecting minority drug resistance variants on treatment outcome is less clear. Criteria for variant detection in population Sanger sequencing have been established; if a mixture is detected above the threshold of about 20%, it is considered clinically important [332]. No definitive standards for NGS and minority variant detection have been agreed upon for use in a clinical setting. These NGS tools need to be established and validated before adapting this strategy to clinical practice. Unfortunately, the nature of the samples selected for this study did not allow validation analyses to be performed here; Canadian samples consisted of an arbitrary cross-section of contemporaneous samples from treated patients with limited subsequent follow-up. Treatment outcome data and MiSeq sequences were available for 212 antiretroviral-naïve participants of the UARTO cohort who initiated NNRTI-based HAART. However, we lacked power to detect any meaningful association of minority resistant variants with virologic outcome as over 95% of these individuals maintained pVL below 50 RNA copies/mL over the course of follow-up. Drug resistance testing provides physicians with clinically-important information, enabling treatment decisions to be made on an individual basis, while assisting surveillance of HIV resistance transmission on a population level [454]. In LMIC, access to HIV drug resistance testing continues to be limited. This has substantial clinical and epidemiological implications; patients in LMIC often remain on sub-optimal ART regimens enabling the transmission of drug-resistant variants [286,287]. 127  In this study, HIV drug resistance testing was attempted on 1108 samples, representing approximately 20% of Canada’s annual HIV drug resistance testing burden, in a single MiSeq run. The potential to use surplus RNA from pVL testing presents the opportunity to further reduce costs and some of the logistical difficulties associated with resistance testing in LMIC. HIV RT from multiple subtypes was successfully and accurately sequenced suggesting that routine individual testing and/or annual population resistance surveillance in LMIC could be performed with this strategy. 128    Hepatitis C Virus (HCV), the cause of hepatitis C disease in humans, is a positive-sense single stranded RNA virus of the Flaviviridae family [455–457]. An estimated 2-3% of the global population has ever been infected with HCV, with 130-150 million individuals being chronically-infected [138,458]. The total number of deaths worldwide due to HCV in 2010 was estimated to be 499,000 [8]. In Canada, over 240,000 individuals are estimated to be HCV-infected [139]. While approximately a quarter of acutely-infected individuals are able to spontaneously clear the virus [161], the remainder develop chronic HCV infection [459–461] and are thus at elevated risk to develop liver cirrhosis and/or hepatocellular carcinoma [140,141]. HCV is classified into 7 major genotypes (GT1-7), each with several subtypes [170]. In North America, HCV GT1 (subtypes 1a and 1b) comprises roughly 70% of infections, with GT2 and GT3 being the next most common [171,172].  Until recently, the standard of care for HCV infection has been combination antiviral therapy with pegylated interferon-alfa and ribavirin (PegIFN/RBV). Sustained virologic response (SVR) rates to PegIFN/RBV therapy vary by HCV genotype, with ~40% of patients with HCV GT1 responding favorably compared to >70% for other genotypes [203]. In 2011, the NS3 protease inhibitors (PI) telaprevir and boceprevir, in combination with PegIFN/RBV, were the first direct-acting antiviral (DAA) agents approved for treatment of chronic HCV GT1 infection [210–213]. The success of HCV therapy with DAA, however, is complicated by the virus’ incredible genetic diversity and its capacity to mutate in response to drug selection pressure [462,463]. Treatment failure is often accompanied by the emergence of resistance mutations in the genes targeted by these drugs [234,235]. Furthermore, 129  certain drug resistance mutations exist as naturally-occurring polymorphisms in a small proportion of treatment-naïve patients and can compromise PI treatment in these individuals [464–467]. In combination with PegIFN/RBV, the second-generation PI simeprevir was approved in Canada in 2013 for the treatment of chronic HCV GT1 infection in adults with compensated liver disease, including cirrhosis [468,469]. The simeprevir combination was shown to be superior to PegIFN/RBV alone with SVR >80% being achieved in both the QUEST-1 and QUEST-2 phase III clinical trials [214,215]. However, SVR rates for the simeprevir combination were reduced to 58% in patients having HCV genotype 1a (GT1a) with the NS3 Q80K polymorphism at baseline; this SVR rate was non-superior to that observed in the placebo arm. Overall, 56% of patients with GT1a who did not achieve SVR in the simeprevir arms had the NS3 Q80K at baseline. In subsequent retrospective genotyping studies, it was discovered that approximately 30% of patients with GT1a enrolled in the phase II and III clinical trials of simeprevir had HCV harboring the Q80K polymorphism at baseline [470]. In addition, a significant geographic bias in the distribution of Q80K was discovered: 48% of patients with HCV GT1a in North America had the Q80K at baseline, compared to 19% of patients in Europe. In contrast, only 0.5% of patients with GT1b HCV were infected with viruses carrying the Q80K polymorphism and no geographical differences were observed. The Q80K polymorphism is stable; viruses carrying the polymorphism are transmissible and are likely descended from a single lineage originating in the United States in which the Q80K substitution occurred around the 1940s [471]. Owing to the stability and high frequency of this polymorphism in Europe and especially North America, screening for the Q80K polymorphism is strongly recommended before initiating simeprevir therapy in patients with HCV GT1a infection [472]. Here, we present methods and demonstrate performance of two independent HCV NS3 Q80K polymorphism assays involving nested RT-PCR and sequencing of a portion of the NS3 protease region: 1) A Sanger sequencing approach incorporating “primary” and “secondary” PCR methods, and 2) a next-generation sequencing approach involving near-whole-genome amplification and sequencing on an Illumina MiSeq sequencer.  130    Janssen Diagnostics BVBA provided frozen plasma samples from 70 treatment-naïve HCV GT1-infected participants from the QUEST-1 and QUEST-2 phase III clinical trials of simeprevir in order to test sequencing accuracy. Median HCV plasma viral load (pVL) was 6.7 log10 IU/mL (IQR: 6.1-6.9 log10 IU/mL; Range: 4.9-7.5 log10 IU/mL). HCV NS3 was previously sequenced in these samples at the Janssen Diagnostics Laboratory in Beerse, Belgium. The BC Centre for Excellence in HIV/AIDS (BCCfE) laboratory remained blinded to the Janssen sequencing results and sample collection details (study arm, timing) throughout assay development and validation.  In addition, archived frozen (-70°C) plasma samples from HCV GT1-infected participants of the Vancouver Injection Drug Users Study (VIDUS) were screened to identify two sample groups: one having wild-type virus and one with the Q80K polymorphism. These sample sets were used for inter- and intra-assay precision studies. HCV pVL was unknown for these samples. A single HCV-positive GT1a plasma sample with pVL 6.6 log10 IU/mL (SeraCare Life Sciences, Milford, MA, USA) was spiked with HIV-positive plasma and with HBV-positive plasma to test potential interference by other viruses. An HCV subtype panel (SeraCare) was used to test cross-reactivity across HCV genotypes (pVL Range 3.7-4.2 log10 IU/mL). Finally, in order to investigate analytical specificity, HCV-negative samples were tested: DEPC-treated water, pooled normal human plasma, HBV-positive/HCV-negative plasma (SeraCare), and in-house HIV-positive/HCV-negative controls.   HCV viral RNA was extracted from 500 μL of frozen plasma using a NucliSENS easyMAG automated nucleic acid extractor (bioMerieux Canada, St-Laurent, QC, Canada) per the manufacturer’s instructions. Extracted RNA was eluted in 60 μL elution buffer and was stored at -20°C until RT-PCR 131  amplification. Where 60 μL of extracted RNA was insufficient for the intended experiments, multiple extractions were performed and eluates pooled prior to further processing.  Extracted HCV RNA was amplified using the QIAGEN OneStep RT-PCR kit (QIAGEN Sciences, Valencia, CA, USA) followed by an in-house nested second round PCR protocol. Two independent nested RT-PCR amplification reactions intended to be “primary” and “secondary” (to be used in case of initial assay failure) methods were designed and tested in parallel. Both amplicons were designed to cover Q80 as well as all the major NS3 protease inhibitor resistance mutations V36A/M, T54A/S, V55A, S122R, R155K/Q, A156S/G, D168A/E/H/T/V/Y, and V/I170A. As simeprevir is approved only for the treatment of HCV GT1 infections and the Q80K polymorphism is mainly limited to HCV GT1a, assay primers were designed to maximize the amplification success rate of HCV GT1 samples. The “primary” ~1.4-kb amplicon was generated using reverse transcription primer NSR1 and two forward first PCR primers NSF1 and 5HCPROT1 (See Appendix IX for RT-PCR primer sequences). Briefly, 8 μL of extracted RNA was used to generate and amplify cDNA into a total reaction volume of 40 μL consisting of 17.35 μL DEPC-treated water, 8 μL 5X OneStep RT-PCR Buffer (QIAGEN), 1.6 μL 10 mM dNTP Mix (QIAGEN), 1.2 μL 25 μM NSR1 primer, 1 μL 25 μM NSF1 primer, 1 μL 25 μM 5HCPROT1 primer, 0.25 μL Protector RNAse Inhibitor (Roche) and 1.6 μL OneStep RT-PCR Enzyme Mix (QIAGEN). Thermal cycling conditions were: 54°C @ 30’; 95°C @ 15’; 8 cycles [94°C @ 30”, 64°C @ 30” (-0.5°C/cycle), 72°C @ 1’20”]; 30 cycles [94°C @ 30”, 60°C @ 30”, 72°C @ 1’25”].  The “secondary” ~800-bp amplicon was generated using reverse transcription primer HCV1NS3SR1 and forward first PCR primer HCV1NS3SF1. Thermal cycling conditions and reaction mixes were modified slightly from the primary method. Briefly, 1.2 μL each of 25 μM HCV1NS3SR1 and 25 μM HCV1NS3SF1 primer were used and an additional 0.8 μL DEPC-treated water used to bring the total reaction volume to 40 μL. RT-PCR thermal cycling conditions for the secondary method were: 54°C @ 132  30’; 95°C @ 15’; 10 cycles [94°C @ 30”, 62°C @ 30” (-0.5°C/cycle), 72°C @ 45”]; 30 cycles [94°C @ 30”, 58°C @ 30”, 72°C @ 50”].  For the primary method, a second nested PCR reaction using forward primer NSF2 and reverse primers and 3HCPROT2 was performed (See Appendix IX for PCR primer sequences). The total reaction volume of 20 μL consisted of 2 μL first round PCR product, 12.39 μL DEPC-treated water, 2 μL 60% sucrose with 0.08% cresol red, 2 μL Expand High Fidelity 10X Buffer with 15 mM MgCl2 (Roche), 0.8 μL 25mM MgCl2 Stock Solution (Roche), 0.16 μL 100 mM dNTP (Roche), 0.29 μL Expand High Fidelity Enzyme Mix (Roche), and 0.12 μL each of 25 μM primers NSF2, NSR2, and 3HCPROT2. Thermal cycling conditions for the nested second round PCR were: 95°C @ 3’; 8 cycles [94°C @ 15”, 64°C @ 30” (-0.5°C/cycle), 72°C @ 1’15”]; 30 cycles [94°C @ 15”, 60°C @ 30”, 72°C @ 1’15” (+3”/cycle)]. The secondary PCR method used nested second round PCR forward primer HCV1NS3SF2 and reverse primer HCV1NS3SR2. PCR primers from the primary nested PCR reaction mix were substituted with 0.12 μL each of 25 μM primers HCV1NS3SF2 and HCV1NS3SR2, and DEPC-treated water. Thermal cycling conditions for the secondary method: 95°C @ 3’; 10 cycles [94°C @ 15”, 64°C @ 30” (-0.5°C/cycle), 72°C @ 45”]; 30 cycles [94°C @ 15”, 58°C @ 30”, 72°C @ 50” (+3”/cycle)]. Bulk (population) sequencing was performed on the amplified products on an ABI 3730xl DNA Analyzer (Life Technologies, Carlsbad, CA, USA). Sequencing primers are listed in Appendix X. Chromatograms were analyzed by the in-house software RECall (version 2.25) [402]. To account for the substantial sequence variation between HCV genotypes sequence reads were aligned against a set of seven HCV NS3 genotype consensus sequences (GT1a, GT1b, GT2a, GT2b, GT3a, GT4, GT6a) with the highest-scoring reference being subsequently chosen for contig assembly. Nucleotide mixtures were called automatically when the secondary peak area exceeded 20% of the major peak. No human editing of base calls was performed. Software settings allowed single primer coverage except at NS3 133  codon 80 and short sections of poor quality sequence in regions not affecting Q80K calls may have been excluded. The final assembled sequences covered 1200-bp (H77 positions 3420-4619) and 564-bp (H77 positions 3420-3983) for the primary and secondary assays, respectively.   In addition to the NS3 amplicon protocol using Sanger sequencing, a separate assay using next-generation (“deep”) sequencing of the whole HCV genome was also investigated as an independent comparison. Briefly, a near-full-length amplicon was generated according to a previously published protocol [473] using an oligo d(A)20 primer for cDNA synthesis. Two nested PCR reactions using primers optimized for GT1a and GT1b were used to generate an 8991-bp amplicon spanning Core to partial NS5B. Following second round PCR amplification, libraries for MiSeq sequencing were prepared using Nextera XT DNA Sample Preparation Kits (Illumina, San Diego, CA, USA) according to the manufacturer’s specifications. Amplicons were uniquely tagged using a dual-indexing approach (Nextera XT Index Kit, Illumina), multiplexed 48-fold and sequenced using MiSeq v2 paired-end kits (2x250-bp).  The resulting reads were processed using an in-house pipeline that incorporated iterative short read mapping with bowtie2 [446] and samtools [447]. Briefly, paired-end reads were initially mapped to the H77 reference using default settings for local alignment [446]. Mapped reads were collapsed into a sample-specific consensus to which all sequenced reads were subsequently re-mapped. Consensus building and re-mapping proceeded in an iterative fashion until ≥95% of reads mapped or no additional reads could be recovered by additional rounds of re-mapping. The overlapping portions of paired-end reads were then merged and error correction rules were applied: Bases with sequencing quality scores <15 were discarded; conflicting base calls in overlapping reads were resolved by retaining the base with the higher quality score. In order to compare MiSeq and Sanger sequencing results, consensus sequences for NS3 were generated from the resulting SAM files with nucleotides 134  >20% frequency being called as mixtures to mimic the expected sensitivity of Sanger sequencing. Samples with <1000-fold coverage at NS3 codon 80 were excluded from downstream analysis.  Three metrics were used to assess the quality of PCR amplification and sequencing between methods and/or replicates: Concordance of nucleotide base calls, concordance of amino acid sequences, and concordance of the presence or absence of the NS3 Q80K polymorphism. Concordance was calculated as the proportion of nucleotide or amino acid agreement observed across all the nucleotide/amino acids sequenced. For the purpose of validation, “partial” discordances in nucleotide or amino acid calls (when one method observed a mixture, while the other method detected only one component thereof; e.g. nucleotides Y vs. C) were weighted equally as complete discordances. In analyses involving amplification and sequencing of multiple replicates of a sample for which no “gold standard” reference sequence was available, a consensus sequence constructed from all available replicates was used as the comparator. Nucleotides appearing in ≥20% of replicate sequences were included (as mixtures) in the consensus. Assay accuracy: The sequences obtained from 70 plasma samples were compared to those previously generated by an independent laboratory (Janssen Diagnostics) who performed the testing for the QUEST clinical trials.  Assay precision: Repeatability (intra-assay variability) was tested in three samples: one commercially prepared HCV-positive plasma and two clinically-derived samples (one with Q80K/R, one without). Twelve replicates of each sample were PCR amplified and sequenced in a single batch. Reproducibility (inter-assay variability) was tested in 11 samples from the VIDUS cohort: four samples with the Q80K/R polymorphism; seven with wild-type Q80. All samples were tested on five different days by two laboratory technicians.  135  Assay sensitivity: Assay sensitivity, defined as the lowest HCV RNA concentration that could be amplified in a minimum of two-thirds of samples attempted, was determined using five patient samples from the Janssen Diagnostics sample set. Samples with HCV pVL ranging between 5.6 – 6.0 log10 IU/mL were serially diluted (1:10) with pooled normal human plasma to obtain concentrations in the ~2 log10 IU/mL range. HCV RNA was extracted from diluted plasma samples and six replicates of three samples per method were tested in a single batch.  Assay specificity: Specificity was assessed via three experiments. First, five HCV-negative samples (pooled normal human plasma, HBV-positive/HCV-negative plasma, two in-house HIV-positive/HCV-negative controls, and DEPC-treated water) were extracted, amplified in triplicate, and gel electrophoresis was run on the “amplified” products. Second, PCR cross-reactivity or interference by other viruses was evaluated in two HCV-positive samples spiked with either HIV-1 or HBV: HCV-positive plasma with pVL 6.6 log10 IU/mL was spiked with a clinically-derived HIV-positive plasma sample (pVL 3.9 log10 HIV RNA copies/mL; primary method), a HIV laboratory clone (PNL4-3, approx. pVL 3.9 log10 HIV RNA copies/mL; secondary method), or HBV-positive plasma (pVL unknown; primary and secondary methods). The HBV- and HIV-spiked plasma sample were amplified and sequenced in triplicate in a single batch. Finally, the ability to amplify products from non-GT1 was evaluated using a commercial panel of HCV subtypes (SeraCare). The panel included samples designated as GT1a, GT1b, GT2a/c, GT2b, GT3a, GT4, GT4a/c/d, GT5a, and GT6a/b; and one HCV-negative plasma sample. Panel samples were amplified in triplicate in a single batch. All validation experiments were performed using both the primary and secondary Sanger sequencing methods. Validation of the MiSeq method was limited to the accuracy, reproducibility and sensitivity experiments.  In this study we developed and characterized two HCV NS3 sequencing assays intended to screen for the Q80K polymorphism prior to the initiation of simeprevir-containing therapy. The Sanger 136  sequencing assay makes use of two independent PCR amplifications; a “primary” method that produces a 1200-bp sequence and a “secondary” method that produces a 564-bp sequence to be used as a backup in the event that amplification with the primary primers fails. For the purpose of this validation, primary and secondary methods were considered independent tests and all validation experiments were performed using both methods. In addition, we developed an independent next-generation sequencing assay involving near-full-genome amplification of HCV GT1 followed by Nextera XT library preparation (Illumina).   Janssen Diagnostics provided 70 frozen plasma samples from treatment-naïve HCV GT1-infected participants of the simeprevir licensing trials (86% GT1a, 14% GT1b). The samples had a median pVL 6.7 log10 IU/mL (Q1-Q3: 6.1-6.9 log10 IU/mL). Nucleotide sequences determined by Janssen – either a 2055-bp sequence covering the entire NS3 and NS4a genes, or a 543-bp fragment covering the first 181 codons of NS3 – served as the comparator. According to the Janssen assay 30 of 70 (42.9%) samples contained the Q80K alone or as part of a mixture. Using the primary and secondary methods described here, all 30 Q80K polymorphisms were identified in HCV GT1a samples.  Amplification and sequencing by the primary 1200-bp assay was successful in 66 (94.3%) samples. When compared to the Janssen sequences we observed 98.8%, 99.6%, and 100% overall concordance in nucleotide sequence, amino acid (AA) sequence and Q80K calls, respectively (Appendix XI-A). When sequences were compared individually, a median 99.3% (Q1-Q3: 98.2-99.8%) nucleotide concordance was observed between sequence pairs. The vast majority of discrepancies (97.8% of nucleotide, 98.9% of AA differences) were due to differences in mixture calls (Figure 7.1), with the BCCfE method calling a marginally higher number of mixed bases overall; a total of 1.5% and 1.3% of all bases were called as mixtures by the BCCfE primary and Janssen methods, respectively. Despite the small difference in the number of mixed bases called, we observed no systematic bias towards over- or under-calling mixtures by either the BCCfE or Janssen assay (Figure 7.2).  137   Three versions of the BCCfE assay were investigated: A “primary” Sanger sequencing method which produces a 1200-bp sequence; a “secondary” Sanger sequencing method which produces a 564-bp sequence (intended to be used as a backup in cases where the primary method fails to amplify a product); and a next-generation sequencing assay in which near-whole-genome HCV amplicons are generated and sequenced on a MiSeq after Nextera XT library preparation. For the MiSeq assay consensus sequences were generated, where all nucleotides observed at >20% frequency at each position were called as mixtures. Sequences obtained by BCCfE methods were compared to the Sanger sequences obtained by Janssen. All assays achieved ≥98.8% concordance in nucleotide base calls. Nearly all discordant base calls (>97.5% Sanger, 91.7% MiSeq) were due to differences in mixture calling between laboratories. All instances of the Q80K polymorphism (N=30) detected by the Janssen method were also detected by all three BCCfE methods. No false positive Q80K calls were reported by any BCCfE method. 138   A portion of HCV NS3 was successfully amplified by at least one of the “primary” or “secondary” Sanger sequencing methods in all 70 samples previously sequenced by Janssen BVBA. To demonstrate that neither method systematically missed minority HCV variants, the percentage of ambiguous nucleotides (“mixed” bases) in the longest fragment sequenced by both laboratories (543, 564 or 1200-bp) was calculated and compared. A modest, but not perfect correlation (panel A), consistent with variation in RNA extraction, RT-PCR amplification and base calling, was observed between sequences collected in the two laboratories. However, inspection of the Bland-Altman plot (panel B) revealed no systematic over- or under-calling of mixtures by either laboratory. Solid red and dashed green lines in panel B indicate the mean and mean ± 1.96SD of the difference in percentage of mixed bases (Janssen-BCCfE), respectively. 139  In total, 68 (97.1%) of samples were successfully amplified and sequenced by the BCCfE secondary method. When compared to the sequences provided by Janssen we observed 98.8%, and 99.5%, and 100% concordance in nucleotide sequence, AA sequence and Q80K calls, respectively (Appendix XI-B). Median nucleotide concordance between sequence pairs was 99.1% (Q1-Q3: 97.8-100%). As in the primary assay, almost all sequence discrepancies were due to differences in mixture calls (98.5% of nucleotide; 100% of AA differences) (Figure 7.1). In one sample the BCCfE secondary method identified a Q80Q/K mixture whereas the Janssen method observed a Q80K; however, this partial AA difference would have no impact on a resistance interpretation as both methods identified the Q80K. For completeness, the sequencing results from the primary and secondary BCCfE amplification methods were compared to each other. In total, 64 (91.4%) of samples gave results by both methods. Overall, we observed 99.0%, 99.7% and 100% concordance in nucleotide sequence, amino acid sequence and Q80K calls, respectively (data not shown). Finally, as the two BCCfE amplicons are intended to serve as “primary” and “backup” methods in a single resistance testing protocol, it should be noted that all 70 samples were successfully sequenced by at least one of the two methods.   Near-full-genome HCV was amplified from frozen plasma samples provided by Janssen Diagnostics. After excluding samples with <1000-fold coverage at NS3 codon 80 following demultiplexing, quality control and iterative mapping, consensus NS3 sequences were successfully obtained for 67 (96%) samples. Median coverage at codon 80 was 8,800 reads/sample (Q1-Q3: 6,500-11,000 reads/sample) and was fairly consistent across the length of NS3. While not the subject of the current study, it should be noted that sequencing coverage across the entire HCV genome was consistently high with the exception of the Core protein and a portion at the N-terminus of E2 (Appendix XII). When assembled consensus NS3 sequences were compared to Sanger sequences obtained by Janssen, 98.8%, 99.6%, and 100% overall concordance in nucleotide, amino acid, and Q80K calls, respectively, were observed 140  (Appendix XI-C). Median pairwise nucleotide concordance was 99.3% (Q1-Q3: 98.1-99.8%) between the Janssen Sanger and BCCfE MiSeq consensus sequences, with 91.7% of observed nucleotide differences being the result of differences in mixture calls between methods (Figure 7.1). When lower-frequency variants were examined, the MiSeq method did not detect any additional samples with minority variants harboring the Q80K polymorphism with frequencies between 2-20%; however, a Q80R resistance variant was observed in a single sample at a prevalence of 3.3%, though the clinical significance of such a low-frequency variant not known.  The ability to produce repeatable results across multiple tests was examined by performing replicate testing of three representative samples per method in a single run. Twelve replicate reverse transcription, PCR amplification and sequencing reactions were performed per sample starting from a single pool of extracted RNA. Intra-assay variability was extremely small, with >99.8% mean concordance observed between replicates when compared to a per-sample consensus (Table 7.1). Only one replicate of a single sample failed to produce a sequence. Assay Sample Success Rate (%) Precision (Mean ± SD) Q80 AA Q80 Concordance Primary 51311A 12/12 (100%) 99.9 ± 0.1% Q 100% 51216A 11/12 (92%) 99.9 ± 0.1% Q 100% 51226A 12/12 (100%) 99.9 ± 0.1% K 100% Secondary 51411A 12/12 (100%) 100 ± 0% Q 100% 51417A 12/12 (100%) 99.8 ± 0.2% Q 100% 51418A 12/12 (100%) 100 ± 0% K/R 100%  Assay repeatability was assessed using a panel of three samples per method which were tested 12 times in a single batch. High concordance (precision) in nucleotide base calls was observed when individual sequences were compared to a per-sample consensus. The “primary” method is a 1200-bp fragment spanning codons 1-400 of HCV NS3. The 564-bp “secondary” method is intended for use in the event of “primary” assay failure. SD: standard deviation; AA: amino acid 141   Assay reproducibility was assessed using a panel of 11 clinically-derived plasma samples for the primary method (four samples with the Q80K polymorphism; seven with wild-type Q80) and 10 plasma samples for the secondary assay (four samples with Q80K; six wild-type). All samples were tested on five different days by two laboratory technicians. An extremely low level of inter-assay variability was observed across all sequenced bases in NS3; over 99.7% concordance was observed between replicates in all samples tested (Table 7.2). No difference in Q80K interpretations were observed between replicates. All five replicates were successfully sequenced in 20 (95.2%) tested samples. The remaining sample was successfully sequenced in only 2 of 5 attempts using the 1200-bp primary method; however, identical nucleotide sequences were obtained in each replicate. Reproducibility of the MiSeq sequencing assay was assessed in a similar manner. RT-PCR and Nextera XT library preparation were attempted on 14 samples on five consecutive days. Successfully amplified libraries were sequenced on two separate MiSeq runs. Two samples failed the full-genome MiSeq assay and were excluded from the analysis of reproducibility: One sample failed to amplify in all five replicates. A PCR product was obtained for the second sample in 4 of 5 replicates; however after sequencing and assembly <1000-fold coverage was obtained in the region surrounding NS3 codon 80 in all replicates. In the remaining 12 samples, no substantial differences were observed in the frequency of the Q80K polymorphism in all recovered reads across the replicates (Figure 7.3). Although coverage <1000 reads at NS3 codon 80 was obtained in one replicate of a single sample, no effect on the detection of the Q80K polymorphism was observed.  142  Assay Sample Success Rate (%) Precision (Mean ± SD) Q80 AA Q80 Concordance Primary 51204A 5/5 (100%) 100 ± 0% K 100% 51206A 5/5 (100%) 99.9 ± 0.1% Q 100% 51208A 5/5 (100%) 99.9 ± 0.1% Q 100% 51210A 5/5 (100%) 99.9 ± 0.1% Q 100% 51212A 5/5 (100%) 99.7 ± 0.1% Q 100% 51218A 5/5 (100%) 99.7 ± 0.2% K/M 100% 51220A 5/5 (100%) 99.9 ± 0.2% K 100% 51222A 5/5 (100%) 100 ± 0% K 100% 51224A 5/5 (100%) 99.9 ± 0.1% Q 100% 51228A 2/5 (40%) 100 ± 0% Q 100% 51230A 5/5 (100%) 100 ± 0% Q 100% Secondary 51411A 5/5 (100%) 99.9 ± 0.1% Q 100% 51412A 5/5 (100%) 100 ± 0% Q 100% 51413A 5/5 (100%) 99.7 ± 0.3% Q 100% 51414A 5/5 (100%) 100 ± 0% K 100% 51415A 5/5 (100%) 100 ± 0% K 100% 51416A 5/5 (100%) 99.9 ± 0.1% K 100% 51446A 5/5 (100%) 100 ± 0% Q 100% 51447A 5/5 (100%) 99.8 ± 0.2% Q 100% 51448A 5/5 (100%) 100 ± 0% Q 100% 51450A 5/5 (100%) 99.9 ± 0.1% K 100%  Assay reproducibility was assessed using a panel of plasma samples which were tested on five separate days by two laboratory technicians. High concordance (precision) was observed when individual sequences were compared to a per-sample consensus.  143   The MiSeq amplification protocol was attempted on 14 samples on five consecutive days. HCV NS3 was successfully amplified in all five replicates in 12 samples (85.7%). The proportion of reads in which wild-type Q80, Q80K, Q80M or other Q80 variants was observed (y-axis, log10 scale) is displayed for each successfully sequenced replicate (x-axis). After read merging, mapping and quality control the proportion of reads exhibiting a Q80K polymorphism was consistent across replicates. No minority (<20% prevalence) variants carrying the Q80K polymorphism were detected in any replicates. Dotted black line indicates the level of sequencing coverage at NS3 codon 80 in each replicate.  144   The lower limit of detection, was determined using serial dilutions (1:10) of five plasma samples (pVL range 5.6 – 6.0 log10 IU/mL). Three samples were tested for each Sanger sequencing method with six replicates performed at each pVL dilution. The lower limit of detection was defined in two ways: 1) PCR amplification success rate among replicates, 2) nucleotide sequence concordance between sequences obtained in the full-strength sample and most dilute sample for which sequencing was successful. For simplicity, one replicate sequence from the highest and lowest concentration for each sample was selected at random for this comparison. Using the primary Sanger method there was 100% and 94% success in amplification and sequencing in samples with >5.0 and 4-5 log10 IU/mL, respectively (Table 7.3A). Amplification and sequencing success rate was <45% in samples with pVL 3-4 log10 IU/mL; however 4 of 6 samples with pVL 7,280 IU/mL were successfully amplified suggesting a limit of detection on the order of 7,000 IU/mL. Nucleotide concordance rates of 98.3-99.8% were observed when sequences from the lowest and highest concentration were compared. No differences were observed at NS3 Q80 in any samples. For the secondary Sanger method, 100% success in amplification and sequencing was observed in samples with >3.0 log10 IU/mL (Table 7.3B). In samples with pVL 2-3 log10 IU/mL the amplification and sequencing success was an acceptable 67%, suggesting a limit of detection of approximately 100 IU/mL or possibly lower. Nucleotide concordance rates of 98.6-99.1% were observed when sequences from the lowest and highest concentration were compared. No differences were observed at NS3 Q80 between concentrations. 145  Sample 1 Sample 2 Sample 3 Summary pVL  (IU/mL) Success Rate (%) pVL  (IU/mL) Success Rate (%) pVL  (IU/mL) Success Rate (%) pVL Range (log10 IU/mL) Success/Reps (%) 1,080,000 6/6 (100%) - - - - >6.0 6/6 (100%) 108,000 6/6 (100%) 728,000 6/6 (100%) 395,000 6/6 (100%) 5 - 6 18/18 (100%) 10,800 5/6 (83%) 72,800 6/6 (100%) 39,500 6/6 (100%) 4 - 5 17/18 (94%) 1,080 2/6 (33%) 7,280 4/6 (67%) 3,950 2/6 (33%) 3 - 4 8/18 (44%) - - 728 1/6 (17%) 395 1/6 (17%) 2 - 3 2/12 (17%)  Sample 1 Sample 2 Sample 3 Summary pVL  (IU/mL) Success Rate (%) pVL  (IU/mL) Success Rate (%) pVL  (IU/mL) Success Rate (%) pVL Range (log10 IU/mL) Success/Reps (%) 1,040,000 6/6 (100%) - - - - >6.0 6/6 (100%) 104,000 6/6 (100%) 695,000 6/6 (100%) 395,000 6/6 (100%) 5 - 6 18/18 (100%) 10,400 6/6 (100%) 69,500 6/6 (100%) 39,500 6/6 (100%) 4 - 5 18/18 (100%) 1,040 6/6 (100%) 6,950 6/6 (100%) 3,950 6/6 (100%) 3 - 4 18/18 (100%) 104 4/6 (67%) 695 4/6 (67%) 395 4/6 (67%) 2 - 3 12/18 (67%)  Estimating the lower limit of detection (LOD) of HCV NS3 Q80K screening assays through serial 1:10 dilution of clinically-derived plasma samples. PCR amplification success rates suggest that the LOD is >3.9 log10 IU/mL for the primary assay (Table A) and >2.0 log10 IU/mL for the secondary assay (Table B). 146   Five HCV-negative samples were tested to demonstrate that no off-target amplicons would be generated. Samples were tested in triplicate and the results were visualized on a 1% agarose gel. As expected, no bands were observed for the negative sample sets for either Sanger or MiSeq assays (data not shown). As no bands were observed, HCV-negative samples were not sequenced by either method.  HCV-positive plasma was spiked with a clinically derived HIV-1 plasma sample, a HIV-1 molecular clone, or a clinically-derived HBV-positive plasma sample and subsequently processed in triplicate by the two Sanger sequencing methods. As expected, nucleotide sequences and NS3 Q80 calls were nearly identical (>99.9% concordance) between HIV/HBV-spiked and un-spiked samples, suggesting no interference by potential coinfecting viruses. Similarly, amplification and sequencing of HCV NS3 were not affected by spiked-in HIV or HBV in the MiSeq assay (data not shown).  The ability to amplify products from non-GT1 HCV was evaluated using a commercial panel of nine plasma samples containing HCV of various genotypes: GT1a, GT1b, GT2a/c, GT2b, GT3a, GT4, GT4a/c/d, GT5a, and GT6a/b. Panel samples were amplified in triplicate by both Sanger sequencing methods and the resulting PCR amplicons were visualized on a 1% agarose gel. All replicates of HCV GT1a and GT1b samples were successfully amplified by both methods; however, non-GT1 HCV samples could not be amplified by either method (data not shown).   We have developed and characterized the performance of two independent HCV GT1 sequencing assays for the detection of the NS3 Q80K polymorphism that can partially compromise the efficacy of 147  some antiviral regimens containing simeprevir. Previous studies have examined the prevalence of the Q80K polymorphism in treatment-naïve and treatment-experienced populations, and have demonstrated a substantial disparity in the prevalence of Q80K in GT1-infected populations in North America versus Europe [467,470,474–477]. We have demonstrated that both Sanger sequencing of a 1200-bp and/or 564-bp fragment of NS3 and next-generation sequencing (MiSeq) of a near-full-genome product can be used to accurately screen for the NS3 Q80K polymorphism. The repeatability and reproducibility of all assays was extremely high, with >99.7% nucleotide concordance observed in all replicate tests. While a subset of samples was effectively clonal across the region sequenced, most samples exhibited a substantial amount of variability as measured by the number of nucleotide mixtures called. The high level of inter- and intra-assay concordance between replicates was therefore not due to a lack of variation within the samples tested. Our sensitivity experiments determined that the lower limit of detection is sufficiently low to allow testing in chronically HCV-infected, treatment-naïve patients. Samples with pVL >3.9 log10 IU/mL were amplified in at least two-thirds of replicates using the primary Sanger method and were sequenced with repeatable results . The secondary method appears to be capable of amplifying samples with pVL >2.0 log10 IU/mL, enabling it to be useful as a backup or “rescue” assay. The lower pVL limit of the MiSeq assay was considerably higher (~5 log10 IU/mL), consistent with the requirement to amplify a >9-kb fragment (data not shown). Nevertheless, the lower pVL limit of either assay is below the expected viral load range of most patients being considered for simeprevir therapy [214,215]. Note that a nested PCR strategy was chosen in order to maximize the future utility of this assay as a resistance test in DAA-treated patients in all stages of treatment, including those with very low viral copy number. Finally, none of the assays appear to be affected by potential coinfecting viruses (HIV, HBV). The Sanger sequencing methods are presently in clinical use at the BC Centre for Excellence in HIV/AIDS and have been made available to interested laboratories worldwide. The RECall analysis software is freely available as a web application (http://pssm.cfenet.ubc.ca/) [402]. 148  Other HCV NS3 Q80K screening assays are available as laboratory services via commercial vendors in the USA (e.g. LabCorp, Quest Diagnostics); however, limited details about assay methods and performance metrics are currently available. Like the methods presented here, these commercially-available tests are validated only for HCV GT1. To our knowledge this study represents the first detailed description of the clinical validation of a HCV NS3 sequencing assay to be used for Q80K screening. It should also be noted that the methods presented here could potentially be used to screen for all major NS3 protease inhibitor resistance mutations known to date, though additional validation studies would be required.  While these methods have produced consistent and accurate results, they are not without limitations. As simeprevir is licensed in Canada for the treatment of only HCV GT1 infection, and the Q80K polymorphism is typically only observed in HCV GT1a, the procedures outlined in this study have been optimized for HCV GT1. Increased amplification and sequencing failure rates may be observed for other HCV genotypes (GT2-7) as a consequence of assay design. In fact, both Sanger sequencing methods failed to amplify HCV NS3 when tested against a panel of non-GT1 samples. It should be noted, however, that the samples in the genotype panel were diluted by the manufacturer to viral loads in the 3-4 log10 IU/mL range. While this is above the limit of detection of the secondary 564-bp method when used for GT1, the performance characteristics may be such that a higher pVL is required to successfully amplify non-GT1 samples. For example, the BCCfE laboratory has subsequently used these protocols to successfully amplify and sequence NS3 from high pVL HCV GT3 samples previously misclassified as GT1 by LiPA (data not shown). If necessary, these procedures could potentially be validated for use in other HCV genotypes; however, this is beyond the scope of the study presented here. Finally, the RT-PCR process is dependent upon an adequate recovery of viral RNA during extraction, as well as proper binding of the primers during amplification. Minority HCV populations carrying the Q80K polymorphism may be missed by either population sequencing method as traditional Sanger sequencing can only detect minority variants (nucleotide mixtures) that exist at a minimum of ~20% 149  of the population. The MiSeq assay could potentially be used in cases where the detection of minority resistance variants is important, though the assay’s performance in that context has not been fully evaluated. However, given the absence of minority variants carrying the Q80K polymorphism detected in this and previous studies [476,477] and that population prevalence of Q80K is likely due to a founder effect rather than active selection [471], deep sequencing specifically for NS3 Q80K may be unnecessary. In summary, we have developed and validated two independent assays on two sequencing platforms to screen for the HCV NS3 Q80K polymorphism. We have demonstrated excellent accuracy, repeatability, reproducibility, and sensitivity of these: all expected Q80K polymorphisms were detected by all methods with no false positive results in samples with pVL at or below levels expected in patients initiating simeprevir-containing regimens. Since these methods cover the amplification of a sufficiently large fragment of HCV NS3 the potential exists for either of these assays to be used in drug resistance testing for all currently approved NS3 protease inhibitors. 150    This thesis described a number of applications of viral genetic sequencing for use in a clinical setting. Both traditional Sanger and multiple “next-generation” sequencing platforms were used to develop novel clinical protocols, evaluate existing clinical tests, and to investigate hypotheses about HIV selection and evolution. Four major aims as outlined in section 1.4 were addressed and provided support for the ongoing and expanding use of viral sequencing in the clinic. Specifically: 1. Inherently variable laboratory tasks, such as calling mixed bases in sequences from diverse virus populations, can be standardized with automated software. This ensures consistently high-quality results and improved process efficiency. 2. The lack of selection of drug resistance mutations in patients with supposed low-level viremia provided supporting evidence that a substantial proportion of detectable plasma viral load measurements by the Roche TaqMan v1 HIV-1 Test were in fact false positive results. 3. Sequencing continues to be the most direct method to measure viral evolution or selection in vivo and in vitro; residual viremia is the main driver of continued HIV evolution in response to host- or drug selective pressures in antiretroviral-treated patients.  4. Viral sequencing will continue to play a major role in infectious disease monitoring and clinical management as new treatment options for additional pathogens continue to be developed. While new sequencing platforms will certainly be introduced, new uses for existing technologies remain to be explored. 151   As a primary focus of this thesis was the development of novel sequencing assays and analysis tools for use in a clinical setting, many of the studies presented here have had direct and immediate applications as clinical protocols. Specifically, the findings in Chapters 2, 5, and 7 have since been translated into routine practice, and/or have been used as laboratory methods in follow-up studies. The methods described in Chapter 6 have implications for future drug resistance surveillance strategies for low and middle-income countries. The remaining chapters generated additional hypotheses. In addition to contributing evidence to support the modification of existing HIV treatment guidelines, the results of Chapter 3 have prompted further research into the clinical significance of low-level HIV viremia. The findings presented in Chapter 4 may inform clinical trial design of therapeutic HIV vaccines or strategies for HIV eradication or cure.  Chapter 2 presented the technical validation of RECall, software used to standardize chromatogram interpretation when sequencing diverse viral populations. As noted in section 2.4, while the validation was limited to HIV-1 PR and RT sequence data generated by an external laboratory, the software can be applied to most any coding region in a pathogen or even host genome. RECall was therefore used as the default analysis software for all Sanger sequencing data in the subsequent studies in this thesis. In fact, RECall is now routinely used to process HIV and HCV sequence data for drug resistance testing, HIV tropism testing, and HLA-B*57:01 screening at the BC Centre for Excellence in HIV/AIDS. Furthermore, the standalone version of RECall is being used in several clinical laboratories, including those of Quest Diagnostics and the US Centers for Disease Control and Prevention, while the web application is used by over 100 registered groups at research and clinical institutions in Canada and around the world [PR Harrigan, personal communication, October 30, 2012].  152  While RECall significantly reduces analysis times by eliminating the mindless drudgery of manual sequence editing, more importantly the software standardizes a highly subjective data interpretation step. Standardization not only removes a potential source of variation in clinical results reporting, it also allows bias-free head-to-head comparisons of results obtained after a significant process change. These could include, for example, validation of new lots of reagents, or changes to standard operating procedures. It also simplifies the interpretation of discrepant sequences determined by different laboratories as part of external quality assessment (EQA) programs. For these reasons, World Health Organization-accredited resistance testing laboratories are required to use RECall as part of their protocols. Given the expansion of next-generation sequencing use in HIV and HCV drug resistance monitoring, an analogous platform to analyze these types of data is urgently needed.  Groups developing novel next-generation viral genetic assays must currently develop both their laboratory methods and data analysis pipelines in parallel [344,346,478–480] as most existing variant calling software was designed with low-coverage sequencing of diploid organisms in mind, rather than deep sequencing of viral quasispecies [447,481–488]. Specifically, for most human re-sequencing data true variants are expected to be observed at discrete frequencies (i.e. 0%, 50%, or 100% of the recovered reads), while a continuum of variant frequencies is possible for pathogen populations or somatic mutations in tumor cells. The problem of distinguishing true low-frequency variants from the noise of PCR and sequencing errors is thus considerably greater for viral deep sequencing [489]. Furthermore, applying different analysis methods to a single dataset can sometimes lead to substantially different results. In human genetic studies where multiple pipelines are used to simultaneously process the same data, poor concordance in single nucleotide variant calls are often seen [490–495]. For example, in a direct comparison of five variant calling methods only 57.4% of the variants called were identified by all methods. When the analysis was restricted to novel variants (i.e. those not found in dbSNP v135) concordance dropped to 11.4%, suggesting an extremely high false positive rate for rare variants [490]. This problem is compounded when differing sample preparation methods are also used [493,494]. In an EQA or proficiency testing setting of next-generation HIV sequencing, standardized software would allow users to interpret discrepant results by disentangling 153  the data generation and data analysis steps. With this in mind, several groups have developed HIV variant-calling tools for next-generation sequence data, but rigorous characterization of their performance in clinical settings has yet to be performed [479,496–498]. Additional software tools aimed at clinical HIV resistance testing have been recently introduced either as cloud-based fee-per-use services [499] or as proprietary software forming part of a larger laboratory-developed test [478]. However, open-source or open-access solutions would be preferred.  Chapter 3 described a systematic deficiency in an approved clinical test. The substantial confusion among clinicians as to the significance of seemingly aberrant test results raises questions about the quantity and quality of evidence required when approving the replacement of a de facto gold-standard diagnostic. The US FDA regulates medical devices through one of two processes: Devices that are regulated through the review of a premarket notification (also known as the 510(k) pathway) are “cleared” by the FDA, while the review of a premarket approval application is required for a device to be FDA “approved”. Currently, a replacement for an existing FDA-cleared test is required only to demonstrate “substantial equivalence” to its predecessor if the replacement has the same indicated use and technological characteristics [500,501]. While the TaqMan v1 assay underwent the more rigorous FDA premarket approval process, including technical validation of the test’s reproducibility, repeatability, linearity, precision and limit of detection, the comparison of its performance relative to existing HIV viral load assays was extremely limited in scope [387]. Specifically, the FDA approval summary and the TaqMan assay product insert indicate than an analysis of the correlation of TaqMan viral load results to those reported by three previously approved tests (COBAS AmpliPrep/COBAS AMPLICOR HIV-1 Test v1.5, COBAS AMPLICOR HIV-1 MONITOR Test v1.5, and VERSANT HIV-1 RNA 3.0 Assay) was limited to only 71 clinically-derived samples. These samples had viral loads that spanned the entire reportable range of the TaqMan assay. Of critical importance, no particular attention was paid to the range of values where important clinical decisions are made. In fact only two of the 71 samples tested had HIV-1 pVL <100 RNA copies/mL by Amplicor. Furthermore, samples that 154  did not give results within the reportable range for both assays compared were systematically ignored, thus potentially obscuring the observations of false-positive test results presented in Chapter 3 [387].  Ultimately, the study presented in Chapter 3, together with previous observations [370], contributed to a revision of the British Columbia HIV/AIDS Therapeutic Guidelines and a modification of the definition of treatment failure [384]. The results, however, do not unequivocally identify the new TaqMan test as the sole source of these discrepancies as a systematic comparison of the Amplicor and TaqMan v1 assays was not performed across the entire reportable ranges of these tests. It is unknown whether similar issues exist at higher viral load strata, or alternatively, if Amplicor systematically underreported viral loads at or around its limit of detection [502]. A randomized clinical trial evaluating HIV therapy outcome where physicians’ treatment decisions are informed either by TaqMan or Amplicor viral load results could establish equivalency (or superiority) of these tests. Unfortunately, given the complexity of such experiments and the fact that the manufacture of the Amplicor assay has been discontinued a direct comparison of the two assays is impossible. Retrospective, observational data from large cohort studies would instead be needed to gain additional insight into the clinical significance of low-level viremia reported by each approved HIV viral load assay.  Establishing the clinical relevance of low level viremia during otherwise suppressive antiretroviral therapy is an active field of research [503]. Studies have consistently demonstrated that persistent low-level HIV viremia is associated with an increased risk of subsequent virological failure, regardless of the viral load assay used [379,385,504–506]; however, the predictive value of transient viral load “blips” is less clear [379,383,507–510]. Furthermore, in patients with low-level viremia, the presence of antiretroviral-resistant variants is well documented. The detection of drug-resistant variants in patients with low-level viremia is predictive of virological failure in the short-term [444,511,512] and the selection of additional resistance mutations is observed in up to 40% of patients with detectable pVL under 1000 copies/mL [444,511–515]. However, despite substantial observational evidence that low-level viremia has a measurable clinical impact, there is a significant dearth of clinical trial data to 155  help inform how low-level viremia should be managed. HIV treatment guidelines are therefore based on observational data and expert opinion and are subject to the available resources in each jurisdiction [503]. Indeed, the BC HIV/AIDS Therapeutic Guidelines were again modified in 2013, recommending further evaluation of patients with sustained HIV RNA levels between 40 and 200 copies/mL including drug resistance genotyping and more frequent viral load testing [191]. The findings presented in Chapter 3, that low-level plasma viral load measurements are often false positive results, may complicate such evaluation. In the absence of an accurate test, methods to distinguish between true cases of intermittently detectable viremia and assay false positives are required. One solution could be the performance of the drug resisting genotyping test itself. A recent study of the predictive merits of drug resistance testing in patients with low viral loads noted that patients for whom genotypic drug resistance testing failed to give a result were at lower risk of virologic failure than both patients with and without detectable drug resistance [512]. As the lower limit of detection of the updated TaqMan v2 test remains poorly calibrated with that of Amplicor [516], the implications and clinical management of low-level viremia will remain an important research topic.  Chapter 4 examined the possibility of continued HIV escape from immune pressures after the initiation of HAART. It concluded that HIV evolution after therapy initiation was dominated largely by the acquisition of new drug resistance mutations rather than CTL escape. Furthermore, the reversion of existing escape mutants was rarely observed. As expected, the primary driver of continued HIV evolution was the incomplete suppression of viral replication. This correlation of HIV evolution with residual viremia further supports the conclusions of Chapter 3 that limited selection of resistant variants in individuals with intermittent low-level viremia increases the likelihood that those detectable viral load results are false positives. Furthermore, these findings may have implications for the design of future therapeutic HIV vaccine trials, or anticipated trials of curative strategies.  156  The goal of therapeutic HIV vaccination is not necessarily the eradication or functional cure of HIV infection. Rather, therapeutic vaccination may help control infection and delay the progression of the disease by stimulating cellular or humoral immune responses to identify and kill activated HIV-infected cells [517]. Therapeutic vaccination is unlikely to replace antiretroviral therapy, but could potentially play a complementary role in suppressing viremia to undetectable levels [517]; however, therapeutic vaccination has been proposed as a component of viral eradication strategies aimed at eliminating the pool of latently infected cells. In the “shock-and-kill” strategy an agent, such as a histone deacetylase inhibitor, would be administered to induce HIV expression in latently infected cells. Reactivated cells would be identified and removed by host immune responses stimulated by therapeutic immunization. Simultaneously, antiretroviral therapy would prevent infection of new target cells [518,519]. To date, however, few clinical trials have demonstrated the efficacy of either preventative [520–522] or therapeutic [523,524] vaccination strategies.  Follow-up studies have identified a number of immune correlates of vaccine efficacy. Of note, “sieve” analyses of breakthrough viruses (the HIV-1 viruses that evade immune responses to establish infection [525]) have shown that sequences from vaccine recipients have larger genetic distances from the vaccine insert [526], or were more likely to contain specific residues at certain positions in env [527] than those from placebo recipients. Sequence analyses of viruses emerging during analytical treatment interruptions in trial of a therapeutic vaccine containing a HIV-1 gag insert [523] have yielded similar results. In the vaccine arm of the AIDS Clinical Trials Group A5197 study, greater gag sequence divergence from the vaccine insert and a lower proportion of HLA-associated sequence polymorphisms was observed in individuals with lower viral loads during the treatment interruption [528,529]. In all of the above studies, the viral “sieving” effects were observed only in HIV regions included in the vaccine inserts. Taken together this sequence-specific effect illustrates that, despite overall poor efficacy at preventing HIV infection or promoting virologic control, these vaccines may have induced protective responses against a narrow set of viruses containing specific sequence motifs. Therefore, future clinical trials of therapeutic vaccines might consider pre-screening patients by sequencing. However, as only patients with consistently undetectable plasma viral loads and high CD4 157  counts are eligible for most therapeutic vaccine trials, sequencing from plasma samples drawn at the time of eligibility screening would be difficult. The findings presented in Chapter 4 therefore support the use of pre-therapy plasma samples should sequence pre-screening be adopted in trial design. Such a strategy is already used in other clinical HIV applications. For example, as HIV co-receptor tropism is rarely observed to change while on suppressive HAART [530,531], HIV tropism testing is often performed using pre-therapy plasma samples for antiretroviral-experienced patients with suppressed viral loads who are considering switching to a regimen containing maraviroc [532–534].   Chapter 5 presented a method to simultaneously assess the replicative capacity of multiple HIV variants using competitive co-culture and by deep sequencing. The pools of virus variants generated as part of this study may have additional applications beyond the study of viral fitness. These materials may be useful in future technical validation and performance studies of clinical testing involving next-generation sequencing. Accepted criteria already exist for the development and validation of clinical assays using Sanger sequencing [500,535–537]. Generally, demonstration of the consistency of generating accurate population sequences is sufficient for validation, with replicate sequencing, either in-house or as part of an inter-lab quality assurance protocol, serving as the gold standard. While guidelines for clinical next-generation sequencing in humans are under development [538–541], no consensus currently exists on the performance thresholds of various deep sequencing platforms for the detection of minority viral variants in a population, or the sample types required to perform these assessments. Clonal laboratory isolates are often used to measure the raw error rates of the extraction, RT-PCR and sequencing steps of a deep sequencing assay [478,542–545]. However, the error rate estimates generated do not address the accuracy and repeatability of quantification of minority variants present in clinical samples [478,542–545]. As such, these studies typically only consider minority variants to be genuinely present if they are observed at a considerably higher frequency (typically 5 to 10-fold) than the technical error rate. Conversely, estimating process error rates using clinical samples often relies on comparing recovered reads to a sample-specific reference in the form 158  of a consensus sequence [343] or a set of sequences generated by Sanger sequencing following limiting dilution of PCR products [345]. Either of these comparisons may overestimate error rates by miscategorizing true minority variants as sequencing errors. Finally, the use of either clones or clinical samples limits the evaluation of error rate to substitution and insertion/deletion errors, but not PCR-induced recombination [546].  The materials presented in Chapter 5, however, offer an alternative platform to validate novel next-generation sequencing-based laboratory developed tests. By using a mixture of 16 different viral clones, the infectious virus pool and the culture supernatants offer the opportunity to simultaneously measure the technical error rate of the RT-PCR and sequencing steps (using the clonal regions), the limit of detection of known minority variants (the 15 individual drug resistance mutations) and the frequency of PCR recombination (as the resistance mutations exist as single mutations in a single viral backbone). Indeed, these materials were subsequently used to perform a head-to-head comparison of error rates and detection threshold of three deep sequencing platforms: 454 GS-Junior, Illumina MiSeq, Pacific Biosciences SMRT [547,548]. The three instruments gave comparable estimates of mutant prevalence, but had substantially different error profiles. As expected, 454 had the lowest substitution error rate overall, but high numbers of insertion/deletion errors were observed in homopolymeric regions. MiSeq substitution error rates were approximately 3-fold higher compared to those of 454; however, these errors were more randomly distributed and few insertion/deletion errors were observed. The highest error rate was observed with the Pacific Biosciences instrument: 9-fold higher than that of the 454. It was also noted that sequencing results by MiSeq were robust up to a ~20,000-fold dilution of the starting material. Therefore, sufficient starting material was available to initiate a next-generation sequencing EQA panel program through the Forum for Collaborative HIV Research. The first iteration of which was initiated in late 2014. It is important to note, however, that these cross-platform and cross-laboratory comparisons were performed starting from viral DNA and therefore could only assess the performance of the sequencing steps themselves. Neither the variability of the reverse transcription step, nor the effect of template oversampling could be evaluated (see section 8.3 for more details). 159   Chapter 6 presented a proof-of-principle study of a novel application of next-generation sequencing technology to HIV drug resistance testing. This study demonstrates that by using a large set of multiplexing indices, the massive data generating capacity of the Illumina MiSeq instrument can be leveraged to simultaneously test hundreds of samples in parallel. The method described is able to replicate the performance of existing Sanger-based protocols and its scalability may make it suitable for high-throughput individual-level drug resistance testing for resource-limited settings.  The scale-up of antiretroviral therapy delivery in resource-limited settings took a public health approach which de-emphasized the need for specialist physician management and routine laboratory monitoring [549]. It was made possible in part by the simplification of clinical decision making into a set of standardized first and second-line regimens which are initiated at set CD4 thresholds. Individual-level viral load and drug resistance testing are not required. Instead, standardized regimens are set based on dosing simplicity, cost and population-level monitoring of efficacy. The emergence of drug resistance is monitored at the population-level using sentinel surveys coordinated by the World Health Organization HIV Resistance Testing Network (WHO HIV ResNet) [454]. Conducted by a network of accredited laboratories [550], the annual surveys consist of drug resistance testing at baseline and at 12 months after therapy initiation for approximately 100 individuals per site [551]. Despite this effort, increases in the population-level prevalence of antiretroviral resistance, particularly to the NNRTI class, have been observed [286,287,552]. These observations emphasize the need for effective population-level resistance surveillance tools and argues for the investigation of simple, practical and cost-effective methods to conduct individual-level monitoring [553]. Given new ambitious treatment coverage targets [554] at a time when funding for WHO HIV ResNet and similar programs is declining [555,556], implementing these tests should be a priority. Implementing individual-level viral load monitoring and drug resistance testing in low and middle income countries is, however, complicated by logistical and technological barriers. For example, access 160  to sequencing technology is limited-to-absent in many parts of Asia, Latin America and Africa. In these settings, mutation-specific PCR assays or similar methods [557] may be alternative options despite their limited scope. In other settings, sample collection and preparation itself (e.g. blood fractionation by centrifugation) is often impossible due to lack of equipment. Transportation of specimens for testing in centralized laboratories is complicated by limited cold-chain infrastructure and restrictive policies on international shipments of biohazardous materials. For these reasons alternative sample collection formats, such as dried blood spots [449,558,559] or absorbent matrix materials [560], have been investigated. These products greatly simplify the logistics of sample collection and shipment and thus may prove useful for expanded testing in LMIC. Viral load and drug resistance testing have been demonstrated to be possible using these starting materials, with some limitations. Dried blood spot assays are less sensitive than methods starting from plasma due to a greatly reduced sample input volume [450]. This results in elevated lower limits of quantification for viral load assays (typically >1000 HIV-1 RNA copies/mL), and higher amplification failure rates in resistance tests [450]. Furthermore, interpretation of results is often complicated due to the amplification of proviral DNA and/or non-integrated cellular HIV RNA. This “contamination” can lead to artificially high viral load measurements, or substantial divergence from sequences collected from circulating plasma viruses [450,451].  Nevertheless, it may be worthwhile to investigate the sequencing method described in Chapter 6 beginning with HIV RNA isolated from dried blood spots in order simultaneously simplify sample collection and decrease sequencing costs. Alternatively, in settings with high HIV prevalence where viral load testing is already available, but access to DNA sequencing is limited, the “wide” sequencing approach may provide a complementary cost-saving paradigm. Laboratories performing viral load testing accession and process samples, and extract HIV RNA from plasma. For certain pVL tests, such as the Abbott RealTime HIV-1 assay, a surplus of recoverable HIV RNA is generated. In most cases, this excess is discarded once testing is complete. However, substantial cost savings and logistical simplification could be achieved if the pVL testing laboratory used this excess RNA to perform a simple, inexpensive RT-PCR reaction. This heat stable and non-biohazardous material could then be 161  shipped by regular post to a centralized laboratory where the remaining PCR and “wide” sequencing steps could be performed. Regardless of the approach chosen, additional work is required to expand the assay to cover all relevant drug resistance mutations affecting common first- and second-line regimens, and validate its performance for clinical use.  The course of thesis research saw the approval of the first direct-acting antiviral agents for HCV treatment and the subsequent rapid expansion of treatment options. Similar to HIV, the selection and transmission of drug resistance are of concern for some of these new drug classes. Genotypic drug resistance testing may therefore be required to identify causes of treatment failure, or to identify treatment options of maximal benefit to the individual patient. Chapter 7 presented the validation of one such test. While the methods described were designed as screening tests for the Q80K polymorphism that confers decreased susceptibility to certain combination regimens containing simeprevir, the coverage of HCV NS3 is sufficient for these assays to serve as resistance tests for all currently available protease inhibitors. Furthermore, the MiSeq method covers nearly the whole HCV genome, making it a potentially useful resistance assay for all current and future direct-acting antivirals. Additional clinical validation is of course required.  Beyond their direct clinical applications, the methods described in Chapter 7 may aid in answering outstanding research questions. For example, as new HCV drugs are approved, further evaluation of the prevalence of primary resistance in untreated populations may be required. As the pace of drug development has been extremely rapid, existing studies are largely limited in scope, either restricted to evaluating single proteins [466,467,475,476], or reporting resistance to investigational drugs that never reached the market [464,465]. Updated studies using more sensitive next-generation sequencing methods covering the whole HCV genome would be beneficial. The lesson of simeprevir and the geographic disparity in the prevalence of Q80K has taught us that these studies should be performed in multiple populations worldwide. Second, while sustained virologic response rates 162  exceeding 95% are now frequently observed in clinical trials of new direct-acting antivirals, a small proportion of patients still fail to respond or relapse with drug-resistant HCV. It will be important to investigate the long-term stability of resistance mutations by longitudinal deep sequencing of post-therapy samples. As HCV does not integrate into the host genome, reversion of selected resistance mutations to wild-type after cessation of therapy may enable re-treatment with the same regimen [213,561,562]. Previous studies using population sequencing suggest that up to 70% of resistant variants selected by first generation NS3 protease inhibitors revert to wild-type within three years [563,564]; however, minority variants may still be detected if more sensitive methods are used [561,565,566]. The potential of these variants to re-emerge upon re-treatment with the same drug, or even with the same class of drugs is largely unknown. The next-generation HCV sequencing assay presented in Chapter 7 could be used to measure the decay of the resistant viral population over time, or quantify the amount of resistant variants prior to re-treatment. With the exception of sofosbuvir, direct-acting antivirals currently approved for use in North America are only active against HCV GT1 (the NS5A inhibitor daclatasvir [567,568] is available in Europe [221], but FDA approval has been delayed). While sofosbuvir has pan-genotypic activity, it has FDA approval only for HCV GT1, GT2, GT3, and GT4. It is therefore essential to accurately determine HCV genotype prior to therapy initiation. In some populations, such as injection drug users, multiple infection with HCV of different genotypes is not uncommon [569–571]. Mixed genotype HCV infection can result in nonresponse to direct-acting antiviral treatment drugs without pan-genotypic activity are used [571]. In such cases, an apparent switch in HCV genotype upon failure can be misinterpreted as reinfection rather than the selection of a preexisting minority population [571]. The prevalence and clinical consequence of mixed HCV infection must be continue to be monitored as new antivirals are approved. Since standard genotyping assays are unable to diagnose mixed genotype infections, alternative approaches, such as deep sequencing of variable HCV regions after amplification with pan-genotypic primers [572], may be required.  163  Direct-acting antivirals are a potential cure for HCV infection. However, the high cost of these drugs, often exceeding $100,000 per treatment course per patient, limits their widespread use [573]. Until such a time that low-cost, all-oral, pan-genotypic, short-course regimens with a high barrier to resistance are available, genetic methods will continue to play a role in personalizing HCV therapy [574]. For example, HCV sequencing could assist in reducing “system waste” and maximize potential for individual patient success by identifying persons who would require more expensive therapy (e.g. in the case of mixed genotype infection), or those who could respond equally well to a lower-cost option (e.g. 8-week sofosbuvir/ledipasvir treatment for low viral load, HCV GT1 infection).  The work presented in this thesis is not without limitations. Globally, HIV and HCV exhibit extreme genetic diversity; however, much of this and previous work has focused on viral subtypes that are common in North America and Europe: HIV subtype B, and HCV GT1. The primary burden of HIV infection worldwide is borne by Sub-Saharan Africa where non-B subtypes predominate. The studies described here may be applicable to these settings, but modifications to laboratory protocols, or data analysis pipelines may be required prior to implementation [575,576]. Insights gained into viral evolution and fitness are limited in context to the predominant viral subtypes and host genetics (e.g. common HLA alleles) in the populations studied, or to the viral genetic backbone into which drug resistance mutations were inserted. Additional work is required to replicate these findings in different populations and/or genetic backgrounds. The same is true for HCV; since the majority of direct-acting antivirals are approved only for use in HCV GT1 infection, additional options for treatment and monitoring of non-GT1 infection are needed. This thesis made extensive use of bioinformatic software for the analysis of raw sequence data. While Chapter 2 presents the clinical validation of software used to analyze Sanger sequence data, the next-generation sequence analysis pipelines used in Chapters 5, 6, and 7 have not been similarly evaluated. Variant calling procedures relied on sequence quality, mapping and alignment parameters that were 164  largely chosen ad hoc, or left at the default values. The relevance of these thresholds has not been validated and the effect of varying parameters on reported results has not been systematically examined. For example, while the analysis pipelines used for the MiSeq data in Chapters 6 and 7 were virtually identical (outside minor versioning differences), the minimum sequence coverage thresholds used as pass/fail criteria were chosen somewhat arbitrarily. This common problem in the literature is not limited to HIV studies [577–580] as data analysis “best practices” defined in human genomic studies may not be applicable to more diverse species [491,540].  Sequencing errors are only one of several factors affecting the accurate detection of rare variants by next-generation sequencing. Multiple steps in library preparation may contribute to false positive variant detection. For example, base misincorporation during reverse transcription or early PCR cycles can lead to artefactual detection of sequence variants [489]. Poor fidelity of the PCR polymerase enzyme results in a base substitution error rate of approximately 0.2%, and recombination during PCR may result in up to 2% of reads being chimeric [581]. These limitations were not addressed by the studies presented here. Oversampling of templates can exacerbate this issue particularly in reactions with low input copy numbers (i.e. plasma samples with low HIV-1 viral load). Briefly, unequal sampling either by biased primers or purely stochastic processes can lead to disproportionate amplification of individual templates. In extreme cases this may result into a 100-fold distortion of variant frequencies [428]. Methods such as “primer ID” tagging have been developed to quantify and correct these issues. Briefly, a degenerate string of nucleotides is added to the 5’ end of the reverse transcription primer such that transcribed cDNA is tagged with a unique string of bases. This tag is carried forward during subsequent rounds of PCR amplification. When sequenced, primer ID tags allow individual sequence reads to be traced back to a single RNA template thus allowing the number of successfully sampled RNA templates to be quantified. Furthermore, if templates tagged with a specific string of nucleotides are observed multiple times, a plurality consensus sequence generated from all of these recovered reads would allow PCR and sequencing errors to be corrected [428]. However, the primer ID approach is not without its own set of limitations. First, primer IDs must be properly designed with input template concentration and sequencing depth in mind in order to 165  prevent multiple cDNA from sharing a single tag. Second, a primer ID must be observed in at least three sequences before error consensus-based correction rules can be applied. However, if samples with high input template numbers are sequenced at relatively low depth of coverage a substantial number of templates are only sequenced once or twice; these singletons and doubleton sequences would be discarded by an error correction strategy [582]. Finally, since current next-generation sequencing reads remain relatively short, multiple reverse transcription reactions would need to be performed in order to generate amplicons spanning relevant drug target genes. The alternate strategy, generating longer amplicons that are subsequently sheared into appropriately-sized fragments (such as with the Illumina Nextera procedure), would produce libraries lacking these primer ID tags. The requirement to perform multiple reverse transcription reactions would substantially increase the complexity and expense of HIV-1 and HCV drug resistance testing. Given these limitations, the primer ID strategy, while theoretically intriguing, might not be feasible for routine clinical testing. Finally, clinically-relevant cutoffs of minor species prevalence must be determined before informed clinical decisions can be made using next-generation sequence data. In order to draw meaningful conclusions, these studies must use treatment response as the outcome variable, similar to how the minority species cutoffs for genotypic HIV-1 co-receptor tropism testing were optimized using outcomes from the MOTIVATE and MERIT clinical trials of maraviroc [322,323,344,480]. For example, antiretroviral naïve patients initiating HAART may be classified as having drug-resistant or susceptible HIV according to a deep-sequencing assay. Performing a receiver-operator characteristic curve analysis by varying the threshold at which minor variants are called, would identify the relevant variant threshold that optimizes the prediction of virologic failure. The “wide” sequencing approach described in Chapter 6 may allow the required data to be rapidly and economically collected. In conclusion, the research presented in this thesis supports the ongoing role of genetic sequencing in infectious disease treatment and monitoring. It presents applications of sequencing to the study of viral diversity, and highlights the benefits of developing new, more sensitive and more cost-effective 166  methods for clinical drug resistance testing. Clearly, next-generation sequencing and the regulation of its use in clinical practice will remain a constantly-evolving field. 167  1.  Centers for Disease Control. Pneumocystis pneumonia--Los Angeles. MMWR Morb Mortal Wkly Rep. 1981;30(21):250–2.  2.  Gottlieb MS, Schroff R, Schanker HM, Weisman JD, Fan PT, Wolf RA, Saxon A. Pneumocystis carinii pneumonia and mucosal candidiasis in previously healthy homosexual men: evidence of a new acquired cellular immunodeficiency. N Engl J Med. 1981;305(24):1425–31.  3.  Alizon M, Sonigo P, Barré-Sinoussi F, Chermann JC, Tiollais P, Montagnier L, Wain-Hobson S. Molecular cloning of lymphadenopathy-associated virus. Nature. 1984;312(5996):757–60.  4.  Broder S, Gallo RC. A pathogenic retrovirus (HTLV-III) linked to AIDS. N Engl J Med. 1984;311(20):1292–7.  5.  Shaw GM, Hahn BH, Arya SK, Groopman JE, Gallo RC, Wong-Staal F. Molecular characterization of human T-cell leukemia (lymphotropic) virus type III in the acquired immune deficiency syndrome. Science. 1984;226(4679):1165–71.  6.  UNAIDS. GLOBAL REPORT: UNAIDS report on the global AIDS epidemic 2013. UNAIDS. 2013.  7.  GBD 2013 Mortality and Causes of Death Collaborators. Global, regional and national levels of age-specific mortality and 240 causes of death, 1990-2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2015;385(9963):117–71.  8.  Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, Abraham J, Adair T, Aggarwal R, Ahn SY, Alvarado M, Anderson HR, Anderson LM, Andrews KG, Atkinson C, Baddour LM, Barker-Collo S, Bartels DH, Bell ML, Benjamin EJ, Bennett D, Bhalla K, Bikbov B, Bin Abdulhak A, Birbeck G, Blyth F, Bolliger I, Boufous S, Bucello C, Burch M, Burney P, Carapetis J, Chen H, Chou D, Chugh SS, Coffeng LE, Colan SD, Colquhoun S, Colson KE, Condon J, Connor MD, Cooper LT, Corriere M, Cortinovis M, de Vaccaro KC, Couser W, Cowie BC, Criqui MH, Cross M, Dabhadkar KC, Dahodwala N, De Leo D, Degenhardt L, Delossantos A, Denenberg J, Des Jarlais DC, Dharmaratne SD, Dorsey ER, Driscoll T, Duber H, Ebel B, Erwin PJ, Espindola P, Ezzati M, Feigin V, Flaxman AD, Forouzanfar MH, Fowkes FGR, Franklin R, Fransen M, Freeman MK, Gabriel SE, Gakidou E, Gaspari F, Gillum RF, Gonzalez-Medina D, Halasa YA, Haring D, Harrison JE, Havmoeller R, Hay RJ, Hoen B, Hotez PJ, Hoy D, Jacobsen KH, James SL, Jasrasaria R, Jayaraman S, Johns N, Karthikeyan G, Kassebaum N, Keren A, Khoo J-P, Knowlton LM, Kobusingye O, Koranteng A, Krishnamurthi R, Lipnick M, Lipshultz SE, Ohno SL, Mabweijano J, MacIntyre MF, Mallinger L, March L, Marks GB, Marks R, Matsumori A, Matzopoulos R, Mayosi BM, McAnulty JH, McDermott MM, McGrath J, Mensah GA, Merriman TR, Michaud C, Miller M, Miller TR, Mock C, Mocumbi AO, Mokdad AA, Moran A, Mulholland K, Nair MN, Naldi L, Narayan KMV, Nasseri K, Norman P, O’Donnell M, Omer SB, Ortblad K, Osborne R, Ozgediz D, Pahari B, Pandian JD, Rivero AP, Padilla RP, Perez-Ruiz F, Perico N, Phillips D, Pierce K, Pope CA, Porrini E, Pourmalek F, Raju M, Ranganathan D, Rehm JT, Rein DB, Remuzzi G, Rivara FP, Roberts T, De León FR, Rosenfeld LC, Rushton L, Sacco RL, Salomon JA, Sampson U, Sanman E, Schwebel DC, Segui-Gomez M, Shepard DS, Singh D, Singleton J, Sliwa K, Smith E, Steer A, Taylor JA, Thomas B, Tleyjeh IM, Towbin JA, Truelsen T, Undurraga EA, Venketasubramanian N, Vijayakumar L, Vos T, Wagner GR, Wang M, Wang W, Watt K, Weinstock MA, Weintraub R, Wilkinson JD, Woolf AD, Wulf S, Yeh P-H, Yip P, Zabetian A, Zheng Z-J, Lopez AD, Murray CJL, AlMazroa MA, Memish ZA. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2095–128.  168  9.  Public Health Agency of Canada. HIV/AIDS Epi Update—Chapter 1: National HIV Prevalence and Incidence Estimates for 2011 [Internet]. 2014 [cited 2015 Mar 17]. Available from: http://www.phac-aspc.gc.ca/aids-sida/publication/epi/2010/pdf/EN_Intro_Web.pdf 10.  Simon V, Ho DD, Karim QA. HIV/AIDS epidemiology, pathogenesis, prevention, and treatment. Lancet. 2006;368(9534):489–504.  11.  Beyrer C, Baral SD, Van Griensven F, Goodreau SM, Chariyalertsak S, Wirtz AL, Brookmeyer R. Global epidemiology of HIV infection in men who have sex with men. Lancet. 2012;380(9839):367–77.  12.  Mathers BM, Degenhardt L, Phillips B, Wiessing L, Hickman M, Strathdee SA, Wodak A, Panda S, Tyndall M, Toufik A, Mattick RP. Global epidemiology of injecting drug use and HIV among people who inject drugs: a systematic review. Lancet. 2008;372(9651):1733–45.  13.  Gouws E, Cuchi P. Focusing the HIV response through estimating the major modes of HIV transmission: a multi-country analysis. Sex Transm Infect. 2012;88(Suppl 2):i76–85.  14.  Gentile M, Adrian T, Scheidler A, Ewald M, Dianzani F, Pauli G, Gelderblom HR. Determination of the size of HIV using adenovirus type 2 as an internal length marker. J Virol Methods. 1994;48(1):43–52.  15.  Ratner L, Haseltine W, Patarca R, Livak KJ, Starcich B, Josephs SF, Doran ER, Rafalski JA, Whitehorn EA, Baumeister K. Complete nucleotide sequence of the AIDS virus, HTLV-III. Nature. 1985;313(6000):277–84.  16.  Coffin JM, Hughes SH, Varmus HE, editors. Retroviruses. Cold Spring Harbor, USA: Cold Spring Harbor Laboratory Press; 1997.  17.  Hallenberger S, Bosch V, Angliker H, Shaw E, Klenk HD, Garten W. Inhibition of furin-mediated cleavage activation of HIV-1 glycoprotein gp160. Nature. 1992;360(6402):358–61.  18.  Jacks T, Power MD, Masiarz FR, Luciw PA, Barr PJ, Varmus HE. Characterization of ribosomal frameshifting in HIV-1 gag-pol expression. Nature. 1988;331(6153):280–3.  19.  Kohl NE, Emini EA, Schleif WA, Davis LJ, Heimbach JC, Dixon RA, Scolnick EM, Sigal IS. Active human immunodeficiency virus protease is required for viral infectivity. Proc Natl Acad Sci U S A. 1988;85(13):4686–90.  20.  Frankel AD, Young JA. HIV-1: fifteen proteins and an RNA. Annu Rev Biochem. 1998;67:1–25.  21.  Wong RW, Balachandran A, Ostrowski MA, Cochrane A. Digoxin suppresses HIV-1 replication by altering viral RNA processing. PLoS Pathog. 2013;9(3):e1003241.  22.  Wyatt R, Sodroski J. The HIV-1 envelope glycoproteins: fusogens, antigens, and immunogens. Science. 1998;280(5371):1884–8.  23.  Dalgleish AG, Beverley PC, Clapham PR, Crawford DH, Greaves MF, Weiss RA. The CD4 (T4) antigen is an essential component of the receptor for the AIDS retrovirus. Nature. 1984;312(5996):763–7.  169  24.  Klatzmann D, Champagne E, Chamaret S, Gruest J, Guetard D, Hercend T, Gluckman JC, Montagnier L. T-lymphocyte T4 molecule behaves as the receptor for human retrovirus LAV. Nature. 1985;312(5996):767–8.  25.  Trkola A, Dragic T, Arthos J, Binley JM, Olson WC, Allaway GP, Cheng-Mayer C, Robinson J, Maddon PJ, Moore JP. CD4-dependent, antibody-sensitive interactions between HIV-1 and its co-receptor CCR-5. Nature. 1996;384(6605):184–7.  26.  Wu L, Gerard NP, Wyatt R, Choe H, Parolin C, Ruffing N, Borsetti A, Cardoso AA, Desjardin E, Newman W, Gerard C, Sodroski J. CD4-induced interaction of primary HIV-1 gp120 glycoproteins with the chemokine receptor CCR-5. Nature. 1996;384(6605):179–83.  27.  Bandres JC, Wang QF, O’Leary J, Baleaux F, Amara A, Hoxie JA, Zolla-Pazner S, Gorny MK. Human immunodeficiency virus (HIV) envelope binds to CXCR4 independently of CD4, and binding can be enhanced by interaction with soluble CD4 or by HIV envelope deglycosylation. J Virol. 1998;72(3):2500–4.  28.  Deng H, Liu R, Ellmeier W, Choe S, Unutmaz D, Burkhart M, Di Marzio P, Marmon S, Sutton RE, Hill CM, Davis CB, Peiper SC, Schall TJ, Littman DR, Landau NR. Identification of a major co-receptor for primary isolates of HIV-1. Nature. 1996;381(6584):661–6.  29.  Lapham CK, Ouyang J, Chandrasekhar B, Nguyen NY, Dimitrov DS, Golding H. Evidence for cell-surface association between fusin and the CD4-gp120 complex in human cell lines. Science. 1996;274(5287):602–5.  30.  Rizzuto CD, Wyatt R, Hernández-Ramos N, Sun Y, Kwong PD, Hendrickson WA, Sodroski J. A conserved HIV gp120 glycoprotein structure involved in chemokine receptor binding. Science. 1998;280(5371):1949–53.  31.  Choe H, Farzan M, Sun Y, Sullivan N, Rollins B, Ponath PD, Wu L, Mackay CR, LaRosa G, Newman W, Gerard N, Gerard C, Sodroski J. The β-chemokine receptors CCR3 and CCR5 facilitate infection by primary HIV-1 isolates. Cell. 1996;85(7):1135–48.  32.  Doranz BJ, Rucker J, Yi Y, Smyth RJ, Samson M, Peiper SC, Parmentier M, Collman RG, Doms RW. A dual-tropic primary HIV-1 isolate that uses fusin and the β-chemokine receptors CKR-5, CKR-3, and CKR-2b as fusion cofactors. Cell. 1996;85(7):1149–58.  33.  Alkhatib G, Liao F, Berger EA, Farber JM, Peden KW. A new SIV co-receptor, STRL33. Nature. 1997;388(6639):238.  34.  Deng HK, Unutmaz D, KewalRamani VN, Littman DR. Expression cloning of new receptors used by simian and human immunodeficiency viruses. Nature. 1997;388(6639):296–300.  35.  Sattentau QJ, Moore JP. Conformational changes induced in the human immunodeficiency virus envelope glycoprotein by soluble CD4 binding. J Exp Med. 1991;174(2):407–15.  36.  Thali M, Moore JP, Furman C, Charles M, Ho DD, Robinson J, Sodroski J. Characterization of conserved human immunodeficiency virus type 1 gp120 neutralization epitopes exposed upon gp120-CD4 binding. J Virol. 1993;67(7):3978–88.  37.  Cocchi F, DeVico AL, Garzino-Demo A, Cara A, Gallo RC, Lusso P. The V3 domain of the HIV-1 gp120 envelope glycoprotein is critical for chemokine-mediated blockade of infection. Nat Med. 1996;2(11):1244–7.  170  38.  Speck RF, Wehrly K, Platt EJ, Atchison RE, Charo IF, Kabat D, Chesebro B, Goldsmith MA. Selective employment of chemokine receptors as human immunodeficiency virus type 1 coreceptors determined by individual amino acids within the envelope V3 loop. J Virol. 1997;71(9):7136–9.  39.  Weiss CD. HIV-1 gp41: Mediator of fusion and target for inhibition. AIDS Rev. 2003;5(4):214–21.  40.  Melikyan GB, Markosyan RM, Hemmati H, Delmedico MK, Lambert DM, Cohen FS. Evidence that the transition of HIV-1 gp41 into a six-helix bundle, not the bundle configuration, induces membrane fusion. J Cell Biol. 2000;151(2):413–23.  41.  Markosyan RM, Cohen FS, Melikyan GB. HIV-1 envelope proteins complete their folding into six-helix bundles immediately after fusion pore formation. Mol Biol Cell. 2003;14(3):926–38.  42.  Ambrose Z, Aiken C. HIV-1 uncoating: connection to nuclear entry and regulation by host proteins. Virology. 2014;454-455:371–9.  43.  Braaten D, Franke EK, Luban J. Cyclophilin A is required for an early step in the life cycle of human immunodeficiency virus type 1 before the initiation of reverse transcription. J Virol. 1996;70(6):3551–60.  44.  Nisole S, Saïb A. Early steps of retrovirus replicative cycle. Retrovirology. 2004;1(9):9.  45.  Engelman A, Cherepanov P. The structural biology of HIV-1: mechanistic and therapeutic insights. Nat Rev Microbiol. 2012;10(4):279–90.  46.  Mansky LM, Temin HM. Lower in vivo mutation rate of human immunodeficiency virus type 1 than that predicted from the fidelity of purified reverse transcriptase. J Virol. 1995;69(8):5087–94.  47.  Perelson AS, Neumann AU, Markowitz M, Leonard JM, Ho DD. HIV-1 dynamics in vivo: virion clearance rate, infected cell life-span, and viral generation time. Science. 1996;271(5255):1582–6.  48.  Coffin JM. HIV population dynamics in vivo: implications for genetic variation, pathogenesis, and therapy. Science. 1995;267(5197):483–9.  49.  Miller MD, Farnet CM, Bushman FD. Human immunodeficiency virus type 1 preintegration complexes: studies of organization and composition. J Virol. 1997;71(7):5382–90.  50.  Kogan M, Rappaport J. HIV-1 accessory protein Vpr: relevance in the pathogenesis of HIV and potential for therapeutic intervention. Retrovirology. 2011;8(1):25.  51.  Hearps AC, Jans DA. HIV-1 integrase is capable of targeting DNA to the nucleus via an importin alpha/beta-dependent mechanism. Biochem J. 2006;398(3):475–84.  52.  Terry LJ, Shows EB, Wente SR. Crossing the nuclear envelope: hierarchical regulation of nucleocytoplasmic transport. Science. 2007;318(2000):1412–6.  53.  Cherepanov P, Maertens G, Proost P, Devreese B, Van Beeumen J, Engelborghs Y, De Clercq E, Debyser Z. HIV-1 integrase forms stable tetramers and associates with LEDGF/p75 protein in human cells. J Biol Chem. 2003;278(1):372–81.  54.  Ruelas DS, Greene WC. An integrated overview of HIV-1 latency. Cell. 2013;155(3):519–29.  171  55.  Schwartz S, Felber BK, Benko DM, Fenyö EM, Pavlakis GN. Cloning and functional analysis of multiply spliced mRNA species of human immunodeficiency virus type 1. J Virol. 1990;64(6):2519–29.  56.  Felber BK, Drysdale CM, Pavlakis GN. Feedback regulation of human immunodeficiency virus type 1 expression by the Rev protein. J Virol. 1990;64(8):3734–41.  57.  Stoltzfus CM, Madsen JM. Role of viral splicing elements and cellular RNA binding proteins in regulation of HIV-1 alternative RNA splicing. Curr HIV Res. 2006;4(1):43–55.  58.  Malim MH, Emerman M. HIV-1 accessory proteins-ensuring viral survival in a hostile environment. Cell Host Microbe. 2008;3(6):388–98.  59.  Roeth JF, Collins KL. Human immunodeficiency virus type 1 Nef: adapting to intracellular trafficking pathways. Microbiol Mol Biol Rev. 2006;70(2):548–63.  60.  Neil SJD, Eastman SW, Jouvenet N, Bieniasz PD. HIV-1 Vpu promotes release and prevents endocytosis of nascent retrovirus particles from the plasma membrane. PLoS Pathog. 2006;2(5):e39.  61.  Sheehy AM, Gaddis NC, Choi JD, Malim MH. Isolation of a human gene that inhibits HIV-1 infection and is suppressed by the viral Vif protein. Nature. 2002;418(6898):646–50.  62.  Goh WC, Rogel ME, Kinsey CM, Michael SF, Fultz PN, Nowak MA, Hahn BH, Emerman M. HIV-1 Vpr increases viral expression by manipulation of the cell cycle: a mechanism for selection of Vpr in vivo. Nat Med. 1998;4(1):65–71.  63.  Sundquist WI, Kräusslich HG. HIV-1 assembly, budding, and maturation. Cold Spring Harb Perspect Med. 2012;2(7):a006924.  64.  Votteler J, Sundquist WI. Virus budding and the ESCRT pathway. Cell Host Microbe. 2013;14(3):232–41.  65.  Sharp PM, Hahn BH. Origins of HIV and the AIDS pandemic. Cold Spring Harb Perspect Med. 2011;1(1):a006841.  66.  Campbell-Yesufu OT, Gandhi RT. Update on human immunodeficiency virus (HIV)-2 infection. Clin Infect Dis. 2011;52(6):780–7.  67.  Robertson DL, Anderson JP, Bradac JA, Carr JK, Foley B, Funkhouser RK, Gao F, Hahn BH, Kalish ML, Kuiken C, Learn GH, Leitner T, McCutchan F, Osmanov S, Peeters M, Pieniazek D, Salminen M, Sharp PM, Wolinsky S, Korber B. HIV-1 nomenclature proposal. Science. 2000;288(5463):55–6.  68.  Plantier J-C, Leoz M, Dickerson JE, De Oliveira F, Cordonnier F, Lemée V, Damond F, Robertson DL, Simon F. A new human immunodeficiency virus derived from gorillas. Nat Med. 2009;15(8):871–2.  69.  Keele BF, Van Heuverswyn F, Li Y, Bailes E, Takehisa J, Santiago ML, Bibollet-Ruche F, Chen Y, Wain LV, Liegeois F, Loul S, Ngole EM, Bienvenue Y, Delaporte E, Brookfield JFY, Sharp PM, Shaw GM, Peeters M, Hahn BH. Chimpanzee reservoirs of pandemic and nonpandemic HIV-1. Science. 2006;313(5786):523–6.  172  70.  Wertheim JO, Worobey M. Dating the age of the SIV lineages that gave rise to HIV-1 and HIV-2. PLoS Comput Biol. 2009;5(5):e1000377.  71.  Faria NR, Rambaut A, Suchard MA, Baele G, Bedford T, Ward MJ, Tatem AJ, Sousa JD, Arinaminpathy N, Pepin J, Posada D, Peeters M, Pybus OG, Lemey P. The early spread and epidemic ignition of HIV-1 in human populations. Science. 2014;346(6205):56–61.  72.  Hemelaar J. The origin and diversity of the HIV-1 pandemic. Trends Mol Med. 2012;18(3):182–92.  73.  D’arc M, Ayouba A, Esteban A, Learn GH, Boué V, Liegeois F, Etienne L, Tagg N, Leendertz FH, Boesch C, Madinda NF, Robbins MM, Gray M, Cournil A, Ooms M, Letko M, Simon VA, Sharp PM, Hahn BH, Delaporte E, Mpoudi Ngole E, Peeters M. Origin of the HIV-1 group O epidemic in western lowland gorillas. Proc Natl Acad Sci U S A. 2015;12(11):E1343–52.  74.  Santiago ML, Range F, Keele BF, Li Y, Bailes E, Bibollet-Ruche F, Fruteau C, Noë R, Peeters M, Brookfield JFY, Shaw GM, Sharp PM, Hahn BH. Simian immunodeficiency virus infection in free-ranging sooty mangabeys (Cercocebus atys atys) from the Taï Forest, Côte d’Ivoire: implications for the origin of epidemic human immunodeficiency virus type 2. J Virol. 2005;79(19):12515–27.  75.  Korber B, Gaschen B, Yusim K, Thakallapally R, Kesmir C, Detours V. Evolutionary and immunological implications of contemporary HIV-1 variation. Br Med Bull. 2001;58:19–42.  76.  Keele BF, Giorgi EE, Salazar-Gonzalez JF, Decker JM, Pham KT, Salazar MG, Sun C, Grayson T, Wang S, Li H, Wei X, Jiang C, Kirchherr JL, Gao F, Anderson JA, Ping L-H, Swanstrom R, Tomaras GD, Blattner WA, Goepfert PA, Kilby JM, Saag MS, Delwart EL, Busch MP, Cohen MS, Montefiori DC, Haynes BF, Gaschen B, Athreya GS, Lee HY, Wood N, Seoighe C, Perelson AS, Bhattacharya T, Korber BT, Hahn BH, Shaw GM. Identification and characterization of transmitted and early founder virus envelopes in primary HIV-1 infection. Proc Natl Acad Sci U S A. 2008;105(21):7552–7.  77.  Abrahams M-R, Anderson JA, Giorgi EE, Seoighe C, Mlisana K, Ping L-H, Athreya GS, Treurnicht FK, Keele BF, Wood N, Salazar-Gonzalez JF, Bhattacharya T, Chu H, Hoffman I, Galvin S, Mapanje C, Kazembe P, Thebus R, Fiscus S, Hide W, Cohen MS, Karim SA, Haynes BF, Shaw GM, Hahn BH, Korber BT, Swanstrom R, Williamson C. Quantitating the multiplicity of infection with human immunodeficiency virus type 1 subtype C reveals a non-poisson distribution of transmitted variants. J Virol. 2009;83(8):3556–67.  78.  Borrow P, Lewicki H, Wei X, Horwitz MS, Peffer N, Meyers H, Nelson JA, Gairin JE, Hahn BH, Oldstone MB, Shaw GM. Antiviral pressure exerted by HIV-1-specific cytotoxic T lymphocytes (CTLs) during primary infection demonstrated by rapid selection of CTL escape virus. Nat Med. 1997;3(2):205–11.  79.  Moore CB, John M, James IR, Christiansen FT, Witt CS, Mallal SA. Evidence of HIV-1 adaptation to HLA-restricted immune responses at a population level. Science. 2002;296(5572):1439–43.  80.  Frost SDW, Wrin T, Smith DM, Kosakovsky Pond SL, Liu Y, Paxinos E, Chappey C, Galovich J, Beauchaine J, Petropoulos CJ, Little SJ, Richman DD. Neutralizing antibody responses drive the evolution of human immunodeficiency virus type 1 envelope during recent HIV infection. Proc Natl Acad Sci U S A. 2005;102(51):18514–9.  81.  Frost SD, Dumaurier MJ, Wain-Hobson S, Brown AJ. Genetic drift and within-host metapopulation dynamics of HIV-1 infection. Proc Natl Acad Sci U S A. 2001;98(12):6975–80.  173  82.  Shankarappa R, Margolick JB, Gange SJ, Rodrigo AG, Upchurch D, Farzadegan H, Gupta P, Rinaldo CR, Learn GH, He X, Huang XL, Mullins JI. Consistent viral evolutionary changes associated with the progression of human immunodeficiency virus type 1 infection. J Virol. 1999;73(12):10489–502.  83.  Patel P, Borkowf CB, Brooks JT, Lasry A, Lansky A, Mermin J. Estimating per-act HIV transmission risk: a systematic review. AIDS. 2014;28(10):1509–19.  84.  Quinn TC, Wawer MJ, Sewankambo N, Serwadda D, Li C, Wabwire-Mangen F, Meehan MO, Lutalo T, Gray RH, the Rakai Project Study Group. Viral load and heterosexual transmission of human immunodeficiency virus type 1. N Engl J Med. 2000;342(13):921–9.  85.  Gray RH, Wawer MJ, Brookmeyer R, Sewankambo NK, Serwadda D, Wabwire-Mangen F, Lutalo T, Li X, VanCott T, Quinn TC. Probability of HIV-1 transmission per coital act in monogamous, heterosexual, HIV-1-discordant couples in Rakai, Uganda. Lancet. 2001;357(9263):1149–53.  86.  Cooper DA, MacLean P, Gold J, Donovan B, Finlayson R, Barnes TG, Michelmore HM, Brooke P, Penny R. Acute AIDS retrovirus infection. Definition of a clinical illness associated with seroconversion. Lancet. 1985;1(8428):537–40.  87.  Guadalupe M, Reay E, Sankaran S, Prindiville T, Flamm J, McNeil A, Dandekar S. Severe CD4+ T-cell depletion in gut lymphoid tissue during primary human immunodeficiency virus type 1 infection and substantial delay in restoration following highly active antiretroviral therapy. J Virol. 2003;77(21):11708–17.  88.  Mehandru S, Poles MA, Tenner-Racz K, Horowitz A, Hurley A, Hogan C, Boden D, Racz P, Markowitz M. Primary HIV-1 infection is associated with preferential depletion of CD4+ T lymphocytes from effector sites in the gastrointestinal tract. J Exp Med. 2004;200(6):761–70.  89.  McMichael AJ, Borrow P, Tomaras GD, Goonetilleke N, Haynes BF. The immune response during acute HIV-1 infection: clues for vaccine development. Nat Rev Immunol. 2010;10(1):11–23.  90.  Schneider E, Whitmore S, Glynn KM, Dominguez K, Mitsch A, McKenna MT. Revised surveillance case definitions for HIV infection among adults, adolescents, and children aged <18 months and for HIV infection and AIDS among children aged 18 months to <13 years--United States, 2008. MMWR Recomm Rep. 2008;57(RR-10):1–12.  91.  Fiebig EW, Wright DJ, Rawal BD, Garrett PE, Schumacher RT, Peddada L, Heldebrant C, Smith R, Conrad A, Kleinman SH, Busch MP. Dynamics of HIV viremia and antibody seroconversion in plasma donors: implications for diagnosis and staging of primary HIV infection. AIDS. 2003;17(13):1871–9.  92.  Koup RA, Safrit JT, Cao Y, Andrews CA, McLeod G, Borkowsky W, Farthing C, Ho DD. Temporal association of cellular immune responses with the initial control of viremia in primary human immunodeficiency virus type 1 syndrome. J Virol. 1994;68(7):4650–5.  93.  Yang OO, Kalams SA, Trocha A, Cao H, Luster A, Johnson RP, Walker BD. Suppression of human immunodeficiency virus type 1 replication by CD8+ cells: evidence for HLA class I-restricted triggering of cytolytic and noncytolytic mechanisms. J Virol. 1997;71(4):3120–8.  94.  Jin X, Bauer DE, Tuttleton SE, Lewin S, Gettie A, Blanchard J, Irwin CE, Safrit JT, Mittler J, Weinberger L, Kostrikis LG, Zhang L, Perelson AS, Ho DD. Dramatic rise in plasma viremia after CD8(+) T cell depletion in simian immunodeficiency virus-infected macaques. J Exp Med. 1999;189(6):991–8.  174  95.  Schmitz JE, Kuroda MJ, Santra S, Sasseville VG, Simon MA, Lifton MA, Racz P, Tenner-Racz K, Dalesandro M, Scallon BJ, Ghrayeb J, Forman MA, Montefiori DC, Rieber EP, Letvin NL, Reimann KA. Control of viremia in simian immunodeficiency virus infection by CD8+ lymphocytes. Science. 1999;283(5403):857–60.  96.  Jenkins MR, Griffiths GM. The synapse and cytolytic machinery of cytotoxic T cells. Curr Opin Immunol. 2010;22(3):308–13.  97.  Busch MP, Lee LL, Satten GA, Henrard DR, Farzadegan H, Nelson KE, Read S, Dodd RY, Petersen LR. Time course of detection of viral and serologic markers preceding human immunodeficiency virus type 1 seroconversion: implications for screening of blood and tissue donors. Transfusion. 1995;35(2):91–7.  98.  Gray ES, Madiga MC, Moore PL, Mlisana K, Abdool Karim SS, Binley JM, Shaw GM, Mascola JR, Morris L. Broad neutralization of human immunodeficiency virus type 1 mediated by plasma antibodies against the gp41 membrane proximal external region. J Virol. 2009;83(21):11265–74.  99.  Bonhoeffer S, Fraser C, Leventhal GE. High heritability is compatible with the broad distribution of set point viral load in HIV carriers. PLOS Pathog. 2015;11(2):e1004634.  100.  Mellors JW, Rinaldo CR, Gupta P, White RM, Todd JA, Kingsley LA. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science. 1996;272(5265):1167–70.  101.  Fraser C, Hollingsworth TD, Chapman R, de Wolf F, Hanage WP. Variation in HIV-1 set-point viral load: epidemiological analysis and an evolutionary hypothesis. Proc Natl Acad Sci U S A. 2007;104(44):17441–6.  102.  Geskus RB, Prins M, Hubert J-B, Miedema F, Berkhout B, Rouzioux C, Delfraissy J-F, Meyer L. The HIV RNA setpoint theory revisited. Retrovirology. 2007;4:65.  103.  Muñoz A, Wang MC, Bass S, Taylor JM, Kingsley LA, Chmiel JS, Polk BF, the Multicenter AIDS Cohort Study Group. Acquired immunodeficiency syndrome (AIDS)-free time after human immunodeficiency virus type 1 (HIV-1) seroconversion in homosexual men. Am J Epidemiol. 1989;130(3):530–9.  104.  Collaborative Group on AIDS Incubation and HIV Survival including the CASCADE EU Concerted Action. Time from HIV-1 seroconversion to AIDS and death before widespread use of highly-active antiretroviral therapy: a collaborative re-analysis. Lancet. 2000;355(9210):1131–7.  105.  Haynes BF, Pantaleo G, Fauci AS. Toward an understanding of the correlates of protective immunity to HIV infection. Science. 1996;271(5247):324–8.  106.  Casado C, Colombo S, Rauch A, Martínez R, Günthard HF, Garcia S, Rodríguez C, del Romero J, Telenti A, López-Galíndez C. Host and viral genetic correlates of clinical definitions of HIV-1 disease progression. PLoS One. 2010;5(6):e11079.  107.  Fowke KR, Nagelkerke NJD, Kimani J, Simonsen JN, Anzala AO, Bwayo JJ, MacDonald KS, Ngugi EN, Plummer FA. Resistance to HIV-1 infection among persistently seronegative prostitutes in Nairobi, Kenya. Lancet. 1996;348(9038):1347–51.  108.  Detels R, Liu Z, Hennessey K, Kan J, Visscher BR, Taylor JM, Hoover DR, Rinaldo CR, Phair JP, Saah AJ, Giorgi JV, the Multicenter AIDS Cohort Study. Resistance to HIV-1 infection. J Acquir Immune Defic Syndr. 1994;7(12):1263–9.  175  109.  Cao Y, Qin L, Zhang L, Safrit J, Ho DD. Virologic and immunologic characterization of long-term survivors of human immunodeficiency virus type 1 infection. N Engl J Med. 1995;332(4):201–8.  110.  Pantaleo G, Menzo S, Vaccarezza M, Graziosi C, Cohen OJ, Demarest JF, Montefiori D, Orenstein JM, Fox C, Schrager LK. Studies in subjects with long-term nonprogressive human immunodeficiency virus infection. N Engl J Med. 1995;332(4):209–16.  111.  Deeks SG, Walker BD. Human immunodeficiency virus controllers: mechanisms of durable virus control in the absence of antiretroviral therapy. Immunity. 2007. p. 406–16.  112.  Phair J, Jacobson L, Detels R, Rinaldo C, Saah A, Schrager L, Muñoz A. Acquired immune deficiency syndrome occurring within 5 years of infection with human immunodeficiency virus type-1: the Multicenter AIDS Cohort Study. J Acquir Immune Defic Syndr. 1992;5(5):490–6.  113.  Markowitz M, Mohri H, Mehandru S, Shet A, Berry L, Kalyanaraman R, Kim A, Chung C, Jean-Pierre P, Horowitz A, La Mar M, Wrin T, Parkin N, Poles M, Petropoulos C, Mullen M, Boden D, Ho DD. Infection with multidrug resistant, dual-tropic HIV-1 and rapid progression to AIDS: a case report. Lancet. 2005;365(9464):1031–8.  114.  Alizon S, von Wyl V, Stadler T, Kouyos RD, Yerly S, Hirschel B, Böni J, Shah C, Klimkait T, Furrer H, Rauch A, Vernazza PL, Bernasconi E, Battegay M, Bürgisser P, Telenti A, Günthard HF, Bonhoeffer S. Phylogenetic approach reveals that virus genotype largely determines HIV set-point viral load. PLoS Pathog. 2010;6(9):e1001123.  115.  Hollingsworth TD, Laeyendecker O, Shirreff G, Donnelly CA, Serwadda D, Wawer MJ, Kiwanuka N, Nalugoda F, Collinson-Streng A, Ssempijja V, Hanage WP, Quinn TC, Gray RH, Fraser C. HIV-1 transmitting couples have similar viral load set-points in Rakai, Uganda. PLoS Pathog. 2010;6(5):e1000876.  116.  Learmont J, Tindall B, Evans L, Cunningham A, Cunningham P, Wells J, Penny R, Kaldor J, Cooper DA. Long-term symptomless HIV-1 infection in recipients of blood products from a single donor. Lancet. 1992;340(8824):863–7.  117.  Deacon NJ, Tsykin A, Solomon A, Smith K, Ludford-Menting M, Hooker DJ, McPhee DA, Greenway AL, Ellett A, Chatfield C, Lawson VA, Crowe S, Maerz A, Sonza S, Learmont J, Sullivan JS, Cunningham A, Dwyer D, Dowton D, Mills J. Genomic structure of an attenuated quasi species of HIV-1 from a blood transfusion donor and recipients. Science. 1995;270(5238):988–91.  118.  Learmont JC, Geczy AF, Mills J, Ashton LJ, Raynes-Greenow CH, Garsia RJ, Dyer WB, McIntyre L, Oelrichs RB, Rhodes DI, Deacon NJ, Sullivan JS. Immunologic and virologic status after 14 to 18 years of infection with an attenuated strain of HIV-1. A report from the Sydney Blood Bank Cohort. N Engl J Med. 1999;340(22):1715–22.  119.  Kouri V, Khouri R, Alemán Y, Abrahantes Y, Vercauteren J, Pineda-Peña A-C, Theys K, Megens S, Moutschen M, Pfeifer N, Van Weyenbergh J, Pérez AB, Pérez J, Pérez L, Van Laethem K, Vandamme A-M. CRF19_cpx is an evolutionary fit HIV-1 variant strongly associated with rapid progression to AIDS in Cuba. EBioMedicine. 2015;2(3):244–54.  120.  Kaleebu P, French N, Mahe C, Yirrell D, Watera C, Lyagoba F, Nakiyingi J, Rutebemberwa A, Morgan D, Weber J, Gilks C, Whitworth J. Effect of human immunodeficiency virus (HIV) type 1 envelope subtypes A and D on disease progression in a large cohort of HIV-1-positive persons in Uganda. J Infect Dis. 2002;185(9):1244–50.  176  121.  Baeten JM, Chohan B, Lavreys L, Chohan V, McClelland RS, Certain L, Mandaliya K, Jaoko W, Overbaugh J. HIV-1 subtype D infection is associated with faster disease progression than subtype A in spite of similar plasma HIV-1 loads. J Infect Dis. 2007;195(8):1177–80.  122.  Kiwanuka N, Laeyendecker O, Robb M, Kigozi G, Arroyo M, McCutchan F, Eller LA, Eller M, Makumbi F, Birx D, Wabwire-Mangen F, Serwadda D, Sewankambo NK, Quinn TC, Wawer M, Gray R. Effect of human immunodeficiency virus Type 1 (HIV-1) subtype on disease progression in persons from Rakai, Uganda, with incident HIV-1 infection. J Infect Dis. 2008;197(5):707–13.  123.  Amornkul PN, Karita E, Kamali A, Rida WN, Sanders EJ, Lakhi S, Price MA, Kilembe W, Cormier E, Anzala O, Latka MH, Bekker L-G, Allen SA, Gilmour J, Fast PE. Disease progression by infecting HIV-1 subtype in a seroconverter cohort in sub-Saharan Africa. AIDS. 2013;27(17):2775–86.  124.  Miura T, Brockman M, Brumme C, Brumme Z, Carlson J, Pereyra F, Trocha A, Addo M, Block B, Rothchild A, Baker B, Flynn T, Schneidewind A, Li B, Wang Y, Heckerman D, Allen T, Walker B. Genetic characterization of human immunodeficiency virus type 1 in elite controllers: lack of gross genetic defects or common amino acid changes. J Virol. 2008;82(17):8422–30.  125.  Fellay J, Shianna KV, Ge D, Colombo S, Ledergerber B, Weale M, Zhang K, Gumbs C, Castagna A, Cossarizza A, Cozzi-Lepri A, De Luca A, Easterbrook P, Francioli P, Mallal S, Martinez-Picado J, Miro JM, Obel N, Smith JP, Wyniger J, Descombes P, Antonarakis SE, Letvin NL, McMichael AJ, Haynes BF, Telenti A, Goldstein DB. A whole-genome association study of major determinants for host control of HIV-1. Science. 2007;317(5840):944–7.  126.  Pereyra F, Jia X, McLaren PJ, Telenti A, de Bakker PIW, Walker BD, Ripke S, Brumme CJ, Pulit SL, Carrington M, Kadie CM, Carlson JM, Heckerman D, Graham RR, Plenge RM, Deeks SG, Gianniny L, Crawford G, Sullivan J, Gonzalez E, Davies L, Camargo A, Moore JM, Beattie N, Gupta S, Crenshaw A, Burtt NP, Guiducci C, Gupta N, Gao X, Qi Y, Yuki Y, Piechocka-Trocha A, Cutrell E, Rosenberg R, Moss KL, Lemay P, O’Leary J, Schaefer T, Verma P, Toth I, Block B, Baker B, Rothchild A, Lian J, Proudfoot J, Alvino DML, Vine S, Addo MM, Allen TM, Altfeld M, Henn MR, Le Gall S, Streeck H, Haas DW, Kuritzkes DR, Robbins GK, Shafer RW, Gulick RM, Shikuma CM, Haubrich R, Riddler S, Sax PE, Daar ES, Ribaudo HJ, Agan B, Agarwal S, Ahern RL, Allen BL, Altidor S, Altschuler EL, Ambardar S, Anastos K, Anderson B, Anderson V, Andrady U, Antoniskis D, Bangsberg D, Barbaro D, Barrie W, Bartczak J, Barton S, Basden P, Basgoz N, Bazner S, Bellos NC, Benson AM, Berger J, Bernard NF, Bernard AM, Birch C, Bodner SJ, Bolan RK, Boudreaux ET, Bradley M, Braun JF, Brndjar JE, Brown SJ, Brown K, Brown ST, Burack J, Bush LM, Cafaro V, Campbell O, Campbell J, Carlson RH, Carmichael JK, Casey KK, Cavacuiti C, Celestin G, Chambers ST, Chez N, Chirch LM, Cimoch PJ, Cohen D, Cohn LE, Conway B, Cooper DA, Cornelson B, Cox DT, Cristofano MV, Cuchural G, Czartoski JL, Dahman JM, Daly JS, Davis BT, Davis K, Davod SM, DeJesus E, Dietz CA, Dunham E, Dunn ME, Ellerin TB, Eron JJ, Fangman JJW, Farel CE, Ferlazzo H, Fidler S, Fleenor-Ford A, Frankel R, Freedberg KA, French NK, Fuchs JD, Fuller JD, Gaberman J, Gallant JE, Gandhi RT, Garcia E, Garmon D, Gathe JC, Gaultier CR, Gebre W, Gilman FD, Gilson I, Goepfert PA, Gottlieb MS, Goulston C, Groger RK, Gurley TD, Haber S, Hardwicke R, Hardy WD, Harrigan PR, Hawkins TN, Heath S, Hecht FM, Henry WK, Hladek M, Hoffman RP, Horton JM, Hsu RK, Huhn GD, Hunt P, Hupert MJ, Illeman ML, Jaeger H, Jellinger RM, John M, Johnson JA, Johnson KL, Johnson H, Johnson K, Joly J, Jordan WC, Kauffman CA, Khanlou H, Killian RK, Kim AY, Kim DD, Kinder CA, Kirchner JT, Kogelman L, Kojic EM, Korthuis PT, Kurisu W, Kwon DS, LaMar M, Lampiris H, Lanzafame M, Lederman MM, Lee DM, Lee JML, Lee MJ, Lee ETY, Lemoine J, Levy JA, Llibre JM, Liguori MA, Little SJ, Liu AY, Lopez AJ, Loutfy MR, Loy D, Mohammed DY, Man A, Mansour MK, Marconi VC, Markowitz M, Marques R, Martin JN, Martin HL, Mayer KH, McElrath MJ, McGhee TA, McGovern BH, McGowan K, McIntyre D, Mcleod GX, Menezes P, Mesa G, Metroka CE, Meyer-Olson D, Miller AO, Montgomery K, Mounzer KC, Nagami EH, Nagin I, Nahass RG, Nelson MO, Nielsen C, Norene DL, O’Connor DH, Ojikutu BO, Okulicz J, Oladehin OO, Oldfield EC, Olender SA, Ostrowski M, 177  Owen WF, Pae E, Parsonnet J, Pavlatos AM, Perlmutter AM, Pierce MN, Pincus JM, Pisani L, Price LJ, Proia L, Prokesch RC, Pujet HC, Ramgopal M, Rathod A, Rausch M, Ravishankar J, Rhame FS, Richards CS, Richman DD, Rodes B, Rodriguez M, Rose RC, Rosenberg ES, Rosenthal D, Ross PE, Rubin DS, Rumbaugh E, Saenz L, Salvaggio MR, Sanchez WC, Sanjana VM, Santiago S, Schmidt W, Schuitemaker H, Sestak PM, Shalit P, Shay W, Shirvani VN, Silebi VI, Sizemore JM, Skolnik PR, Sokol-Anderson M, Sosman JM, Stabile P, Stapleton JT, Starrett S, Stein F, Stellbrink H-J, Sterman FL, Stone VE, Stone DR, Tambussi G, Taplitz RA, Tedaldi EM, Theisen W, Torres R, Tosiello L, Tremblay C, Tribble MA, Trinh PD, Tsao A, Ueda P, Vaccaro A, Valadas E, Vanig TJ, Vecino I, Vega VM, Veikley W, Wade BH, Walworth C, Wanidworanun C, Ward DJ, Warner DA, Weber RD, Webster D, Weis S, Wheeler DA, White DJ, Wilkins E, Winston A, Wlodaver CG, van’t Wout A, Wright DP, Yang OO, Yurdin DL, Zabukovic BW, Zachary KC, Zeeman B, Zhao M. The major genetic determinants of HIV-1 control affect HLA class I peptide presentation. Science. 2010;330(6010):1551–7.  127.  Fellay J, Ge D, Shianna KV, Colombo S, Ledergerber B, Cirulli ET, Urban TJ, Zhang K, Gumbs CE, Smith JP, Castagna A, Cozzi-Lepri A, De Luca A, Easterbrook P, Günthard HF, Mallal S, Mussini C, Dalmau J, Martinez-Picado J, Miro JM, Obel N, Wolinsky SM, Martinson JJ, Detels R, Margolick JB, Jacobson LP, Descombes P, Antonarakis SE, Beckmann JS, O’Brien SJ, Letvin NL, McMichael AJ, Haynes BF, Carrington M, Feng S, Telenti A, Goldstein DB. Common genetic variation and the control of HIV-1 in humans. PLoS Genet. 2009;5(12):e1000791.  128.  Limou S, Le Clerc S, Coulonges C, Carpentier W, Dina C, Delaneau O, Labib T, Taing L, Sladek R, Deveau C, Ratsimandresy R, Montes M, Spadoni J-L, Lelièvre J-D, Lévy Y, Therwath A, Schächter F, Matsuda F, Gut I, Froguel P, Delfraissy J-F, Hercberg S, Zagury J-F. Genomewide association study of an AIDS-nonprogression cohort emphasizes the role played by HLA genes (ANRS Genomewide Association Study 02). J Infect Dis. 2009;199(3):419–26.  129.  Apps R, Qi Y, Carlson JM, Chen H, Gao X, Thomas R, Yuki Y, Prete GQ Del, Goulder P, Brumme ZL, Brumme CJ, John M, Mallal S, Nelson G, Bosch R, Heckerman D, Stein JL, Soderberg KA, Moody MA, Denny TN, Zeng X, Fang J, Moffett A, Lifson JD, Goedert JJ, Buchbinder S, Kirk GD, Fellay J, Mclaren P, Deeks SG, Pereyra F, Walker B, Michael NL, Weintrob A, Wolinsky S, Liao W, Carrington M. Influence of HLA-C expression level on HIV control. Science. 2013;87(6128):1–18.  130.  Draenert R, Allen TM, Liu Y, Wrin T, Chappey C, Verrill CL, Sirera G, Eldridge RL, Lahaie MP, Ruiz L, Clotet B, Petropoulos CJ, Walker BD, Martinez-Picado J. Constraints on HIV-1 evolution and immunodominance revealed in monozygotic adult twins infected with the same virus. J Exp Med. 2006;203(3):529–39.  131.  Feeney ME, Tang Y, Roosevelt KA, Leslie AJ, McIntosh K, Karthas N, Walker BD, Goulder PJR. Immune escape precedes breakthrough human immunodeficiency virus type 1 viremia and broadening of the cytotoxic T-lymphocyte response in an HLA-B27-positive long-term-nonprogressing child. J Virol. 2004;78(16):8927–30.  132.  Goulder PJ, Phillips RE, Colbert RA, McAdam S, Ogg G, Nowak MA, Giangrande P, Luzzi G, Morgan B, Edwards A, McMichael AJ, Rowland-Jones S. Late escape from an immunodominant cytotoxic T-lymphocyte response associated with progression to AIDS. Nat Med. 1997;3(2):212–7.  133.  Schneidewind A, Brockman MA, Yang R, Adam RI, Li B, Le Gall S, Rinaldo CR, Craggs SL, Allgaier RL, Power KA, Kuntzen T, Tung C-S, LaBute MX, Mueller SM, Harrer T, McMichael AJ, Goulder PJR, Aiken C, Brander C, Kelleher AD, Allen TM. Escape from the dominant HLA-B27-restricted cytotoxic T-lymphocyte response in Gag is associated with a dramatic reduction in human immunodeficiency virus type 1 replication. J Virol. 2007;81(22):12382–93.  134.  Brockman MA, Schneidewind A, Lahaie M, Schmidt A, Miura T, Desouza I, Ryvkin F, Derdeyn CA, Allen S, Hunter E, Mulenga J, Goepfert PA, Walker BD, Allen TM. Escape and compensation from 178  early HLA-B57-mediated cytotoxic T-lymphocyte pressure on human immunodeficiency virus type 1 Gag alter capsid interactions with cyclophilin A. J Virol. 2007;81(22):12608–18.  135.  Goepfert PA, Lumm W, Farmer P, Matthews P, Prendergast A, Carlson JM, Derdeyn CA, Tang J, Kaslow RA, Bansal A, Yusim K, Heckerman D, Mulenga J, Allen S, Goulder PJR, Hunter E. Transmission of HIV-1 Gag immune escape mutations is associated with reduced viral load in linked recipients. J Exp Med. 2008;205(5):1009–17.  136.  Prado JG, Prendergast A, Thobakgale C, Molina C, Tudor-Williams G, Ndung’u T, Walker BD, Goulder P. Replicative capacity of human immunodeficiency virus type 1 transmitted from mother to child is associated with pediatric disease progression rate. J Virol. 2010;84(1):492–502.  137.  Schneidewind A, Tang Y, Brockman MA, Ryland EG, Dunkley-Thompson J, Steel-Duncan JC, St John MA, Conrad JA, Kalams SA, Noel F, Allen TM, Christie CD, Feeney ME. Maternal transmission of human immunodeficiency virus escape mutations subverts HLA-B57 immunodominance but facilitates viral control in the haploidentical infant. J Virol. 2009;83(17):8616–27.  138.  Lavanchy D. The global burden of hepatitis C. Liver Int. 2009;29(Suppl. 1):74–81.  139.  Remis RS. Modelling the Incidence and Prevalence of Hepatitis C Infection and its Sequelae in Canada, 2007. Public Health Agency of Canada; 2007.  140.  The Global Burden of Hepatitis C Working Group. Global burden of disease (GBD) for hepatitis C. J Clin Pharmacol. 2004;44(1):20–9.  141.  Perz JF, Armstrong GL, Farrington LA, Hutin YJF, Bell BP. The contributions of hepatitis B virus and hepatitis C virus infections to cirrhosis and primary liver cancer worldwide. J Hepatol. 2006;45(4):529–38.  142.  Nelson PK, Mathers BM, Cowie B, Hagan H, Des Jarlais D, Horyniak D, Degenhardt L. Global epidemiology of hepatitis B and hepatitis C in people who inject drugs: results of systematic reviews. Lancet. 2011;378(9791):571–83.  143.  Soriano V, Vispo E, Labarga P, Medrano J, Barreiro P. Viral hepatitis and HIV co-infection. Antiviral Res. 2010;85(1):303–15.  144.  Alter MJ. Epidemiology of viral hepatitis and HIV co-infection. J Hepatol. 2006;44(SUPPL. 1):6–9.  145.  Rao VB, Johari N, du Cros P, Messina J, Ford N, Cooke GS. Hepatitis C seroprevalence and HIV co-infection in sub-Saharan Africa: a systematic review and meta-analysis. Lancet Infect Dis. 2015;3099(15):6–7.  146.  Terrault NA, Dodge JL, Murphy EL, Tavis JE, Kiss A, Levin TR, Gish RG, Busch MP, Reingold AL, Alter MJ. Sexual transmission of hepatitis C virus among monogamous heterosexual couples: the HCV partners study. Hepatology. 2013;57(3):881–9.  147.  Hahn JA. Sex, drugs, and hepatitis C virus. J Infect Dis. 2007;195(11):1556–9.  148.  Urbanus AT, van de Laar TJ, Stolte IG, Schinkel J, Heijman T, Coutinho RA, Prins M. Hepatitis C virus infections among HIV-infected men who have sex with men: an expanding epidemic. AIDS. 2009;23(12):F1–7.  179  149.  Alter HJ, Purcell RH, Holland PV, Alling DW, Koziol DE. Donor transaminase and recipient hepatitis. Impact on blood transfusion services. JAMA. 1981;246(6):630–4.  150.  Stevens CE, Aach RD, Hollinger FB, Mosley JW, Szmuness W, Kahn R, Werch J, Edwards V. Hepatitis B virus antibody in blood donors and the occurrence of non-A, non-B hepatitis in transfusion recipients. An analysis of the transfusion-transmitted viruses study. Ann Intern Med. 1984;101(6):733–8.  151.  Tobler LH, Busch MP. History of posttransfusion hepatitis. Clin Chem. 1997;43(8 Pt. 2):1487–93.  152.  Krever H. Final Report: Commission of Inquiry on the Blood System in Canada. [Internet]. 1997 [cited 2015 May 7]. p. 708–18. Available from: http://publications.gc.ca/pub?id=9.646740&sl=0 153.  Trubnikov M, Yan P, Njihia J, Archibald C. Identifying and describing a cohort effect in the national database of reported cases of hepatitis C virus infection in Canada (1991-2010): an age-period-cohort analysis. CMAJ Open. 2014;2(4):E281–7.  154.  Frank C, Mohamed MK, Strickland GT, Lavanchy D, Arthur RR, Magder LS, El Khoby T, Abdel-Wahab Y, Aly Ohn ES, Anwar W, Sallam I. The role of parenteral antischistosomal therapy in the spread of hepatitis C virus in Egypt. Lancet. 2000;355(9207):887–91.  155.  Rao MR, Naficy AB, Darwish MA, Darwish NM, Schisterman E, Clemens JD, Edelman R. Further evidence for association of hepatitis C infection with parenteral schistosomiasis treatment in Egypt. BMC Infect Dis. 2002;2:29.  156.  Pybus OG, Drummond AJ, Nakano T, Robertson BH, Rambaut A. The epidemiology and iatrogenic transmission of hepatitis C virus in egypt: A Bayesian coalescent approach. Mol Biol Evol. 2003;20(3):381–7.  157.  Maegraith BG. Treatment of bilharziasis in Egypt, UAR. Geneva, Switzerland; 1964.  158.  Lindenbach BD, Rice CM. The ins and outs of hepatitis C virus entry and assembly. Nat Rev Microbiol. 2013;11(10):688–700.  159.  Bartenschlager R, Lohmann V, Penin F. The molecular and structural basis of advanced antiviral therapy for hepatitis C virus infection. Nat Rev Microbiol. 2013;11(7):482–96.  160.  Moradpour D, Penin F, Rice CM. Replication of hepatitis C virus. Nat Rev Microbiol. 2007;5(6):453–63.  161.  Micallef JM, Kaldor JM, Dore GJ. Spontaneous viral clearance following acute hepatitis C infection: a systematic review of longitudinal studies. J Viral Hepat. 2006;13(1):34–41.  162.  Mehta SH, Cox A, Hoover DR, Wang XH, Mao Q, Ray S, Strathdee SA, Vlahov D, Thomas DL. Protection against persistence of hepatitis C. Lancet. 2002;359(9316):1478–83.  163.  Bassett SE, Guerra B, Brasky K, Miskovsky E, Houghton M, Klimpel GR, Lanford RE. Protective immune response to hepatitis C virus in chimpanzees rechallenged following clearance of primary infection. Hepatology. 2001;33(6):1479–87.  164.  Thomas DL, Thio CL, Martin MP, Qi Y, Ge D, O’Huigin C, Kidd J, Kidd K, Khakoo SI, Alexander G, Goedert JJ, Kirk GD, Donfield SM, Rosen HR, Tobler LH, Busch MP, McHutchison JG, Goldstein 180  DB, Carrington M. Genetic variation in IL28B and spontaneous clearance of hepatitis C virus. Nature. 2009;461(7265):798–801.  165.  O’Brien SJ, Gao X, Carrington M. HLA and AIDS: a cautionary tale. Trends Mol Med. 2001;7(9):379–81.  166.  McKiernan SM, Hagan R, Curry M, McDonald GSA, Kelly A, Nolan N, Walsh A, Hegarty J, Lawlor E, Kelleher D. Distinct MHC class I and II alleles are associated with hepatitis C viral clearance, originating from a single source. Hepatology. 2004;40(1):108–14.  167.  Kim AY, Kuntzen T, Timm J, Nolan BE, Baca MA, Reyor LL, Berical AC, Feller AJ, Johnson KL, Schulze zur Wiesch J, Robbins GK, Chung RT, Walker BD, Carrington M, Allen TM, Lauer GM. Spontaneous control of HCV is associated with expression of HLA-B *57 and preservation of targeted epitopes. Gastroenterology. 2011;140(2):686–96.  168.  Lange CM, Roomp K, Dragan A, Nattermann J, Michalk M, Spengler U, Weich V, Lengauer T, Zeuzem S, Berg T, Sarrazin C. HLA class I allele associations with HCV genetic variants in patients with chronic HCV genotypes 1a or 1b infection. J Hepatol. 2010;53(6):1022–8.  169.  Kuntzen T, Timm J, Berical A, Lewis-Ximenez LL, Jones A, Nolan B, Schulze zur Wiesch J, Li B, Schneidewind A, Kim AY, Chung RT, Lauer GM, Allen TM. Viral sequence evolution in acute hepatitis C virus infection. J Virol. 2007;81(21):11658–68.  170.  Smith DB, Bukh J, Kuiken C, Muerhoff AS, Rice CM, Stapleton JT, Simmonds P. Expanded classification of hepatitis C virus into 7 genotypes and 67 subtypes: updated criteria and genotype assignment web resource. Hepatology. 2014;59(1):318–27.  171.  Chaudhary RK, Tepper M, Eisaadany S, Gully PR. Distribution of hepatitis C virus genotypes in Canada: results from the LCDC Sentinel Health Unit Surveillance System. Can J Infect Dis. 1999;10(1):53–6.  172.  Messina JP, Humphreys I, Flaxman A, Brown A, Cooke GS, Pybus OG, Barnes E. Global distribution and prevalence of hepatitis C virus genotypes. Hepatology. 2015;61(1):77–87.  173.  Selik RM, Chu SY, Buehler JW. HIV infection as leading cause of death among young adults in US cities and states. JAMA. 1993;269(23):2991–4.  174.  Samji H, Cescon A, Hogg RS, Modur SP, Althoff KN, Buchacz K, Burchell AN, Cohen M, Gebo KA, Gill MJ, Justice A, Kirk G, Klein MB, Korthuis PT, Martin J, Napravnik S, Rourke SB, Sterling TR, Silverberg MJ, Deeks S, Jacobson LP, Bosch RJ, Kitahata MM, Goedert JJ, Moore R, Gange SJ. Closing the gap: increases in life expectancy among treated HIV-positive individuals in the United States and Canada. PLoS One. 2013;8(12):e81355.  175.  Wada N, Jacobson LP, Cohen M, French A, Phair J, Muñoz A. Cause-specific mortality among HIV-infected individuals, by CD4(+) cell count at HAART initiation, compared with HIV-uninfected individuals. AIDS. 2014;28(2):257–65.  176.  Fischl MA, Richman DD, Grieco MH, Gottlieb MS, Volberding PA, Laskin OL, Leedom JM, Groopman JE, Mildvan D, Schooley RT. The efficacy of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex. A double-blind, placebo-controlled trial. N Engl J Med. 1987;317(4):185–91.  177.  Volberding PA, Lagakos SW, Koch MA, Pettinelli C, Myers MW, Booth DK, Balfour HH, Reichman RC, Bartlett JA, Hirsch MS. Zidovudine in asymptomatic human immunodeficiency virus 181  infection. A controlled trial in persons with fewer than 500 CD4-positive cells per cubic millimeter. The AIDS Clinical Trials Group of the National Institute of Allergy and Infectious Diseases. N Engl J Med. 1990;322(14):941–9.  178.  Mulder JW, Cooper DA, Mathiesen L, Sandström E, Clumeck N, Gatell JM, French M, Donovan B, Gray F, Yeo JM, Lange JMA, The European-Australian Collaborative Group (Study 017). Zidovudine twice daily in asymptomatic subjects with HIV infection and a high risk of progression to AIDS: a randomized, double-blind placebo-controlled study. AIDS. 1994;8(3):313–21.  179.  Vella S, Giuliano M, Pezzotti P, Agresti MG, Tomino C, Floridia M, Greco D, Moroni M, Visco G, Milazzo F. Survival of zidovudine-treated patients with AIDS compared with that of contemporary untreated patients. Italian Zidovudine Evaluation Group. JAMA. 1992;267(9):1232–6.  180.  Hammer SM, Katzenstein DA, Hughes MD, Gundacker H, Schooley RT, Haubrich RH, Henry WK, Lederman MM, Phair JP, Niu M, Hirsch MS, Merigan TC, the AIDS Clinical Trials Group Study 175 Study Team. A trial comparing nucleoside monotherapy with combination therapy in HIV-infected adults with CD4 cell counts from 200 to 500 per cubic millimeter. N Engl J Med. 1996;335(15):1081–90.  181.  Collier AC, Coombs RW, Fischl MA, Skolnik PR, Northfelt D, Boutin P, Hooper CJ, Kaplan LD, Volberding PA, Davis LG, Henrard DR, Weller S, Corey L. Combination therapy with zidovudine and didanosine compared with zidovudine alone in HIV-1 infection. Ann Intern Med. 1993;119(8):786–93.  182.  Rooke R, Tremblay M, Soudeyns H, DeStephano L, Yao XJ, Fanning M, Montaner JS, O’Shaughnessy M, Gelmon K, Tsoukas C, the Canadian Zidovudine Multi-Centre Study Group. Isolation of drug-resistant variants of HIV-1 from patients on long-term zidovudine therapy. AIDS. 1989;3(7):411–5.  183.  Kozal MJ, Kroodsma K, Winters MA, Shafer RW, Efron B, Katzenstein DA, Merigan TC. Didanosine resistance in HIV-infected patients switched from zidovudine to didanosine monotherapy. Ann Intern Med. 1994;121(4):263–8.  184.  Shirasaka T, Kavlick MF, Ueno T, Gao WY, Kojima E, Alcaide ML, Chokekijchai S, Roy BM, Arnold E, Yarchoan R. Emergence of human immunodeficiency virus type 1 variants with resistance to multiple dideoxynucleosides in patients receiving therapy with dideoxynucleosides. Proc Natl Acad Sci U S A. 1995;92(6):2398–402.  185.  Shafer RW, Iversen AK, Winters MA, Aguiniga E, Katzenstein DA, Merigan TC, the AIDS Clinical Trials Group 143 Virology Team. Drug resistance and heterogeneous long-term virologic responses of human immunodeficiency virus type 1-infected subjects to zidovudine and didanosine combination therapy. J Infect Dis. 1995;172(1):70–8.  186.  D’Aquila RT, Hughes MD, Johnson VA, Fischl MA, Sommadossi J-P, Liou SH, Timpone J, Myers M, Basgoz N, Niu M, Hirsch MS, The National Institute of Allergy and Infectious Diseases AIDS Clinical Trials Group Protocol 214 Investigators. Nevirapine, zidovudine, and didanosine compared with zidovudine and didanosine in patients with HIV-1 infection. A randomized, double-blind, placebo-controlled trial. Ann Intern Med. 1996;124(12):1019–30.  187.  Hammer SM, Squires KE, Hughes MD, Grimes JM, Demeter LM, Currier JS, Eron JJ, Feinberg JE, Balfour HH, Deyton LR, Chodakewitz JA, Fischl MA, the AIDS Clinical Trials Group 320 Study Team. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less. N Engl J Med. 1997;337(11):725–33.  182  188.  Montaner JS, Reiss P, Cooper D, Vella S, Harris M, Conway B, Wainberg MA, Smith D, Robinson P, Hall D, Myers M, Lange JMA, the INCAS Study Group. A randomized, double-blind trial comparing combinations of nevirapine, didanosine, and zidovudine for HIV-infected patients: the INCAS Trial. JAMA. 1998;279(12):930–7.  189.  Palella FJ, Delaney KM, Moorman AC, Loveless MO, Fuhrer J, Satten GA, Aschman DJ, Holmberg SD, the HIV Outpatient Study Investigators. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. N Engl J Med. 1998;338(13):853–60.  190.  Hogg RS, Heath KV, Yip B, Craib KJ, O’Shaughnessy MV, Schechter MT, Montaner JS. Improved survival among HIV-infected individuals following initiation of antiretroviral therapy. JAMA. 1998;279(6):450–4.  191.  Montaner JSG, Guillemi S, Harris M, BC Centre for Excellence in HIV/AIDS Therapeutic Guidelines Committee. Therapeutic Guidelines: Antiretroviral (ARV) Treatment of Adult HIV Infection [Internet]. 2013 [cited 2015 Jan 30]. Available from: http://www.cfenet.ubc.ca/sites/default/files/uploads/Therapeutic Guidelines 2013-Feb-final.pdf 192.  Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents. February 5, 2001 [Internet]. 2012 [cited 2015 Jan 16]. Available from: http://aidsinfo.nih.gov/contentfiles/lvguidelines/adultandadolescentgl.pdf 193.  Richman DD. HIV chemotherapy. Nature. 2001;410(6831):995–1001.  194.  Pomerantz RJ, Horn DL. Twenty years of therapy for HIV-1 infection. Nat Med. 2003;9(7):867–73.  195.  Zeldin RK, Petruschke RA. Pharmacological and therapeutic properties of ritonavir-boosted protease inhibitor therapy in HIV-infected patients. J Antimicrob Chemother. 2004;53(1):4–9.  196.  Hicks C, Gulick RM. Raltegravir: the first HIV type 1 integrase inhibitor. Clin Infect Dis. 2009;48(7):931–9.  197.  Dorr P, Westby M, Dobbs S, Griffin P, Irvine B, Macartney M, Mori J, Rickett G, Smith-Burchnell C, Napier C, Webster R, Armour D, Price D, Stammen B, Wood A, Perros M. Maraviroc (UK-427,857), a potent, orally bioavailable, and selective small-molecule inhibitor of chemokine receptor CCR5 with broad-spectrum anti-human immunodeficiency virus type 1 activity. Antimicrob Agents Chemother. 2005;49(11):4721–32.  198.  Kohli A, Shaffer A, Sherman A, Kottilil S. Treatment of Hepatitis C. JAMA. 2014;312(6):631.  199.  Swain MG, Lai M, Shiffman ML, Cooksley WGE, Zeuzem S, Dieterich DT, Abergel A, Pessôa MG, Lin A, Tietz A, Connell EV, Diago M. A sustained virologic response is durable in patients with chronic hepatitis C treated with peginterferon Alfa-2a and ribavirin. Gastroenterology. 2010;139(5):1593–601.  200.  Chen J, Florian J, Carter W, Fleischer RD, Hammerstrom TS, Jadhav PR, Zeng W, Murray J, Birnkrant D. Earlier sustained virologic response end points for regulatory approval and dose selection of hepatitis C therapies. Gastroenterology. 2013;144(7):1450–5.e2.  201.  Yoshida EM, Sulkowski MS, Gane EJ, Herring RW, Ratziu V, Ding X, Wang J, Chuang S-M, Ma J, McNally J, Stamm LM, Brainard DM, Symonds WT, McHutchison JG, Beavers KL, Jacobson IM, 183  Reddy KR, Lawitz E. Concordance of sustained virological response 4, 12, and 24 weeks post-treatment with sofosbuvir-containing regimens for hepatitis C virus. Hepatology. 2015;61(1):41–5.  202.  Di Bisceglie AM, Martin P, Kassianides C, Lisker-Melman M, Murray L, Waggoner J, Goodman Z, Banks SM, Hoofnagle JH. Recombinant interferon alfa therapy for chronic hepatitis C. A randomized, double-blind, placebo-controlled trial. N Engl J Med. 1989;321(22):1506–10.  203.  Di Bisceglie AM, Hoofnagle JH. Optimal therapy of hepatitis C. Hepatology. 2002;36(5 Suppl 1):S121–7.  204.  Dusheiko G, Main J, Thomas H, Reichard O, Lee C, Dhillon A, Rassam S, Fryden A, Reesink H, Bassendine M, Norkrans G, Cuypers T, Lelie N, Telfer P, Watson J, Weegink C, Sillikens P, Weiland O. Ribavirin treatment for patients with chronic hepatitis C: results of a placebo-controlled study. J Hepatol. 1996;25(5):591–8.  205.  Poynard T, Marcellin P, Lee SS, Niederau C, Minuk GS, Ideo G, Bain V, Heathcote J, Zeuzem S, Trepo C, Albrecht J. Randomised trial of interferon alpha2b plus ribavirin for 48 weeks or for 24 weeks versus interferon alpha2b plus placebo for 48 weeks for treatment of chronic infection with hepatitis C virus. International Hepatitis Interventional Therapy Group (IHIT). Lancet. 1998;352(9138):1426–32.  206.  Zeuzem S, Feinman SV, Rasenack J, Heathcote EJ, Lai MY, Gane E, O’Grady J, Reichen J, Diago M, Lin A, Hoffman J, Brunda MJ. Peginterferon alfa-2a in patients with chronic hepatitis C. N Engl J Med. 2000;343(23):1666–72.  207.  Manns MP, McHutchison JG, Gordon SC, Rustgi VK, Shiffman M, Reindollar R, Goodman ZD, Koury K, Ling M, Albrecht JK. Peginterferon alfa-2b plus ribavirin compared with interferon alfa-2b plus ribavirin for initial treatment of chronic hepatitis C: a randomised trial. Lancet. 2001;358(9286):958–65.  208.  Fried MW, Shiffman ML, Reddy KR, Smith C, Marinos G, Gonçales FL, Häussinger D, Diago M, Carosi G, Dhumeaux D, Craxì A, Lin A, Hoffman J, Yu J. Peginterferon alfa-2a plus ribavirin for chronic hepatitis C virus infection. N Engl J Med. 2002;347(13):975–82.  209.  Hadziyannis SJ, Sette H, Morgan TR, Balan V, Diago M, Marcellin P, Ramadori G, Bodenheimer H, Bernstein D, Rizzetto M, Zeuzem S, Pockros PJ, Lin A, Ackrill AM. Peginterferon-α2a and ribavirin combination therapy in chronic hepatitis C: a randomized study of treatment duration and ribavirin dose. Ann Intern Med. 2004;140(5):346–55.  210.  Poordad F, McCone J, Bacon BR, Bruno S, Manns MP, Sulkowski MS, Jacobson IM, Reddy KR, Goodman ZD, Boparai N, DiNubile MJ, Sniukiene V, Brass CA, Albrecht JK, Bronowicki J-P. Boceprevir for untreated chronic HCV genotype 1 infection. N Engl J Med. 2011;364(13):1195–206.  211.  Bacon BR, Gordon SC, Lawitz E, Marcellin P, Vierling JM, Zeuzem S, Poordad F, Goodman ZD, Sings HL, Boparai N, Burroughs M, Brass CA, Albrecht JK, Esteban R. Boceprevir for previously treated chronic HCV genotype 1 infection. N Engl J Med. 2011;364(13):1207–17.  212.  Jacobson IM, McHutchison JG, Dusheiko G, Di Bisceglie AM, Reddy KR, Bzowej NH, Marcellin P, Muir AJ, Ferenci P, Flisiak R, George J, Rizzetto M, Shouval D, Sola R, Terg RA, Yoshida EM, Adda N, Bengtsson L, Sankoh AJ, Kieffer TL, George S, Kauffman RS, Zeuzem S. Telaprevir for previously untreated chronic hepatitis C virus infection. N Engl J Med. 2011;364(25):2405–16.  184  213.  Zeuzem S, Andreone P, Pol S, Lawitz E, Diago M, Roberts S, Focaccia R, Younossi Z, Foster GR, Horban A, Ferenci P, Nevens F, Müllhaupt B, Pockros P, Terg R, Shouval D, van Hoek B, Weiland O, Van Heeswijk R, De Meyer S, Luo D, Boogaerts G, Polo R, Picchio G, Beumont M. Telaprevir for retreatment of HCV infection. N Engl J Med. 2011;364(25):2417–28.  214.  Manns M, Marcellin P, Poordad F, de Araujo ESA, Buti M, Horsmans Y, Janczewska E, Villamil F, Scott J, Peeters M, Lenz O, Ouwerkerk-Mahadevan S, De La Rosa G, Kalmeijer R, Sinha R, Beumont-Mauviel M. Simeprevir with pegylated interferon alfa 2a or 2b plus ribavirin in treatment-naive patients with chronic hepatitis C virus genotype 1 infection (QUEST-2): a randomised, double-blind, placebo-controlled phase 3 trial. Lancet. 2014;384(9941):414–26.  215.  Jacobson IM, Dore GJ, Foster GR, Fried MW, Radu M, Rafalsky VV, Moroz L, Craxì A, Peeters M, Lenz O, Ouwerkerk-Mahadevan S, De La Rosa G, Kalmeijer R, Scott J, Sinha R, Beumont-Mauviel M. Simeprevir with pegylated interferon alfa 2a plus ribavirin in treatment-naive patients with chronic hepatitis C virus genotype 1 infection (QUEST-1): a phase 3, randomised, double-blind, placebo-controlled trial. Lancet. 2014;384(9941):403–13.  216.  Lawitz E, Mangia A, Wyles D, Rodriguez-Torres M, Hassanein T, Gordon SC, Schultz M, Davis MN, Kayali Z, Reddy KR, Jacobson IM, Kowdley KV, Nyberg L, Subramanian GM, Hyland RH, Arterburn S, Jiang D, McNally J, Brainard D, Symonds WT, McHutchison JG, Sheikh AM, Younossi Z, Gane EJ. Sofosbuvir for previously untreated chronic hepatitis C infection. N Engl J Med. 2013;368(20):1878–87.  217.  Gane EJ, Stedman CA, Hyland RH, Ding X, Svarovskaia E, Symonds WT, Hindes RG, Berrey MM. Nucleotide polymerase inhibitor sofosbuvir plus ribavirin for hepatitis C. N Engl J Med. 2013;368(1):34–44.  218.  Sulkowski MS, Naggie S, Lalezari J, Fessel WJ, Mounzer K, Shuhart M, Luetkemeyer AF, Asmuth D, Gaggar A, Ni L, Svarovskaia E, Brainard DM, Symonds WT, Subramanian GM, McHutchison JG, Rodriguez-Torres M, Dieterich D. Sofosbuvir and ribavirin for hepatitis C in patients with HIV coinfection. JAMA. 2014;312(4):353–61.  219.  Kowdley KV, Gordon SC, Reddy KR, Rossaro L, Bernstein DE, Lawitz E, Shiffman ML, Schiff E, Ghalib R, Ryan M, Rustgi V, Chojkier M, Herring R, Di Bisceglie AM, Pockros PJ, Subramanian GM, An D, Svarovskaia E, Hyland RH, Pang PS, Symonds WT, McHutchison JG, Muir AJ, Pound D, Fried MW. Ledipasvir and sofosbuvir for 8 or 12 weeks for chronic HCV without cirrhosis. N Engl J Med. 2014;370(20):1879–88.  220.  Gane EJ, Stedman CA, Hyland RH, Ding X, Svarovskaia E, Subramanian GM, Symonds WT, McHutchison JG, Pang PS. Efficacy of nucleotide polymerase inhibitor sofosbuvir plus the NS5A inhibitor ledipasvir or the NS5B non-nucleoside inhibitor GS-9669 against HCV genotype 1 infection. Gastroenterology. 2014;146(3):736–43.e1.  221.  European Association for the Study of the Liver. EASL Clinical Practice Guidelines: management of hepatitis C virus infection. J Hepatol. 2011;55(2):245–64.  222.  Hepburn MJ, Hepburn LM, Cantu NS, Lapeer MG, Lawitz EJ. Differences in treatment outcome for hepatitis C among ethnic groups. Am J Med. 2004;117(3):163–8.  223.  Jensen DM, Morgan TR, Marcellin P, Pockros PJ, Reddy KR, Hadziyannis SJ, Ferenci P, Ackrill AM, Willems B. Early identification of HCV genotype 1 patients responding to 24 weeks peginterferon α-2a (40 kd)/ribavirin therapy. Hepatology. 2006;43(5):954–60.  185  224.  Davis GL, Wong JB, McHutchison JG, Manns MP, Harvey J, Albrecht J. Early virologic response to treatment with peginterferon alfa-2b plus ribavirin in patients with chronic hepatitis C. Hepatology. 2003;38(3):645–52.  225.  Ferenci P, Fried MW, Shiffman ML, Smith CI, Marinos G, Gonçales FL, Häussinger D, Diago M, Carosi G, Dhumeaux D, Craxì A, Chaneac M, Reddy KR. Predicting sustained virological responses in chronic hepatitis C patients treated with peginterferon alfa-2a (40 KD)/ribavirin. J Hepatol. 2005;43(3):425–33.  226.  Volk ML, Tocco R, Saini S, Lok ASF. Public health impact of antiviral therapy for hepatitis C in the United States. Hepatology. 2009;50(6):1750–5.  227.  Lettmeier B, Mühlberger N, Schwarzer R, Sroczynski G, Wright D, Zeuzem S, Siebert U. Market uptake of new antiviral drugs for the treatment of hepatitis C. J Hepatol. 2008;49(4):528–36.  228.  Cotler SJ, Patil R, McNutt RA, Speroff T, Banaad-Omiotek G, Ganger DR, Rosenblate H, Kaur S, Cotler S, Jensen DM. Patients’ values for health states associated with hepatitis C and physicians' estimates of those values. Am J Gastroenterol. 2001;96(9):2730–6.  229.  Stein MD, Maksad J, Clarke J. Hepatitis C disease among injection drug users: Knowledge, perceived risk and willingness to receive treatment. Drug Alcohol Depend. 2001;61(3):211–5.  230.  Grebely J, Genoway KA, Raffa JD, Dhadwal G, Rajan T, Showler G, Kalousek K, Duncan F, Tyndall MW, Fraser C, Conway B, Fischer B. Barriers associated with the treatment of hepatitis C virus infection among illicit drug users. Drug Alcohol Depend. 2008;93(1-2):141–7.  231.  Edlin BR, Kresina TF, Raymond DB, Carden MR, Gourevitch MN, Rich JD, Cheever LW, Cargill VA. Overcoming barriers to prevention, care, and treatment of hepatitis C in illicit drug users. Clin Infect Dis. 2005;40 Suppl 5(Suppl 5):S276–85.  232.  Thompson V V, Ragland KE, Hall CS, Morgan M, Bangsberg DR. Provider assessment of eligibility for hepatitis C treatment in HIV-infected homeless and marginally housed persons. AIDS. 2005;19 Suppl 3:S208–14.  233.  Lohmann V, Bartenschlager R. On the history of hepatitis C virus cell culture systems. J Med Chem. 2014;57(5):1627–42.  234.  Kieffer TL, De Meyer S, Bartels DJ, Sullivan JC, Zhang EZ, Tigges A, Dierynck I, Spanks J, Dorrian J, Jiang M, Adiwijaya B, Ghys A, Beumont M, Kauffman RS, Adda N, Jacobson IM, Sherman KE, Zeuzem S, Kwong AD, Picchio G. Hepatitis C viral evolution in genotype 1 treatment-naïve and treatment-experienced patients receiving telaprevir-based therapy in clinical trials. PLoS One. 2012;7(4):e34372.  235.  Barnard RJO, Howe JA, Ogert RA, Zeuzem S, Poordad F, Gordon SC, Ralston R, Tong X, Sniukiene V, Strizki J, Ryan D, Long J, Qiu P, Brass CA, Albrecht J, Burroughs M, Vuocolo S, Hazuda DJ. Analysis of boceprevir resistance associated amino acid variants (RAVs) in two phase 3 boceprevir clinical studies. Virology. 2013;444(1-2):329–36.  236.  Feld JJ, Kowdley KV, Coakley E, Sigal S, Nelson DR, Crawford D, Weiland O, Aguilar H, Xiong J, Pilot-Matias T, DaSilva-Tillmann B, Larsen L, Podsadecki T, Bernstein B. Treatment of HCV with ABT-450/r-ombitasvir and dasabuvir with ribavirin. N Engl J Med. 2014;370(17):1594–603.  237.  Zeuzem S, Jacobson IM, Baykal T, Marinho RT, Poordad F, Bourlière M, Sulkowski MS, Wedemeyer H, Tam E, Desmond P, Jensen DM, Di Bisceglie AM, Varunok P, Hassanein T, Xiong 186  J, Pilot-Matias T, DaSilva-Tillmann B, Larsen L, Podsadecki T, Bernstein B. Retreatment of HCV with ABT-450/r–ombitasvir and dasabuvir with ribavirin. N Engl J Med. 2014;370(17):1604–14.  238.  Peto T. Surrogate markers in HIV disease. J Antimicrob Chemother. 1996;37(Suppl B):161–70.  239.  Teitel JM, Freedman JJ, Garvey MB, Kardish M. Two-year evaluation of clinical and laboratory variables of immune function in 117 hemophiliacs seropositive or seronegative for HIV-1. Am J Hematol. 1989;32(4):262–72.  240.  Eyster ME, Ballard JO, Gail MH, Drummond JE, Goedert JJ. Predictive markers for the acquired immunodeficiency syndrome (AIDS) in hemophiliacs: persistence of p24 antigen and low T4 cell count. Ann Intern Med. 1989;110(12):963–9.  241.  Cooley TP, Kunches LM, Saunders CA, Ritter JK, Perkins CJ, McLaren C, McCaffrey RP, Liebman HA. Once-daily administration of 2’,3'-dideoxyinosine (ddI) in patients with the acquired immunodeficiency syndrome or AIDS-related complex. Results of a Phase I trial. N Engl J Med. 1990;322(19):1340–5.  242.  Lambert JS, Seidlin M, Reichman RC, Plank CS, Laverty M, Morse GD, Knupp C, McLaren C, Pettinelli C, Valentine FT. 2’,3'-dideoxyinosine (ddI) in patients with the acquired immunodeficiency syndrome or AIDS-related complex. A phase I trial. N Engl J Med. 1990;322(19):1333–40.  243.  Darbyshire JH, Aboulker JP. Didanosine for zidovudine-intolerant patients with HIV disease. Lancet. 1992;340(8831):1346–7.  244.  Seligmann M, Warrell DA, Aboulker JP, Carbon C, Darbyshire JH, Dormont J, Eschwege E, Girling DJ, James DR, Levy JP, Peto TEA, Schwarz D, Stone AB, Weller IVD, Withnall R, Gelmon K, Lafon E, Swart AM, Aber VR, Babiker AG, Lhoro S, Nunn AJ, Vray M. Concorde: MRC/ANRS randomised double-blind controlled trial of immediate and deferred zidovudine in symptom-free HIV infection. Lancet. 1994;343(8902):871–81.  245.  Hughes MD, Daniels MJ, Fischl MA, Kim S, Schooley RT. CD4 cell count as a surrogate endpoint in HIV clinical trials: a meta-analysis of studies of the AIDS Clinical Trials Group. AIDS. 1998;12(14):1823–32.  246.  Mulder J, McKinney N, Christopherson C, Sninsky J, Greenfield L, Kwok S. Rapid and simple PCR assay for quantitation of human immunodeficiency virus type 1 RNA in plasma: application to acute retroviral infection. J Clin Microbiol. 1994;32(2):292–300.  247.  Mulder J, Resnick R, Saget B, Scheibel S, Herman S, Payne H, Harrigan R, Kwok S. A rapid and simple method for extracting human immunodeficiency virus type 1 RNA from plasma: enhanced sensitivity. J Clin Microbiol. 1997;35(5):1278–80.  248.  Kievits T, van Gemen B, van Strijp D, Schukkink R, Dircks M, Adriaanse H, Malek L, Sooknanan R, Lens P. NASBA isothermal enzymatic in vitro nucleic acid amplification optimized for the diagnosis of HIV-1 infection. J Virol Methods. 1991;35(3):273–86.  249.  Todd J, Pachl C, White R, Yeghiazarian T, Johnson P, Taylor B, Holodniy M, Kern D, Hamren S, Chernoff D. Performance characteristics for the quantitation of plasma HIV-1 RNA using branched DNA signal amplification technology. J Acquir Immune Defic Syndr Hum Retrovirol. 1995;10(Suppl 2):S35–44.  187  250.  Kern D, Collins M, Fultz T, Detmer J, Hamren S, Peterkin JJ, Sheridan P, Urdea M, White R, Yeghiazarian T, Todd J. An enhanced-sensitivity branched-DNA assay for quantification of human immunodeficiency virus type 1 RNA in plasma. J Clin Microbiol. 1996;34(12):3196–202.  251.  Mellors JW, Kingsley LA, Rinaldo CR, Todd JA, Hoo BS, Kokka RP, Gupta P. Quantitation of HIV-1 RNA in plasma predicts outcome after seroconversion. Ann Intern Med. 1995;122(8):573–9.  252.  Mellors JW, Muñoz A, Giorgi JV, Margolick JB, Tassoni CJ, Gupta P, Kingsley LA, Todd JA, Saah AJ, Detels R, Phair JP, Rinaldo CR. Plasma viral load and CD4+ lymphocytes as prognostic markers of HIV-1 infection. Ann Intern Med. 1997;126(12):946–54.  253.  De Wolf F, Spijkerman I, Schellekens PT, Langendam M, Kuiken C, Bakker M, Roos M, Coutinho R, Miedema F, Goudsmit J. AIDS prognosis based on HIV-1 RNA, CD4+ T-cell count and function: markers with reciprocal predictive value over time after seroconversion. AIDS. 1997;11(15):1799–806.  254.  Katzenstein DA, Hammer SM, Hughes MD, Gundacker H, Jackson JB, Fiscus S, Rasheed S, Elbeik T, Reichman R, Japour A, Merigan TC, Hirsch MS, the AIDS Clinical Trials Group Study 175 Virology Study Team. The relation of virologic and immunologic markers to clinical outcomes after nucleoside therapy in HIV-infected adults with 200 to 500 CD4 cells per cubic millimeter. N Engl J Med. 1996;335(15):1091–8.  255.  Marschner IC, Collier AC, Coombs RW, D’Aquila RT, DeGruttola V, Fischl MA, Hammer SM, Hughes MD, Johnson VA, Katzenstein DA, Richman DD, Smeaton LM, Spector SA, Saag MS. Use of changes in plasma levels of human immunodeficiency virus type 1 RNA to assess the clinical benefit of antiretroviral therapy. J Infect Dis. 1998;177(1):40–7.  256.  Egger M, May M, Chêne G, Phillips AN, Ledergerber B, Dabis F, Costagliola D, D’Arminio Monforte A, de Wolf F, Reiss P, Lundgren JD, Justice AC, Staszewski S, Leport C, Hogg RS, Sabin CA, Gill MJ, Salzberger B, Sterne JAC. Prognosis of HIV-1-infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet. 2002;360(9327):119–29.  257.  Chêne G, Sterne JAC, May M, Costagliola D, Ledergerber B, Phillips AN, Dabis F, Lundgren J, D’Arminio Monforte A, de Wolf F, Hogg R, Reiss P, Justice A, Leport C, Staszewski S, Gill J, Fatkenheuer G, Egger ME. Prognostic importance of initial response in HIV-1 infected patients starting potent antiretroviral therapy: analysis of prospective studies. Lancet. 2003;362(9385):679–86.  258.  Ledergerber B, Lundgren JD, Sarah Walker A, Sabin C, Justice A, Reiss P, Mussini C, Wit F, d’Arminio Monforte A, Weber R, Fusco G, Staszewski S, Law M, Hogg R, Lampe F, John Gill M, Castelli F, Phillips AN, Rooney P, Taylor S, Couldwell D, Austin D, Block M, Clemons J, Finlayson R, Petoumenos K, Quan D, Smith D, O’Connor C, Gorton C, Allen D, Mulhall B, Mutimer K, Keeffe N, Cooper D, Carr A, Miller J, Pell C, Ellis D, Baker D, Kidd J, McFarlane R, Liang MT, Brown K, Huffam S, Savage J, Morgan S, Knibbs P, Sowden D, Walker A, Orth D, Lister G, Chuah J, Fankhauser W, Dickson B, Bradford D, Wilson C, Ree H, Magon H, Anderson J, Moore R, Russell D, McGovern G, McNair R, Bal J, Fairley K, Roth N, Eu B, Strecker S, Wood H, Mijch A, Hoy J, Pierce A, McCormack C, Watson K, Medland N, Daye J, Mallal S, French M, Skett J, Maxwel D, Cain A, Montroni M, Scalise G, Costantini A, Giacometti A, Tirelli U, Nasti G, Pastore G, Ladisa N, Perulli ML, Suter F, Arici C, Maggiolo F, Chiodo F, Gritti FM, Colangeli V, Fiorini C, Guerra L, Carosi G, Cadeo GP, Minardi C, Vangi D, Paraninfo G, Casari S, Pan A, Patroni A, Torti C, Quiros-Roldan E, Tomasoni L, Moretti F, Nasta P, Uccelli MC, Bertelli D, Rizzardini G, Migliorino G, Migliorino M, Abeli C, Manconi PE, Piano P, Ferraro T, Scerbo A, Pizzigallo E, Ricci F, Santoro D, Pusterla L, Carnevale G, Galloni D, Viganò P, Mena M, Ghinelli F, Sighinolfi L, Leoncini F, Mazzotta F, Pozzi M, Lo Caputo S, Angarano G, Grisorio B, Ferrara S, Grima P, Tundo P, Pagano 188  G, Piersantelli N, Alessandrini A, Piscopo R, Toti M, Chigiotti S, Soscia F, Tacconi L, Orani A, Perini P, Nigro M, Scasso A, Vincenti A, Scalzini A, Fibbia G, Moroni M, Lazzarin A, Cargnel A, Vigevani GM, Caggese L, Tordato F, Novati R, Galli A, Merli S, Pastecchia C, Moioli C, Esposito R, Abrescia N, Chirianni A, Izzo C, Piazza M, De Marco M, Montesarchio V, Manzillo E, Nappa S, Colomba A, Abbadessa V, Prestileo T, Mancuso S, Ferrari C, Pzzaferri P, Filice G, Minoli L, Bruno R, Maserati R, Novati S, Tinelli C, Pauluzzi S, Baldelli F, Petrelli E, Cioppi A, Alberici F, Ruggieri A, Menichetti F, Martinelli C, De Stefano C, La Gala A, Zauli T, Ballardini G, Magnani G, Ursitti MA, Arlotti M, Ortolani P, Ortona L, Dianzani F, Ippolito G, Antinori A, Antonucci G, D’Elia S, Narciso P, Petrosillo N, Vullo V, De Luca A, Del Forno L, Zaccarelli M, De Longis P, Ciardi M, D’Offizi G, Noto P, Lichtner M, Capobianchi MR, Girardi E, Pezzotti P, Rezza G, Mura MS, Mannazzu M, Caramello P, Sinicco A, Soranzo ML, Gennero L, Sciandra M, Salassa B, Grossi PA, Basilico C, Poggio A, Bottari G, Raise E, Pasquinucci S, De Lalla F, Tositti G, Resta F, Chimienti A, Lepri Cozzi A, Bachmann S, Battegay M, Bernasconi E, Bucher H, Bürgisser P, Cattacin S, Egger M, Erb P, Fierz W, Fischer M, Flepp M, Fontana A, Francioli P, Furrer HJ, Gorgievski M, Günthard H, Hirschel B, Kaiser L, Kind C, Klimkait T, Lauper U, Opravil M, Paccaud F, Pantaleo G, Perrin L, Piffaretti JC, Rickenbach M, Rudin C, Schüpbach J, Speck R, Tarr P, Telenti A, Trkola A, Vernazza P, Yerly S, de Wolf F, van Sighem AI, van Valkengoed I, Gras L, Bronsveld W, Veldkamp A, Prins JM, Bos JC, Schattenkerk Eeftinck JKM, Godfried MH, Lange JMA, Lowe SH, van der Meer JTM, Nellen FJB, Pogany K, van der Poll T, Ruys TA, Sankatsing S, van der Valk M, van Vonderen MGA, Wit FWMN, van Eeden A, ten Veen JH, van Dam PS, Hillebrand-Haverkort ME, Brinkman K, Frissen PHJ, Weigel HM, Mulder JW, van Gorp ECM, Meenhorst PL, Mairuhu ATA, Veenstra J, Danner SA, Van Agtmael MA, Claessen FAP, Geerlings SE, Perenboom RM, Jurriaans S, Back NKT, Cuijpers T, Rietra PJGM, Roozendaal KJ, Pauw W, van Zanten AP, Smits PHM, von Blomberg BME, Savelkoul P, Zaaijer H, Beijnen JH, Crommentuyn KML, Huitema ADR, Kappelhoff B, de Maat MMR, Richter C, van der Berg J, van Leusen R, Swanink C, Vriesendorp R, Jeurissen FJF, Kauffmann RH, Koger ELW, Franck PFH, Lampe AS, Jansen CL, Bravenboer B, ten Napel CHH, Mudrikova T, Hendriks R, Sprenger HG, Miesen WMAJ, Schirm J, Benne D, ten Kate RW, Veenendaal D, van Houte DPF, Leemhuis MP, Pole M, Storm H, van Zeijl JH, Kroon FP, Schippers EF, Kroes ACM, Claas HCJ, Schreij G, van de Geest S, Verbon A, Bruggeman CAMVA, Goossens VJ, Koopmans PP, Telgt M, van der Ven AJAM, Burger DM, Hugen PWH, Galama JMD, Poort YAGM, van der Ende ME, Gyssens IC, de Marie S, Nouwen JL, Niesters MG, Osterhaus ADME, Schutten M, Juttmann JR, Buiting AGM, Swaans CAM, Schneider MME, Bonten MJM, Borleffs JCC, Hoepelman IM, Jaspers CAJJ, Schouten I, Schurink CAM, Boucher CAB, Schuurman R, Blok WL, Groeneveld PHP, Boel E, Jansz AF, Dabis F, Thiebaut R, Chêne G, Lawson-Ayayi S, Meyer L, Boufassa F, Hamouda O, Touloumi G, Hatzakis A, Karafoulidou A, Katsarou O, Brettle R, Del Amo J, del Romero J, van Asten L, van Benthem B, Prins M, Coutinho R, Kirk O, Pedersen C, Hernández Aguado I, Pérez-Hoyos S, Eskild A, Bruun JN, Sannes M, Lee C, Johnson AM, Babiker A, Darbyshire J, Gill N, Porter K, Vanhems P, Kaldor J, Ashton L, Vizzard J, Muga R, Gill J, Cayla J, Garcia de Olalla P, Day NE, De Angelis D, Walker S, Tyrer F, Beral V, Raffanti S, Becker S, Scarsella A, Braun J, Most B, Balu R, Gilbert L, Fleenor R, Ising T, Dieterich D, Fusco J, Losso M, Duran A, Vetter N, Clumeck N, De Wit S, Kabeya K, Poll B, Colebunders R, Machala L, Rozsypal H, Nielsen J, Lundgren J, Olsen CH, Gerstoft J, Katzenstein T, Hansen ABE, Skinhøj P, Zilmer K, Rauka M, Katlama C, De Sa M, Viard JP, Saint-Marc T, Pradier C, Dietrich M, Manegold C, van Lunzen J, Stellbrink HJ, Miller V, Goebel FD, Salzberger B, Rockstroh J, Kosmidis J, Gargalianos P, Sambatakou H, Perdios J, Panos G, Filandras A, Banhegyi D, Mulcahy F, Yust I, Burke M, Pollack S, Hassoun J, Sthoeger Z, Maayan S, Vella S, Chiesi A, Pristerá R, Gabbuti A, Bedini A, Montesarchio E, Santopadre P, Franci P, Castagna A, Viksna L, Rozentale B, Chaplinskas S, Hemmer R, Staub T, Bruun J, Maeland A, Ormaasen V, Knysz B, Gasiorowski J, Horban A, Prokopowicz D, Wiercinska-Drapalo A, Boron-Kaczmarska A, Pynka M, Beniowski M, Trocha H, Smiatacz T, Antunes F, Mansinho K, Maltez F, Duiculescu D, Streinu-Cercel A, Mokrás M, Staneková D, González-Lahoz J, Diaz B, García-Benayas T, Martin-Carbonero L, Soriano V, Clotet B, Jou A, Conejero J, Tural C, Gatell JM, Miró JM, Zamora L, Blaxhult A, Karlsson A, Pehrson P, Schiffer V, Furrer H, Chentsova N, Barton S, Mercey D, Phillips A, Youle M, Johnson MA, Mocroft A, Murphy M, Weber J, Scullard G, Fisher M, Loveday C, Ruiz L, Helm EB, Carlebach A, Mösch M, Müller A, Haberl A, Korn S, Stephan C, Bickel M, Gute P, Locher L, Lutz T, Klauke S, Doerr HW, Stürmer M, Dauer B, Jennings B, Alexander C, Braitstein P, Chan K, Cote H, Gataric N, Harrigan PR, Harris M, Bonner S, Montaner J, O’Shaughnessy M, Wood E, Yip B, Chaloner C, Gumley H, 189  Ransom D, Sabin CA, Lipman M, Johnson M, Read R, Riccio G, Borghi V. Predictors of trend in CD4-positive T-cell count and mortality among HIV-1-infected individuals with virological failure to all three antiretroviral-drug classes. Lancet. 2004;364(9428):51–62.  259.  Murri R, Lepri AC, Cicconi P, Poggio A, Arlotti M, Tositti G, Santoro D, Soranzo ML, Rizzardini G, Colangeli V, Montroni M, Monforte AD. Is moderate HIV viremia associated with a higher risk of clinical progression in HIV-infected people treated with highly active antiretroviral therapy: evidence from the Italian cohort of antiretroviral-naive patients study. J Acquir Immune Defic Syndr. 2006;41(1):23–30.  260.  El-Sadr WM, Lundgren JD, Neaton JD, Gordin F, Abrams D, Arduino RC, Babiker A, Burman W, Clumeck N, Cohen CJ, Cohn D, Cooper D, Darbyshire J, Emery S, Fätkenheuer G, Gazzard B, Grund B, Hoy J, Klingman K, Losso M, Markowitz N, Neuhaus J, Phillips A, Rappoport C. CD4+ count-guided interruption of antiretroviral treatment. N Engl J Med. 2006;355(22):2283–96.  261.  Lundgren JD, Babiker A, El-Sadr W, Emery S, Grund B, Neaton JD, Neuhaus J, Phillips AN. Inferior clinical outcome of the CD4+ cell count-guided antiretroviral treatment interruption strategy in the SMART study: role of CD4+ Cell counts and HIV RNA levels during follow-up. J Infect Dis. 2008;197(8):1145–55.  262.  DART Trial Team. Fixed duration interruptions are inferior to continuous treatment in African adults starting therapy with CD4 cell counts < 200 cells/µl. AIDS. 2008;22(2):237–47.  263.  Ananworanich J, Gayet-Ageron A, Le Braz M, Prasithsirikul W, Chetchotisakd P, Kiertiburanakul S, Munsakul W, Raksakulkarn P, Tansuphasawasdikul S, Sirivichayakul S, Cavassini M, Karrer U, Genné D, Nüesch R, Vernazza P, Bernasconi E, Leduc D, Satchell C, Yerly S, Perrin L, Hill A, Perneger T, Phanuphak P, Furrer H, Cooper D, Ruxrungtham K, Hirschel B. CD4-guided scheduled treatment interruptions compared with continuous therapy for patients infected with HIV-1: results of the Staccato randomised trial. Lancet. 2006;368(9534):459–65.  264.  Danel C, Moh R, Minga A, Anzian A, Ba-Gomis O, Kanga C, Nzunetu G, Gabillard D, Rouet F, Sorho S, Chaix ML, Eholié S, Menan H, Sauvageot D, Bissagnene E, Salamon R, Anglaret X. CD4-guided structured antiretroviral treatment interruption strategy in HIV-infected adults in west Africa (Trivacan ANRS 1269 trial): a randomised trial. Lancet. 2006;367(9527):1981–9.  265.  Phillips AN, Carr A, Neuhaus J, Visnegarwala F, Prineas R, Burman WJ, Williams I, Drummond F, Duprez D, Belloso WH, Goebel FD, Grund B, Hatzakis A, Vera J, Lundgren JD. Interruption of antiretroviral therapy and risk of cardiovascular disease in persons with HIV-1 infection: exploratory analyses from the SMART trial. Antivir Ther. 2008;13(2):177–87.  266.  Duprez DA, Neuhaus J, Kuller LH, Tracy R, Belloso W, De Wit S, Drummond F, Lane HC, Ledergerber B, Lundgren J, Nixon D, Paton NI, Prineas RJ, Neaton JD. Inflammation, coagulation and cardiovascular disease in HIV-infected individuals. PLoS One. 2012;7(9):e44454.  267.  Carpenter CC, Cooper DA, Fischl MA, Gatell JM, Gazzard BG, Hammer SM, Hirsch MS, Jacobsen DM, Katzenstein DA, Montaner JS, Richman DD, Saag MS, Schechter M, Schooley RT, Thompson MA, Vella S, Yeni PG, Volberding PA. Antiretroviral therapy in adults: updated recommendations of the International AIDS Society-USA Panel. JAMA. 2000;283(3):381–90.  268.  Thompson MA, Aberg JA, Cahn P, Montaner JSG, Rizzardini G, Telenti A, Gatell JM, Günthard HF, Hammer SM, Hirsch MS, Jacobsen DM, Reiss P, Richman DD, Volberding PA, Yeni P, Schooley RT. Antiretroviral treatment of adult HIV infection: 2010 recommendations of the International AIDS Society-USA Panel. JAMA. 2010;304(3):321–33.  190  269.  Church JD, Jones D, Flys T, Hoover D, Marlowe N, Chen S, Shi C, Eshleman JR, Guay LA, Jackson JB, Kumwenda N, Taha TE, Eshleman SH. Sensitivity of the ViroSeq HIV-1 genotyping system for detection of the K103N resistance mutation in HIV-1 subtypes A, C, and D. J Mol Diagn. 2006;8(4):430–2.  270.  Boltz VF, Bao Y, Lockman S, Halvas EK, Kearney MF, McIntyre JA, Schooley RT, Hughes MD, Coffin JM, Mellors JW. Low-frequency nevirapine (NVP)-resistant HIV-1 variants are not associated with failure of antiretroviral therapy in women without prior exposure to single-dose NVP. J Infect Dis. 2014;209(5):703–10.  271.  Gill VS, Lima VD, Zhang W, Wynhoven B, Yip B, Hogg RS, Montaner JSG, Harrigan PR. Improved virological outcomes in British Columbia concomitant with decreasing incidence of HIV type 1 drug resistance detection. Clin Infect Dis. 2010;50(1):98–105.  272.  Cescon A, Kanters S, Brumme CJ, Lepik KJ, Forrest JI, Hull M, Samji H, Nosyk B, Harrigan PR, Hogg RS, Montaner JSG. Trends in plasma HIV-RNA suppression and antiretroviral resistance in British Columbia, 1997-2010. J Acquir Immune Defic Syndr. 2014;65(1):107–14.  273.  Tam LWY, Chui CKS, Brumme CJ, Bangsberg DR, Montaner JSG, Hogg RS, Harrigan PR. The relationship between resistance and adherence in drug-naive individuals initiating HAART is specific to individual drug classes. J Acquir Immune Defic Syndr. 2008;49(3):266–71.  274.  Harrigan PR, Hogg RS, Dong WWY, Yip B, Wynhoven B, Woodward J, Brumme CJ, Brumme ZL, Mo T, Alexander CS, Montaner JSG. Predictors of HIV drug-resistance mutations in a large antiretroviral-naive cohort initiating triple antiretroviral therapy. J Infect Dis. 2005;191(3):339–47.  275.  Palmer S, Maldarelli F, Wiegand A, Bernstein B, Hanna GJ, Brun SC, Kempf DJ, Mellors JW, Coffin JM, King MS. Low-level viremia persists for at least 7 years in patients on suppressive antiretroviral therapy. Proc Natl Acad Sci U S A. 2008;105(10):3879–84.  276.  Shen L, Siliciano RF. Viral reservoirs, residual viremia, and the potential of highly active antiretroviral therapy to eradicate HIV infection. J Allergy Clin Immunol. 2008;122(1):22–8.  277.  Zhang L, Ramratnam B, Tenner-Racz K, He Y, Vesanen M, Lewin S, Talal A, Racz P, Perelson a S, Korber BT, Markowitz M, Ho DD. Quantifying residual HIV-1 replication in patients receiving combination antiretroviral therapy. N Engl J Med. 1999;340(21):1605–13.  278.  Dinoso JB, Kim SY, Wiegand AM, Palmer SE, Gange SJ, Cranmer L, O’Shea A, Callender M, Spivak A, Brennan T, Kearney MF, Proschan MA, Mican JM, Rehm CA, Coffin JM, Mellors JW, Siliciano RF, Maldarelli F. Treatment intensification does not reduce residual HIV-1 viremia in patients on highly active antiretroviral therapy. Proc Natl Acad Sci U S A. 2009;106(23):9403–8.  279.  Kieffer TL, Finucane MM, Nettles RE, Quinn TC, Broman KW, Ray SC, Persaud D, Siliciano RF. Genotypic analysis of HIV-1 drug resistance at the limit of detection: virus production without evolution in treated adults with undetectable HIV loads. J Infect Dis. 2004;189(8):1452–65.  280.  Hermankova M, Ray SC, Ruff C, Powell-Davis M, Ingersoll R, D’Aquila RT, Quinn TC, Siliciano JD, Siliciano RF, Persaud D. HIV-1 drug resistance profiles in children and adults with viral load of <50 copies/ml receiving combination therapy. JAMA. 2001;286(2):196–207.  281.  Erice A, Mayers DL, Strike DG, Sannerud KJ, McCutchan FE, Henry K, Balfour HH. Brief report: primary infection with zidovudine-resistant human immunodeficiency virus type 1. N Engl J Med. 1993;328(16):1163–5.  191  282.  Little SJ, Holte S, Routy J-P, Daar ES, Markowitz M, Collier AC, Koup RA, Mellors JW, Connick E, Conway B, Kilby M, Wang L, Whitcomb JM, Hellmann NS, Richman DD. Antiretroviral-drug resistance among patients recently infected with HIV. N Engl J Med. 2002;347(6):385–94.  283.  Vercauteren J, Wensing AMJ, van de Vijver DAMC, Albert J, Balotta C, Hamouda O, Kücherer C, Struck D, Schmit J-C, Asjö B, Bruckova M, Camacho RJ, Clotet B, Coughlan S, Grossman Z, Horban A, Korn K, Kostrikis L, Nielsen C, Paraskevis D, Poljak M, Puchhammer-Stöckl E, Riva C, Ruiz L, Salminen M, Schuurman R, Sonnerborg A, Stanekova D, Stanojevic M, Vandamme A-M, Boucher CAB. Transmission of drug-resistant HIV-1 is stabilizing in Europe. J Infect Dis. 2009;200(10):1503–8.  284.  Wensing AMJ, van de Vijver DA, Angarano G, Asjö B, Balotta C, Boeri E, Camacho R, Chaix M-L, Costagliola D, De Luca A, Derdelinckx I, Grossman Z, Hamouda O, Hatzakis A, Hemmer R, Hoepelman A, Horban A, Korn K, Kücherer C, Leitner T, Loveday C, MacRae E, Maljkovic I, de Mendoza C, Meyer L, Nielsen C, Op de Coul EL, Ormaasen V, Paraskevis D, Perrin L, Puchhammer-Stöckl E, Ruiz L, Salminen M, Schmit J-C, Schneider F, Schuurman R, Soriano V, Stanczak G, Stanojevic M, Vandamme A-M, Van Laethem K, Violin M, Wilbe K, Yerly S, Zazzi M, Boucher CA. Prevalence of drug-resistant HIV-1 variants in untreated individuals in Europe: implications for clinical management. J Infect Dis. 2005;192(6):958–66.  285.  Yerly S, Vora S, Rizzardi P, Chave JP, Vernazza PL, Flepp M, Telenti A, Battegay M, Veuthey AL, Bru JP, Rickenbach M, Hirschel B, Perrin L. Acute HIV infection: impact on the spread of HIV and transmission of drug resistance. AIDS. 2001;15(17):2287–92.  286.  Gupta RK, Jordan MR, Sultan BJ, Hill A, Davis DHJ, Gregson J, Sawyer AW, Hamers RL, Ndembi N, Pillay D, Bertagnolio S. Global trends in antiretroviral resistance in treatment-naïve individuals with HIV after rollout of antiretroviral treatment in resource-limited settings: a global collaborative study and meta-regression analysis. Lancet. 2012;380(9849):1250–8.  287.  Hamers RL, Wallis CL, Kityo C, Siwale M, Mandaliya K, Conradie F, Botes ME, Wellington M, Osibogun A, Sigaloff KCE, Nankya I, Schuurman R, Wit FW, Stevens WS, van Vugt M, de Wit TFR. HIV-1 drug resistance in antiretroviral-naive individuals in sub-Saharan Africa after rollout of antiretroviral therapy: a multicentre observational study. Lancet Infect Dis. 2011;11(10):750–9.  288.  Hamers RL, Kityo C, Lange JMA, Wit TFRD, Mugyenyi P. Global threat from drug resistant HIV in sub-Saharan Africa. BMJ. 2012;344:e4159.  289.  Hogg RS, Bangsberg DR, Lima VD, Alexander C, Bonner S, Yip B, Wood E, Dong WWY, Montaner JSG, Harrigan PR. Emergence of drug resistance is associated with an increased risk of death among patients first starting HAART. PLoS Med. 2006;3(9):1570–8.  290.  DeGruttola V, Dix L, D’Aquila R, Holder D, Phillips A, Ait-Khaled M, Baxter J, Clevenbergh P, Hammer S, Harrigan R, Katzenstein D, Lanier R, Miller M, Para M, Yerly S, Zolopa A, Murray J, Patick A, Miller V, Castillo S, Pedneault L, Mellors J. The relation between baseline HIV drug resistance and response to antiretroviral therapy: re-analysis of retrospective and prospective studies using a standardized data analysis plan. Antivir Ther. 2000;5(1):41–8.  291.  Durant J, Clevenbergh P, Halfon P, Delgiudice P, Porsin S, Simonet P, Montagne N, Boucher CAB, Schapiro JM, Dellamonica P. Drug-resistance genotyping in HIV-1 therapy: The VIRADAPT randomised controlled trial. Lancet. 1999;353(9171):2195–9.  292.  Baxter JD, Mayers DL, Wentworth DN, Neaton JD, Hoover ML, Winters MA, Mannheimer SB, Thompson MA, Abrams DI, Brizz BJ, Ioannidis JP, Merigan TC, the CPCRA 046 Study Team for the Terry Beirn Community Programs for Clinical Research on AIDS. A randomized study of 192  antiretroviral management based on plasma genotypic antiretroviral resistance testing in patients failing therapy. AIDS. 2000;14(9):F83–93.  293.  Tural C, Ruiz L, Holtzer C, Schapiro J, Viciana P, González J, Domingo P, Boucher C, Rey-Joly C, Clotet B. Clinical utility of HIV-1 genotyping and expert advice: the Havana trial. AIDS. 2002;16(2):209–18.  294.  Cingolani A, Antinori A, Rizzo MG, Murri R, Ammassari A, Baldini F, Di Giambenedetto S, Cauda R, De Luca A. Usefulness of monitoring HIV drug resistance and adherence in individuals failing highly active antiretroviral therapy: a randomized study (ARGENTA). AIDS. 2002;16(3):369–79.  295.  Sax PE, Islam R, Walensky RP, Losina E, Weinstein MC, Goldie SJ, Sadownik SN, Freedberg KA. Should resistance testing be performed for treatment-naive HIV-infected patients? A cost-effectiveness analysis. Clin Infect Dis. 2005;41(9):1316–23.  296.  Weinstein MC, Goldie SJ, Losina E, Cohen CJ, Baxter JD, Zhang H, Kimmel AD, Freedberg KA. Use of genotypic resistance testing to guide HIV therapy: clinical impact and cost-effectiveness. Ann Intern Med. 2001;134(6):440–50.  297.  Kellam P, Larder BA. Recombinant virus assay: A rapid, phenotypic assay for assessment of drug susceptibility of human immunodeficiency virus type 1 isolates. Antimicrob Agents Chemother. 1994;38(1):23–30.  298.  Zhang J, Rhee S-Y, Taylor J, Shafer RW. Comparison of the precision and sensitivity of the Antivirogram and PhenoSense HIV drug susceptibility assays. J Acquir Immune Defic Syndr. 2005;38(4):439–44.  299.  Johnson VA, Calvez V, Gunthard HF, Paredes R, Pillay D, Shafer RW, Wensing AM, Richman DD. Update of the drug resistance mutations in HIV-1: March 2013. Top Antivir Med. 2013;21(1):6–14.  300.  Cockerill 3rd FR. Genetic methods for assessing antimicrobial resistance. Antimicrob Agents Chemother. 1999;43(2):199–212.  301.  Liu TF, Shafer RW. Web resources for HIV type 1 genotypic-resistance test interpretation. Clin Infect Dis. 2006;42(11):1608–18.  302.  Bennett DE, Camacho RJ, Otelea D, Kuritzkes DR, Fleury H, Kiuchi M, Heneine W, Kantor R, Jordan MR, Schapiro JM, Vandamme A-M, Sandstrom P, Boucher CAB, van de Vijver D, Rhee S-YY, Liu TF, Pillay D, Shafer RW. Drug resistance mutations for surveillance of transmitted HIV-1 drug-resistance: 2009 update. PLoS One. 2009;4(3):e4724.  303.  Vermeiren H, Van Craenenbroeck E, Alen P, Bacheler L, Picchio G, Lecocq P. Prediction of HIV-1 drug susceptibility phenotype from the viral genotype using linear regression modeling. J Virol Methods. 2007;145(1):47–55.  304.  Winters B, Montaner J, Harrigan PR, Gazzard B, Pozniak A, Miller MD, Emery S, van Leth F, Robinson P, Baxter JD, Perez-Elias M, Castor D, Hammer S, Rinehart A, Vermeiren H, Van Craenenbroeck E, Bacheler L. Determination of clinically relevant cutoffs for HIV-1 phenotypic resistance estimates through a combined analysis of clinical trial and cohort data. J Acquir Immune Defic Syndr. 2008;48(1):26–34.  305.  Puchhammer-Stöckl E, Steininger C, Geringer E, Heinz FX. Comparison of virtual phenotype and HIV-SEQ program (Stanford) interpretation for predicting drug resistance of HIV strains. HIV Med. 2002;3(3):200–6.  193  306.  Kijak GH, Rubio AE, Pampuro SE, Zala C, Cahn P, Galli R, Montaner JS, Salomón H. Discrepant results in the interpretation of HIV-1 drug-resistance genotypic data among widely used algorithms. HIV Med. 2003;4(1):72–8.  307.  Jiamsakul A, Kantor R, Li PC, Sirivichayakul S, Sirisanthana T, Kantipong P, Lee CK, Kamarulzaman A, Ratanasuwan W, Ditangco R, Singtoroj T, Sungkanuparph S. Comparison of predicted susceptibility between genotype and virtual phenotype HIV drug resistance interpretation systems among treatment-naive HIV-infected patients in Asia: TASER-M cohort analysis. BMC Res Notes. 2012;5(1):582.  308.  Rhee S-Y, Fessel WJ, Liu TF, Marlowe NM, Rowland CM, Rode RA, Vandamme A-M, Van Laethem K, Brun-Vézinet F, Calvez V, Taylor J, Hurley L, Horberg M, Shafer RW. Predictive value of HIV-1 genotypic resistance test interpretation algorithms. J Infect Dis. 2009;200(3):453–63.  309.  Frentz D, Boucher CAB, Assel M, de Luca A, Fabbiani M, Incardona F, Libin P, Manca N, Müller V, Nualláin BÓ, Paredes R, Prosperi M, Quiros-Roldan E, Ruiz L, Sloot PMA, Torti C, Vandamme A-M, van Laethem K, Zazzi M, van de Vijver DAMC. Comparison of HIV-1 genotypic resistance test interpretation systems in predicting virological outcomes over time. PLoS One. 2010;5(7):e11505.  310.  Korn K, Reil H, Walter H, Schmidt B. Quality control trial for human immunodeficiency virus type 1 drug resistance testing using clinical samples reveals problems with detecting minority species and interpretation of test results. J Clin Microbiol. 2003;41(8):3559–65.  311.  Schuurman R, Brambilla D, de Groot T, Huang D, Land S, Bremer J, Benders I, Boucher CAB. Underestimation of HIV type 1 drug resistance mutations: results from the ENVA-2 genotyping proficiency program. AIDS Res Hum Retroviruses. 2002;18(4):243–8.  312.  Shafer RW, Warford A, Winters MA, Gonzales MJ. Reproducibility of human immunodeficiency virus type 1 (HIV-1) protease and reverse transcriptase sequencing of plasma samples from heavily treated HIV-1-infected individuals. J Virol Methods. 2000;86(2):143–53.  313.  Johnson JA, Li JF, Wei X, Lipscomb J, Irlbeck D, Craig C, Smith A, Bennett DE, Monsour M, Sandstrom P, Lanier ER, Heneine W. Minority HIV-1 drug resistance mutations are present in antiretroviral treatment-naïve populations and associate with reduced treatment efficacy. PLoS Med. 2008;5(7):1112–22.  314.  Paredes R, Lalama CM, Ribaudo HJ, Schackman BR, Shikuma C, Giguel F, Meyer WA, Johnson VA, Fiscus SA, D’Aquila RT, Gulick RM, Kuritzkes DR. Pre-existing minority drug-resistant HIV-1 variants, adherence, and risk of antiretroviral treatment failure. J Infect Dis. 2010;201(5):662–71.  315.  Simen BB, Simons JF, Hullsiek KH, Novak RM, Macarthur RD, Baxter JD, Huang C, Lubeski C, Turenchalk GS, Braverman MS, Desany B, Rothberg JM, Egholm M, Kozal MJ. Low-abundance drug-resistant viral variants in chronically HIV-infected, antiretroviral treatment-naive patients significantly impact treatment outcomes. J Infect Dis. 2009;199(5):693–701.  316.  Li JZ, Paredes R, Ribaudo HJ, Svarovskaia ES, Metzner KJ, Kozal MJ, Hullsiek KH, Balduin M, Jakobsen MR, Geretti AM, Thiebaut R, Ostergaard L, Masquelier B, Johnson JA, Miller MD, Kuritzkes DR. Low-frequency HIV-1 drug resistance mutations and risk of NNRTI-based antiretroviral treatment failure: a systematic review and pooled analysis. JAMA. 2011;305(13):1327–35.  317.  Cozzi-Lepri A, Noguera-Julian M, Di Giallonardo F, Schuurman R, Däumer M, Aitken S, Günthard HF, Brun-Vézinet F, Metzner K, Paredes R, Group and the CMH-1 VW. Low-frequency drug-resistant HIV-1 and risk of virological failure to first-line NNRTI-based ART: a multi-cohort 194  European case-control study using centralized ultrasensitive 454 sequencing. J Antimicrob Chemother. 2015;70(3):930–40.  318.  Peuchant O, Thiébaut R, Capdepont S, Lavignolle-Aurillac V, Neau D, Morlat P, Dabis F, Fleury H, Masquelier B. Transmission of HIV-1 minority-resistant variants and response to first-line antiretroviral therapy. AIDS. 2008;22(12):1417–23.  319.  Lataillade M, Chiarella J, Yang R, Schnittman S, Wirtz V, Uy J, Seekins D, Krystal M, Mancini M, McGrath D, Simen B, Egholm M, Kozal M. Prevalence and clinical significance of HIV drug resistance mutations by ultra-deep sequencing in antiretroviral-naïve subjects in the CASTLE study. PLoS One. 2010;5(6):e10952.  320.  Charpentier C, Laureillard D, Piketty C, Tisserand P, Batisse D, Karmochkine M, Si-Mohamed A, Weiss L. High frequency of integrase Q148R minority variants in HIV-infected patients naive of integrase inhibitors. AIDS. 2010;24(6):867–73.  321.  Ross LL, Weinberg WG, DeJesus E, Fischl MA, Horton JH, Pappa KA, Lancaster CT, Pakes GE, Smith KY. Impact of low abundance HIV variants on response to ritonavir-boosted atazanavir or fosamprenavir given once daily with tenofovir/emtricitabine in antiretroviral-naive HIV-infected patients. AIDS Res Hum Retroviruses. 2010;26(4):407–17.  322.  Swenson LC, Mo T, Dong WWY, Zhong X, Woods CK, Jensen MA, Thielen A, Chapman D, Lewis M, James I, Heera J, Valdez H, Harrigan PR. Deep sequencing to infer HIV-1 co-receptor usage: application to three clinical trials of maraviroc in treatment-experienced patients. J Infect Dis. 2011;203(2):237–45.  323.  Swenson LC, Mo T, Dong WWY, Zhong X, Woods CK, Thielen A, Jensen MA, Knapp DJHF, Chapman D, Portsmouth S, Lewis M, James I, Heera J, Valdez H, Harrigan PR. Deep V3 sequencing for HIV type 1 tropism in treatment-naive patients: a reanalysis of the MERIT trial of maraviroc. Clin Infect Dis. 2011;53(7):732–42.  324.  Sanger F, Nicklen S, Coulson AR. DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci U S A. 1977;74(12):5463–7.  325.  Smith LM, Sanders JZ, Kaiser RJ, Hughes P, Dodd C, Connell CR, Heiner C, Kent SB, Hood LE. Fluorescence detection in automated DNA sequence analysis. Nature. 1986;321(6071):674–9.  326.  Prober JM, Trainor GL, Dam RJ, Hobbs FW, Robertson CW, Zagursky RJ, Cocuzza AJ, Jensen MA, Baumeister K. A system for rapid DNA sequencing with fluorescent chain-terminating dideoxynucleotides. Science. 1987;238(4825):336–41.  327.  Smith LM, Fung S, Hunkapiller MW, Hunkapiller TJ, Hood LE. The synthesis of oligonucleotides containing an aliphatic amino group at the 5’ terminus: synthesis of fluorescent DNA primers for use in DNA sequence analysis. Nucleic Acids Res. 1985;13(7):2399–412.  328.  Swerdlow H, Wu S, Harke H, Dovichi NJ. Capillary gel electrophoresis for DNA sequencing. Laser-induced fluorescence detection with the sheath flow cuvette. J Chromatogr. 1990;516(1):61–7.  329.  Ewing B, Hillier L, Wendl MC, Green P. Base-calling of automated sequencer traces using phred. I. Accuracy assessment. Genome Res. 1998;8(3):175–85.  330.  Ewing B, Green P. Base-calling of automated sequencer traces using phred. II. Error probabilities. Genome Res. 1998;8(3):186–94.  195  331.  Knapp DJHF, McGovern RA, Dong W, Poon AFY, Swenson LC, Zhong X, Woods CK, Harrigan PR. Factors influencing the sensitivity and specificity of conventional sequencing in human immunodeficiency virus type 1 tropism testing. J Clin Microbiol. 2013;51(2):444–51.  332.  Larder BA, Kohli A, Kellam P, Kemp SD, Kronick M, Henfrey RD. Quantitative detection of HIV-1 drug resistance mutations by automated DNA sequencing. Nature. 1993;365(6447):671–3.  333.  Shafer RW, Hertogs K, Zolopa AR, Warford A, Bloor S, Betts BJ, Merigan TC, Harrigan R, Larder BA. High degree of interlaboratory reproducibility of human immunodeficiency virus type 1 protease and reverse transcriptase sequencing of plasma samples from heavily treated patients. J Clin Microbiol. 2001;39(4):1522–9.  334.  Galli RA, Sattha B, Wynhoven B, O’Shaughnessy MV, Harrigan PR. Sources and magnitude of intralaboratory variability in a sequence-based genotypic assay for human immunodeficiency virus type 1 drug resistance. J Clin Microbiol. 2003;41(7):2900–7.  335.  Huang DD, Eshleman SH, Brambilla DJ, Palumbo PE, Bremer JW. Evaluation of the editing process in human immunodeficiency virus type 1 genotyping. J Clin Microbiol. 2003;41(7):3265–72.  336.  Struck D, Wallis CL, Denisov G, Lambert C, Servais JY, Viana RV, Letsoalo E, Bronze M, Aitken SC, Schuurman R, Stevens W, Schmit JC, Rinke de Wit T, Perez Bercoff D. Automated sequence analysis and editing software for HIV drug resistance testing. J Clin Virol. 2012;54(1):30–5.  337.  Voelkerding KV, Dames SA, Durtschi JD. Next-generation sequencing: from basic research to diagnostics. Clin Chem. 2009;55(4):641–58.  338.  Nyrén P. Enzymatic method for continuous monitoring of DNA polymerase activity. Anal Biochem. 1987;167(2):235–8.  339.  Ronaghi M, Karamohamed S, Pettersson B, Uhlén M, Nyrén P. Real-time DNA sequencing using detection of pyrophosphate release. Anal Biochem. 1996;242(1):84–9.  340.  Margulies M, Egholm M, Altman WE, Attiya S, Bader JS, Bemben LA, Berka J, Braverman MS, Chen Y-J, Chen Z, Dewell SB, Du L, Fierro JM, Gomes XV, Godwin BC, He W, Helgesen S, Ho CH, Irzyk GP, Jando SC, Alenquer MLI, Jarvie TP, Jirage KB, Kim J-B, Knight JR, Lanza JR, Leamon JH, Lefkowitz SM, Lei M, Li J, Lohman KL, Lu H, Makhijani VB, McDade KE, McKenna MP, Myers EW, Nickerson E, Nobile JR, Plant R, Puc BP, Ronan MT, Roth GT, Sarkis GJ, Simons JF, Simpson JW, Srinivasan M, Tartaro KR, Tomasz A, Vogt KA, Volkmer GA, Wang SH, Wang Y, Weiner MP, Yu P, Begley RF, Rothberg JM. Genome sequencing in microfabricated high-density picolitre reactors. Nature. 2005;437(7057):376–80.  341.  Dressman D, Yan H, Traverso G, Kinzler KW, Vogelstein B. Transforming single DNA molecules into fluorescent magnetic particles for detection and enumeration of genetic variations. Proc Natl Acad Sci U S A. 2003;100(15):8817–22.  342.  Archer J, Rambaut A, Taillon BE, Harrigan PR, Lewis M, Robertson DL. The evolutionary analysis of emerging low frequency HIV-1 CXCR4 using variants through time-an ultra-deep approach. PLoS Comput Biol. 2010;6(12):e1001022.  343.  Archer J, Weber J, Henry K, Winner D, Gibson R, Lee L, Paxinos E, Arts EJ, Robertson DL, Mimms L, Quiñones-Mateu ME. Use of four next-generation sequencing platforms to determine HIV-1 coreceptor tropism. PLoS One. 2012;7(11):e49602.  196  344.  Swenson LC, Moores A, Low AJ, Thielen A, Dong W, Woods C, Jensen MA, Wynhoven B, Chan D, Glascock C, Harrigan PR. Improved detection of CXCR4-using HIV by V3 genotyping: application of population-based and “deep” sequencing to plasma RNA and proviral DNA. J Acquir Immune Defic Syndr. 2010;54(5):506–10.  345.  Wang C, Mitsuya Y, Gharizadeh B, Ronaghi M, Shafer RW. Characterization of mutation spectra with ultra-deep pyrosequencing: application to HIV-1 drug resistance. Genome Res. 2007;17(8):1195–201.  346.  Henn MR, Boutwell CL, Charlebois P, Lennon NJ, Power KA, Macalalad AR, Berlin AM, Malboeuf CM, Ryan EM, Gnerre S, Zody MC, Erlich RL, Green LM, Berical A, Wang Y, Casali M, Streeck H, Bloom AK, Dudek T, Tully D, Newman R, Axten KL, Gladden AD, Battis L, Kemper M, Zeng Q, Shea TP, Gujja S, Zedlack C, Gasser O, Brander C, Hess C, Günthard HF, Brumme ZL, Brumme CJ, Bazner S, Rychert J, Tinsley JP, Mayer KH, Rosenberg E, Pereyra F, Levin JZ, Young SK, Jessen H, Altfeld M, Birren BW, Walker BD, Allen TM. Whole genome deep sequencing of HIV-1 reveals the impact of early minor variants upon immune recognition during acute infection. PLoS Pathog. 2012;8(3):e1002529.  347.  Glenn TC. Field guide to next-generation DNA sequencers. Mol Ecol Resour. 2011;11(5):759–69.  348.  Bentley DR, Balasubramanian S, Swerdlow HP, Smith GP, Milton J, Brown CG, Hall KP, Evers DJ, Barnes CL, Bignell HR, Boutell JM, Bryant J, Carter RJ, Cheetham RK, Cox AJ, Ellis DJ, Flatbush MR, Gormley NA, Humphray SJ, Irving LJ, Karbelashvili MS, Kirk SM, Li H, Liu X, Maisinger KS, Murray LJ, Obradovic B, Ost T, Parkinson ML, Pratt MR, Rasolonjatovo IMJ, Reed MT, Rigatti R, Rodighiero C, Ross MT, Sabot A, Sankar SV, Scally A, Schroth GP, Smith ME, Smith VP, Spiridou A, Torrance PE, Tzonev SS, Vermaas EH, Walter K, Wu X, Zhang L, Alam MD, Anastasi C, Aniebo IC, Bailey DMD, Bancarz IR, Banerjee S, Barbour SG, Baybayan PA, Benoit VA, Benson KF, Bevis C, Black PJ, Boodhun A, Brennan JS, Bridgham JA, Brown RC, Brown AA, Buermann DH, Bundu AA, Burrows JC, Carter NP, Castillo N, Catenazzi MCE, Chang S, Cooley RN, Crake NR, Dada OO, Diakoumakos KD, Dominguez-Fernandez B, Earnshaw DJ, Egbujor UC, Elmore DW, Etchin SS, Ewan MR, Fedurco M, Fraser LJ, Fuentes Fajardo KV, Furey WS, George D, Gietzen KJ, Goddard CP, Golda GS, Granieri PA, Green DE, Gustafson DL, Hansen NF, Harnish K, Haudenschild CD, Heyer NI, Hims MM, Ho JT, Horgan AM, Hoschler K, Hurwitz S, Ivanov DV, Johnson MQ, James T, Jones TAH, Kang G-D, Kerelska TH, Kersey AD, Khrebtukova I, Kindwall AP, Kingsbury Z, Kokko-Gonzales PI, Kumar A, Laurent MA, Lawley CT, Lee SE, Lee X, Liao AK, Loch JA, Lok M, Luo S, Mammen RM, Martin JW, McCauley PG, McNitt P, Mehta P, Moon KW, Mullens JW, Newington T, Ning Z, Ng BL, Novo SM, O’Neill MJ, Osborne MA, Osnowski A, Ostadan O, Paraschos LL, Pickering L, Pike AC, Pike AC, Pinkard DC, Pliskin DP, Podhasky J, Quijano VJ, Raczy C, Rae VH, Rawlings SR, Rodriguez AC, Roe PM, Rogers J, Rogert Bacigalupo MC, Romanov N, Romieu A, Roth RK, Rourke NJ, Ruediger ST, Rusman E, Sanches-Kuiper RM, Schenker MR, Seoane JM, Shaw RJ, Shiver MK, Short SW, Sizto NL, Sluis JP, Smith MA, Sohna Sohna JE, Spence EJ, Stevens K, Sutton N, Szajkowski L, Tregidgo CL, Turcatti G, VandeVondele S, Verhovsky Y, Virk SM, Wakelin S, Walcott GC, Wang J, Worsley GJ, Yan J, Yau L, Zuerlein M, Rogers J, Mullikin JC, Hurles ME, McCooke NJ, West JS, Oaks FL, Lundberg PL, Klenerman D, Durbin R, Smith AJ. Accurate whole human genome sequencing using reversible terminator chemistry. Nature. 2008;456(7218):53–9.  349.  Shendure J, Porreca GJ, Reppas NB, Lin X, McCutcheon JP, Rosenbaum AM, Wang MD, Zhang K, Mitra RD, Church GM. Accurate multiplex polony sequencing of an evolved bacterial genome. Science. 2005;309(5741):1728–32.  350.  McKernan KJ, Peckham HE, Costa GL, McLaughlin SF, Fu Y, Tsung EF, Clouser CR, Duncan C, Ichikawa JK, Lee CC, Zhang Z, Ranade SS, Dimalanta ET, Hyland FC, Sokolsky TD, Zhang L, Sheridan A, Fu H, Hendrickson CL, Li B, Kotler L, Stuart JR, Malek JA, Manning JM, Antipova AA, Perez DS, Moore MP, Hayashibara KC, Lyons MR, Beaudoin RE, Coleman BE, Laptewicz MW, 197  Sannicandro AE, Rhodes MD, Gottimukkala RK, Yang S, Bafna V, Bashir A, MacBride A, Alkan C, Kidd JM, Eichler EE, Reese MG, De La Vega FM, Blanchard AP. Sequence and structural variation in a human genome uncovered by short-read, massively parallel ligation sequencing using two-base encoding. Genome Res. 2009;19(9):1527–41.  351.  Van Dijk EL, Auger H, Jaszczyszyn Y, Thermes C. Ten years of next-generation sequencing technology. Trends Genet. 2014;30(9):418–26.  352.  Eid J, Fehr A, Gray J, Luong K, Lyle J, Otto G, Peluso P, Rank D, Baybayan P, Bettman B, Bibillo A, Bjornson K, Chaudhuri B, Christians F, Cicero R, Clark S, Dalal R, Dewinter A, Dixon J, Foquet M, Gaertner A, Hardenbol P, Heiner C, Hester K, Holden D, Kearns G, Kong X, Kuse R, Lacroix Y, Lin S, Lundquist P, Ma C, Marks P, Maxham M, Murphy D, Park I, Pham T, Phillips M, Roy J, Sebra R, Shen G, Sorenson J, Tomaney A, Travers K, Trulson M, Vieceli J, Wegener J, Wu D, Yang A, Zaccarin D, Zhao P, Zhong F, Korlach J, Turner S. Real-time DNA sequencing from single polymerase molecules. Science. 2009;323(5910):133–8.  353.  Travers KJ, Chin CS, Rank DR, Eid JS, Turner SW. A flexible and efficient template format for circular consensus sequencing and SNP detection. Nucleic Acids Res. 2010;38(15):e159.  354.  Grant RM, Kuritzkes DR, Johnson VA, Mellors JW, Sullivan JL, Swanstrom R, D’Aquila RT, Van Gorder M, Holodniy M, Lloyd Jr RM, Reid C, Morgan GF, Winslow DL. Accuracy of the TRUGENE HIV-1 genotyping kit. J Clin Microbiol. 2003;41(4):1586