UBC Faculty Research and Publications

Comparative genomic, transcriptomic, and proteomic reannotation of human herpesvirus 6 Greninger, Alexander L; Knudsen, Giselle M; Roychoudhury, Pavitra; Hanson, Derek J; Sedlak, Ruth H; Xie, Hong; Guan, Jon; Nguyen, Thuy; Peddu, Vikas; Boeckh, Michael; Huang, Meei-Li; Cook, Linda; Depledge, Daniel P; Zerr, Danielle M; Koelle, David M; Gantt, Soren; Yoshikawa, Tetsushi; Caserta, Mary; Hill, Joshua A; Jerome, Keith R Mar 20, 2018

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

Item Metadata


52383-12864_2018_Article_4604.pdf [ 2.19MB ]
JSON: 52383-1.0364401.json
JSON-LD: 52383-1.0364401-ld.json
RDF/XML (Pretty): 52383-1.0364401-rdf.xml
RDF/JSON: 52383-1.0364401-rdf.json
Turtle: 52383-1.0364401-turtle.txt
N-Triples: 52383-1.0364401-rdf-ntriples.txt
Original Record: 52383-1.0364401-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessComparative genomic, transcriptomic, andproteomic reannotation of humanherpesvirus 6Alexander L. Greninger1,2* , Giselle M. Knudsen3, Pavitra Roychoudhury1,2, Derek J. Hanson1, Ruth Hall Sedlak1,Hong Xie1, Jon Guan1, Thuy Nguyen1, Vikas Peddu1, Michael Boeckh2, Meei-Li Huang1, Linda Cook1,Daniel P. Depledge4, Danielle M. Zerr5, David M. Koelle1, Soren Gantt6, Tetsushi Yoshikawa7, Mary Caserta8,Joshua A. Hill2 and Keith R. Jerome1,2AbstractBackground: Human herpesvirus-6A and -6B (HHV-6) are betaherpesviruses that reach > 90% seroprevalence in theadult population. Unique among human herpesviruses, HHV-6 can integrate into the subtelomeric regions ofhuman chromosomes; when this occurs in germ line cells it causes a condition called inherited chromosomallyintegrated HHV-6 (iciHHV-6). Only two complete genomes are available for replicating HHV-6B, leading to numerousconflicting annotations and little known about the global genomic diversity of this ubiquitous virus.Results: Using a custom capture panel for HHV-6B, we report complete genomes from 61 isolates of HHV-6B fromactive infections (20 from Japan, 35 from New York state, and 6 from Uganda), and 64 strains of iciHHV-6B (mostlyfrom North America). HHV-6B sequence clustered by geography and illustrated extensive recombination. MultipleiciHHV-6B sequences from unrelated individuals across the United States were found to be completely identical,consistent with a founder effect. Several iciHHV-6B strains clustered with strains from recent active pediatricinfection. Combining our genomic analysis with the first RNA-Seq and shotgun proteomics studies of HHV-6B, wecompletely reannotated the HHV-6B genome, altering annotations for more than 10% of existing genes, withmultiple instances of novel splicing and genes that hitherto had gone unannotated.Conclusion: Our results are consistent with a model of intermittent de novo integration of HHV-6B into host germlinecells during active infection with a large contribution of founder effect in iciHHV-6B. Our data provide a significantadvance in the genomic annotation of HHV-6B, which will contribute to the detection, diversity, and control of this virus.Keywords: Herpesvirus, Human herpesvirus 6, HHV-6B, HHV-6A, Betaherpesvirus, Viral genomics, Herpesvirus genomics,Comparative genomics, Genomic annotation, iciHHV-6BackgroundHHV-6 is a ubiquitous betaherpesvirus that is dividedinto two species (HHV-6A and -6B) [1]. HHV-6Binfects > 90% of children by 2 years of age, causingroseola, also called exanthem subitem or sixth disease,which is the leading cause of febrile seizures among chil-dren [2–5]. The virus persists in multiple cell types withconsistent detectable viral DNA in saliva. HHV-6B reacti-vates in approximately 50% of allogeneic hematopoietic celltransplant (HCT) patients and is the most common causeof encephalitis in this setting. HHV-6B has also been associ-ated with graft-versus-host disease, hepatitis, pneumonitis,and mortality after HCT, although causality remains to beproven [2].Like other human herpesviruses, HHV-6A and -6Bestablish lifelong latency, but unique among humanherpesviruses, they have the ability to integrate intohuman chromosomes. When this integration occurs in agerm cell, the virus can be passed to offspring and* Correspondence: agrening@uw.edu1Department of Laboratory Medicine, University of Washington, Seattle, WA,USA2Fred Hutchinson Cancer Research Center, Seattle, WA, USAFull list of author information is available at the end of the article© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Greninger et al. BMC Genomics  (2018) 19:204 https://doi.org/10.1186/s12864-018-4604-2results in inherited chromosomally integrated HHV6(iciHHV6). Affected individuals have a copy of the virusin each of their cells and the ability to pass on the inte-grated state to 50% of their offspring. IciHHV-6 ispresent in 0.5-2% of the population, constituting almost70 million people worldwide, with the majority of thesebeing iciHHV-6B [6]. iciHHV-6 can also be passed be-tween individuals via transplantation [7, 8]. IciHHV-6was recently associated with an increased risk of acutegraft versus host disease and CMV viremia in HCT pa-tients [9]. Integrated virus has been shown to reactivateboth in vitro and in vivo and can confound assays foractive HHV-6 infection [10]. The mechanism of integra-tion and viral proteins required for integration areunclear [11].Clinical testing for HHV-6 has hitherto been reservedto large academic medical centers and reference labsdue to concerns over reactivation in HCT patients. Un-biased metagenomic sequencing has uncovered HHV-6infection in a number of cases of encephalitis and febrileillness that were previously “unsolved” [12–14]. Of note,HHV-6 has recently been included in new, rapid, point-of-care multiplex PCR panels for meningitis/encephalitisand febrile illness [15]. Given the ease of use and extra-ordinarily rapid turn-around time of these multiplexPCR panels, they have already been adopted by thou-sands of hospitals across the world [15–17]. Because it isnot uncommon for children to have HHV-6B in theircerebrospinal fluid around the time of primary infection,we expect the coming years to see hundreds of thou-sands of HHV-6 infections detected that previouslywould have gone undetected based on the sheer numberof samples that will be tested for HHV-6 [18]. With somany new infections detected, there is an increasingneed to understand the clinical associations, sequencediversity, and basic biology of this virus.To date, only two complete genomes from replicatingHHV-6B are available – the Z29 type strain from Zaireand the HST strain from Japan – and limited compara-tive genomics studies have been conducted for HHV-6B[19–21]. These two genomes have multiple conflictingannotations for gene and protein products. In addition,the annotated gene functions are mostly based on hom-ology from cytomegalovirus, another human betaherpes-virus. Gene boundaries, protein sequences, and diversityof strains across time, place, and iciHHV-6 status arerelatively unknown. These factors are critical for beingable to perform molecular mechanism studies of viralpathogenesis [22].Given that so little is known about HHV-6 genomediversity, gene/protein annotation, and gene/proteinfunction despite its clinical disease associations, thereis an opportunity to use agnostic technologies to rap-idly annotate the HHV-6 genome. Large scale genomesequencing, RNA-Seq, and ribosome profiling have previ-ously been conducted in other human herpesviruses todiscover new genes and proteins and to ascribe novelfunctions to known genes of these obligate intracellularparasites [23–27]. Here we report the results of the firstlarge-scale genome sequencing effort for HHV-6B with125 near complete genomes along with reannotation ofthe genome with comparative genomics, transcriptomics,and proteomics. The results reveal limited sequence diver-sity among HHV-6B sequences with geographical cluster-ing of HHV-6B sequences from acute infections andidentical iciHHV-6B sequences among individuals withoutknown recent common ancestry. RNA sequencing andshotgun proteomics combined with comparative genomicanalysis enabled a consensus re-annotation of HHV-6Bgene products that will serve as a resource for futureclinical and basic science studies of HHV-6B.ResultsGlobal genomic diversity of HHV-6In order to understand the genomic diversity of HHV-6,we performed capture sequencing of 125 strains ofHHV-6B, comprised of 20 viral isolates from Japan, 35isolates from New York, 6 strains from Uganda, and 74strains of iciHHV-6 (64 species B, 10 species A) fromHCT recipients or donors in Seattle (Fig. 1a, Table 1).The HHV-6B oligonucleotides designed for capture se-quencing could retrieve > 99% of the HHV-6B genome,with less than 1% unresolved due to repetitive elements.The same panel was able to retrieve approximately 80%of the HHV-6A genome, again due to repetitive ele-ments and in this case the reduced sequence identitywith the HHV-6B oligonucleotide set (Fig. 1b). Acrossthe HHV-6B strains, the recoverable contiguous HHV-6B unique (U) region measured 119.6 kb, the N-terminal U86 contig measured 3.1 kb, the U90/91contig measured 6.0 kb, and the U94-U100 contigmeasured 10.2 kb. The 10 HHV-6A strains assembledranged from lengths of 60 kb to 119 kb with a medianlength of 118 kb.Demographic characteristics of cohortsThe median age of the roseola cohort from Japan was12 months [8 - 24 months], the New York febrile in-fant cohort was 11 months [1 – 25 months], and theUganda cohort was 25 months. All 20 patients fromthe two cohorts from Japan were of Japanese ancestryand all 6 patients from the Uganda cohort were BlackAfricans. In the New York cohort 16/35 (45.7%) ofpatients were Caucasian, 8/35 (22.9%) were African-American, 4/35 (11.4) were Hispanic, 1/35 (2.9%) wereAsian, and 6/35 (17.1%) were of unknown ethnicity. Ofthe iciHHV-6 samples sequenced, 68/74 (91.9%) ofpatients came from the United States, while 2 patientsGreninger et al. BMC Genomics  (2018) 19:204 Page 2 of 17came from the United Kingdom, 2 patients from Germany,and 1 from Australia (Additional file 1: Table S1). The me-dian age of iciHHV-6B individuals sequenced was 40 years[1 – 68 years] and 57 years [21 – 63 years] for iciHHV-6Aindividuals (Table 1).Comparison of HHV-6A and HHV-6BPhylogenetic analysis of a 40.2 kb segment ranging fromU18 to U41 that could be captured in both the HHV-6Aand HHV-6B strains sequenced in this study demon-strated separate clustering of the HHV-6A and HHV-6Bstrains, consistent with their designation as uniquespecies of human herpesviruses (Fig. 1c). Recombinationanalyses using all 10 iciHHV-6A partial sequences,HHV-6A type strain, and 14 selected HHV-6B sequencescbaFig. 1 Experimental set up and HHV-6 genome calling mock up. a A total of 129 HHV-6 specimens comprised of 55 cultured HHV-6B strains fromacute infections, 6 clinical samples from acute infections, and 64 iciHHV-6B and 10 iciHHV-6A cell lines were sequenced using a capture panelbased on the HHV-6B reference genome (NC_000898). b Consensus genomes used for phylogenetic analysis were called for regions outside ofthe repeat regions for the HHV-6B specimens, including the unique long region (119 kb), U90/91 region (6 kb, between R2 and R3 repeats), andU94-100 (10 kb, between R3 and DR-R repeat). The HHV-6B capture panel recovered much of the unique long region from HHV-6A specimens. cOverall phylogeny of 40.2 kb sequence that was recovered from HHV-6A and HHV-6B strains sequenced reveals separation of HHV-6A and HHV-6B as separate herpesvirus species. Location images purchased from Adobe StockTable 1 Summary of Samples Sequenced in This StudyLocation Number Species Clinical Median AgeJapan 10 HHV-6B Acute, BMT 28 yrs [3 - 64 yrs]Japan 10 HHV-6B Acute, Exanthemsubitum1 yr [8 mo - 3 yrs]New York 35 HHV-6B Acute, fever 11 mo [1-25 mo]Fred Hutch 64 HHV-6B iciHHV-6, BMT 40 yrs [1 - 68 yrs]Fred Hutch 10 HHV-6A iciHHV-6, BMT 57 yrs [21 - 63 yrs]UW Virology 11 HHV-6B reactivation, acute unknownUganda 6 HHV-6B Primary unknownAbbreviations: BMT bone marrow transplantGreninger et al. BMC Genomics  (2018) 19:204 Page 3 of 17revealed no recombination sites between HHV-6A andHHV-6B sequences. Two individuals from Germany andthe United States who shared no relations were found tohave identical iciHHV-6A sequences. HHV-6A sequencesshowed little divergence in this 40.2 kb region with 98.4%of sites having no nucleotide variants. When just compar-ing the maximum divergence between iciHHV-6A andici-HHV6B sequences across the 40.2 kb region, iciHHV-6A strains showed greater maximal pairwise divergencethan iciHHV-6B strains (354 versus 68 SNPs).Sequence divergence in HHV-6BPhylogenetic analysis of the unique long region revealeda cluster of two viruses from Uganda and one from NewYork NY310 that comprise the most divergent HHV-6Bviruses sequenced to date (Fig. 1c). NY310 most closelyaligned to the Z29 strain, differing from the Z29 strainby 644 of 119,635 sites (0.54%). NY310 showed greatergenetic distance to the next closest American strainNY434 (703 sites, 0.59%). This strain had no obviousunique demographic or clinical characteristics, havingbeen derived from an 18-month old white male withfever after only 2 passages in culture (Additional file 1:Table S1). NY310 served as the outgroup for all subse-quent phylogenetic analyses of HHV-6B genomes.Overall, the 119.6 kb HHV-6B unique long contigshowed remarkably little sequence divergence with98.1% of sites being identical among all 127 HHV-6B ge-nomes and an overall pairwise identity of > 99.9%between strains. The prototypical typing gene U90 wasthe most divergent with changes at 6.14% of total sites,while U15 and B6 repeat genes had the least amount ofdivergence with only 0.52% and 0.41% of sites being di-vergent (Fig. 2). Average nucleotide diversity of theUgandan HHV-6B U sequences was 7 times that of theiciHHV-6B strains, which had the least amount of diver-sity of all strains profiled (Table 2). The Japanese isolateswere approximately 40% less diverse than New York iso-lates. Both Achaz and Tajima neutrality tests to test fornon-random sequence evolution across the entire U re-gion were strongly negative in all cohorts, likely due tothe population structure and demographic history of thesamples analyzed [28, 29].Phylogenetic analysis of HHV-6B sequences from acuteinfectionsThe continual replication of HHV-6B virus present inacute infections suggests a different evolutionary historythan that of iciHHV-6B sequences, which are potentiallypreserved over time due to the high fidelity of humangenomic replication. Phylogenies of the 119.6 kb uniquelong region of the HHV-6B genomes revealed clusteringby geography as well as by sample type (i.e. acute HHV-6B infection versus iciHHV-6B) (Fig. 3). Strains fromactive New York infections demonstrated the greatestamount of sequence divergence with at least two differentclusters while Japanese strains all clustered together, in-cluding the HST strain reference sequence. Three of theFig. 2 Nucleotide diversity by gene. Non-identical sites listed by gene by percent of sites with any variance. The prototypical typing gene U90 isthe most divergent with changes at 6.14% of total sites, while U15 and B6 repeat genes had the least amount of divergence with only 0.52% and0.41% of sites being divergentGreninger et al. BMC Genomics  (2018) 19:204 Page 4 of 17Uganda sequences clustered among one of the New Yorkstrain clusters, while three others formed a unique clade(Fig. 3). Sequence from the only known patient of Asiandescent from New York (NY379) fell into the Japanesecluster along with two additional New York HHV-6Bstrains.Identical iciHHV-6B strains in unrelated individualsIciHHV-6B strains showed remarkable relatedness amongunrelated individuals. Across 62 of the 64 iciHHV-6B Uregions sequenced here, only 334 of 119,635 (0.28%) siteshad polymorphisms. Among iciHHV-6B HCT recipientswhose donors were first-degree relatives, all 6 pairs hadiciHHV-6B strains that were found to be identical (Fig. 3).Identical iciHHV-6B strains were also found between un-related individuals from Germany and the United States.Notably, resequencing of several of the identicaliciHHV-6B strains from unrelated individuals gaveidentical sequence in 11 of 12 samples, controlling forpossibility of laboratory contamination or sample mix-up (Additional file 2: Figure S1). The lone outlier was astrain with a singular base with a variant allele fre-quency of almost exactly 50% in each sequencing repli-cate. Analysis of the off-target human mitochondrialreads from unrelated individuals revealed unique mito-chondrial SNPs, confirming that these are from unre-lated individuals (data not shown). IciHHV-6B strainswere found intersperse among a New York cluster ofacute infections as well as the Japanese cluster of acuteinfections, and several New York acute infection strainsfell in the iciHHV-6B clusters. Branch lengths weregenerally longer for most of the acute infection strainsindicating greater sequence divergence from commonancestor compared to iciHHV-6B strains.Sequence diversity of HHV-6B in non-U regions, U90-91and U94-100Based on our data showing U90 to be the most divergentgene in HHV-6B, we sequenced an additional 11 U90 se-quences from HHV-6 positive clinical specimens presentin the UW Virology clinical lab. Phylogenies from theU90-91 and U94-100 regions revealed a similar topologyto that of the unique long phylogeny with a few notableexceptions. New York strains again showed the greatestdiversity and the Japanese strains again clustered together.The U90-91 phylogeny showed two Japanese strains (B1and B4), a New York strain (NY40), a UW clinical isolate(UW_AH1), and one iciHHV-6B strain (61C11) that clus-tered with the Z29 type strain and four strains with U90sequence present in Genbank (Fig. 4, Additional file 3:Figure S2B). Additional recent UW clinical strains forwhich U90 sequence was available clustered throughoutthe HHV-6B U90 tree recovered from the whole genomesequence, with one additional prominent outgroup(UW_BF2). The Japanese and New York strains were eachlocated in their respective unique long cluster, while theiciHHV-6B-61C11 strain fell in the Japanese cluster. Thedisparity between the U and U90 phylogenies is evidenceof potential recombination in these strains with HHV-6Bthat is closer to the Zairian Z29 strain or the NY310 out-group. Of note, the Japan-B1 strain also fell in a uniqueposition in the U94-100 region among an iciHHV-6Bcluster, while the NY40 strain was located in the U94-100Japanese cluster (Additional file 3: Figure S2B). In additionto the phylogenetic analysis indicative of recombination,Hudson-Kaplan RM estimates of parsimonious recombin-ation events across the U region ranged from 20 recom-bination sites for iciHHV-6B strains to 103 sites for NewYork strains (Table 2), suggesting widespread recombin-ation within HHV-6B species. No interspecies HHV-6A xHHV-6B recombinants were observed [30].Annotation of HHV-6B via comparative genomicsMultiple gene annotation discrepancies exist betweenthe published HHV-6B Z29 and HST genomes. With theavailability of 125 new HHV-6B genomes, we nextexamined the sites of these annotation discrepancies inour new HHV-6B genome sequences. For instance, theU91 gene contains an annotated splice site in the Z29while no such annotation is found in the HST assembly.Sanger sequencing of U91 cDNA from our lab’s culturedZ29 strain (Z29-1) revealed a different splice site 13 bpaway from the annotated Z29 splice site, adding 5 add-itional amino acids to the middle of the U91 proteinTable 2 Population genomics statistics for U region by cohort sequencedLocus Cohort Samples Nucleotide diversity Achaz Y Tajima’s D H-K sites119 kb U Japan 20 109.5 −0.75 −0.85 25New York 35 177.3 −0.33 −1.72 103Uganda 6 357.3 −0.39 − 0.11 34iciHHV-6B 64 51.1 −0.94 −1.86 20All Us 127 144.9 −1.59 −2.02 16240 kb U 11 HHV-6A129 HHV-6B140 256.0 −1.59 − 1.47 144Greninger et al. BMC Genomics  (2018) 19:204 Page 5 of 17Fig. 3 (See legend on next page.)Greninger et al. BMC Genomics  (2018) 19:204 Page 6 of 17(Fig. 5a). Both the cloned splice site and the annotatedsplice site contained canonical intronic splice sequencing(GU…AG). Cloning of the Z29-1 cDNA with the newsplice site revealed an early stop codon that woulddisrupt the annotated C-terminal half of the protein inZ29 strains. Shotgun genomic sequencing of the cul-tured HHV-6B Z29-1 strain matched the Z29-1 cDNAsequence. Of note, Z29 is the only HHV-6B strain in ourgenomic sequencing with a single adenine insertion nearthe start of the second exon. Using the cloned splice site,all other U91 genes sequenced in this study would be in-frame to the end of the annotated U91, revealing thatZ29 is likely unique among HHV-6B strains in missingthe C-terminal half of U91.Several other annotation discrepancies between exisit-ing HHV-6B sequences could be reconciled with ournew genome sequences. The U12 gene in the Z29 strainis interrupted by a stop codon while the HST strain con-tains one long ORF (Fig. 5b). Comparison with the 126U genome sequences in this study show that for U12,the HST CDS should be considered the more represen-tative of the original two genomes. Alternatively for U27and U52, homopolymeric SNPs in HST creates abnor-mally long and short annotated ORFs, respectively,that are not reflected in the newly sequenced genomes(Fig. 5c/d). Homopolymeric SNPs are also found inthe U83 gene resulting in a polymorphic annotationacross many of the sequence genomes (Fig. 5e).(See figure on previous page.)Fig. 3 Phylogenetic tree of unique long region from HHV-6B samples. HHV-6B genomes were aligned using MAFFT, curated for sequence outsideof repeat regions, and phylogenetic trees were constructed using MrBayes along the 119 kb unique long region. HHV6-6B NY310 was used as anoutgroup. Samples are colored and labeled for origin based on New York (green), Japan (blue), Uganda (purple), or iciHHV6 from HSCT recipientsor their donors in Seattle (black), as well as whether two genomes were recovered from first-degree relatives (red). Location images purchasedfrom Adobe StockFig. 4 Phylogenetic trees of HHV-6B samples in U90 region including UW clinical isolates. HHV-6B U90 sequence captured from the 125 completegenomes and directly PCR-amplified from the UW cohort specimens were aligned using MAFFT and phylogenetic trees were constructed usingMrBayes. Samples are colored and labeled for origin based on New York (green), Japan (blue), Uganda (purple), UW Virology clinical specimens(gold), or iciHHV6/FHCRC (black), as well as whether two genomes were recovered from first-degree relatives (red). Location images purchasedfrom Adobe StockGreninger et al. BMC Genomics  (2018) 19:204 Page 7 of 17Reannotation of HHV-6B genome through RNA-sequencingand shotgun proteomicsBased on the number of discrepancies between HSTand Z29 strain annotation that could be resolved bycomparative genomics, we pursued RNA sequencingof the transcriptome of the HHV-6B Z29 type strain tomore exhaustively reannotate the HHV-6B genome. Twobiological replicates were prepared for the HHV-6B Z29abcdeFig. 5 HHV-6B annotation based on comparative genomics. Differences in annotation between HHV-6B Z29 and HST sequences are comparedwith a subset of the 119 genomes sequenced in this study. a Sanger sequencing of U91 cDNA revealed a different splice site 13-bp upstreamthan that which is annotated in the reference Z29 strain. The aberrant splice-site annotation in Z29 is likely due to a single base insertion foundonly in Z29 that alters the reading frame in the second exon. Genome sequencing of our cultured HHV-6B strain (Z29-1) confirmed the Z29-1cDNA sequence. The reading frame depicted for Z29 is as annotated in the NCBI reference genome (NC_000898). Based on the newly discoveredsplice site, the Z29 U91 would contain an early stop codon while all other U91 sequences obtained in this study would continue the readingframe to the end of U91 as annotated in Z29. Several key different loci in U12 (b), U27 (c), and US52 genes (d) that alter the length of openreading frames in Z29 and HST are depicted. e A homopolymeric polymorphism in U83 changes in expected length and sequence of its openreading frame between different strainsGreninger et al. BMC Genomics  (2018) 19:204 Page 8 of 17RNA-Seq library in SupT1 cells and were sequenced atcoverages of 266X and 3600X, while one strand-specificlibrary was prepared for a HHV-6B Z29 infected MOLT3cells at an average coverage of 5751X. RPKM values forHHV-6 genes from SupT1 replicates were highly reprodu-cible (r2 = 0.92) (Fig. 6a). Compared to the Z29 transcrip-tome in SupT1 cells, the Z29 transcriptome in MOLT3cells demonstrated significantly less correlation (r2 = 0.66)(Fig. 6b). While only 3/104 (2.9%) HHV-6B CDS had 2-fold higher expression in in SupT1 cells compared toMOLT3 cells, 19/104 (18.2%) CDS had greater expressionin MOLT3 cell lines (Fig. 6c).Analysis of the mapped reads revealed a number ofnovel spliceoforms that were present. All splice sitesmapped were perfectly conserved in the 127 HHV-6Bgenomes analyzed. Five of 43 (11.6%) total splice sitesrecovered were non-canonical with 4/5 (80%) non-canonical splice sites occurring in U7-U9 transcripts. Tovalidate these novel spliceoforms and extensions thataffected coding sequences, we performed shotgun massspectometry on 1D gel-separated proteins from HHV-6BZ29 cultured in SupT1 cells (Additional file 4: Figure S3,Additional file 5: Figure S4). Shotgun proteomic analysisproduced 350 unique spectra covering 39 different HHV-6proteins that may be viewed inMS Viewer (Additional file 6:Table S2 and Additional file 7: Table S3).Intriguingly, three novel U79 mRNA isoforms werefound, one of which also demonstrated divergent spli-cing based on culture in SupT1 versus MOLT3 cell lines(Fig. 7). Peptide confirmation of the novel U79 spliceo-form present in SupT1 cells was confirmed with two pep-tides – LSTCEYLK with m/z 507.25 (2+), and YLCVR355.68 (2+) – from shotgun proteomics analysis (Add-itional file 7: Table S3). The U19 gene demonstrated anunannotated splice junction just prior to the annotatedstop codon, extending the C-terminus of the protein by 13amino acids (Fig. 8). Peptides immediately before and afterthe splice junction were recovered, confirming the expres-sion of the C-terminal extension (DFLEEIAN 475.72 (2+)and SPENAVHESAAVLR 493.92 (3+) in Additional file 7:Table S3). Antisense reads along with a novel stop codonwere recovered to the existing U83 annotation (Fig. 9).DiscussionIn this study we sequenced 125 HHV-6B genomes and10 partial HHV-6A genomes, increasing the full genomedata available for HHV-6 by more than an order ofmagnitude. We found remarkably little sequence diver-sity among HHV-6B strains sampled from New York,Seattle, and Japan, with the average strain having fewerthan 150 differences across the 119 kb unique longregion relative to any other strain sequenced here.IciHHV-6B from across the United States had consider-ably less diversity than other cohorts of HHV-6 sampled.HHV-6A and HHV-6B strains sequenced here showedno overlap or recombination between species and themost divergent HHV-6B strain identified to date wasisolated and sequenced. Viral sequences clustered bygeographical origin and identical iciHHV-6B strains werefound among many apparently unrelated individuals.These results suggest that HHV-6B integration is arelatively infrequent event, that iciHHV-6B does notgeneral reflect strains circulating in community causingacute infection, and that sequence diversity may bedriven by a founder effect. Alternatively, certain strainscould be prone to integration. At the same time,iciHHV-6B sequences were found admixed with HHV-6B strains from acute infection, suggesting that integra-tion events are not uncommon. The hypothesis thatHHV-6 integration into the germline is an infrequentevent, however, would be consistent with a foundereffect for each clade of identical iciHHV-6B found acrossour North American patients and account for the identi-cal iciHHV-6 sequences found between two pairs ofa b cFig. 6 RNA Sequencing of Sup-T1 and MOLT3 cell lines asynchronously infected with HHV-6B Z29 type strain. RPKM values for HHV-6B Z29transcripts in biological replicates of virus grown in Sup-T1 cells show excellent reproducibility (a). RPKM values of HHV-6B Z29 transcripts forvirus grown in MOLT3 cells show differences in expression compared to virus grown in Sup-T1 cells (b). List of HHV-6B CDS with > 2-fold absolutevariation in expression in Sup-T1 and MOLT3 cell lines (c). Substantially more HHV-6B genes had higher expression in MOLT3 cells than inSup-T1 cellsGreninger et al. BMC Genomics  (2018) 19:204 Page 9 of 17individuals from different sides of the Atlantic Ocean. Itwould also suggest that chromosomal integration ofHHV-6 into the germ line is an extraordinarily rare eventand most iciHHV-6 individuals acquired their virus froma remote integration event [31]. More sequencing of bothcirculating HHV-6 strains and iciHHV-6 individuals isneeded to test this hypothesis and will no doubt becomeavailable as more human genomes are sequenced. Thehypothesis that integration bias due to viral sequence isthe cause of the degeneracy of iciHHV-6 genomes is diffi-cult to separate from founder effect and would only betestable in vitro or by following many individuals acutelyinfected with different strains of HHV-6B.Despite widespread recombination, phylogenetic ana-lyses demonstrated geographical clustering of HHV-6Bstrains with unique clades for Japanese strains and forab cFig. 7 Alternative and differential splicing of HHV-6B U79 transcripts in Sup-T1 versus MOLT3 cells. Strand-specific RNA sequencing reveals threeadditional spliceoforms of the U79 gene in HHV-6B Z29 strain cultured in Sup-T1 cells compared with the Z29 reference annotation inNC_000898 (a). Reads depicted in orange are positive-sense reads, while negative-sense reads are shown in blue. The highlighted peptide fromthe U79a2 transcript in red was confirmed by shotgun proteomics of the Sup-T1 cultured HHV-6B. While in SupT1 four total spliceoforms arefound (b), in MOLT3 cells, only two forms of splicing in U79 are detected (c)Fig. 8 Unannotated splicing leading to C-terminal extension of HHV-6B U19 protein. Strand-specific RNA sequencing of HHV-6B cultured inSup-T1 cells demonstrated a novel splice site at the 3′ end of the U19 transcript in the codon immediately before the annotated stop codon. Thenew splice site leads to a 13 amino acid C-terminal extension, which was confirmed by shotgun proteomicsGreninger et al. BMC Genomics  (2018) 19:204 Page 10 of 17several of the New York strains. Of note, the only pa-tient of Asian descent in the New York cohort alignedbest to the Japanese strains. These data would be con-sistent with the hypothesis of a familial source of trans-mission of acute HHV-6B. Because of the clustering ofNew York and Japan HHV-6 sequences, we are unableto ascertain whether strain differences can account forthe striking differences in reported rates of encephalitisin infants with primary HHV-6 infection between Japanand the United States [32].The geographical cluster of HHV-6B is similar tothat seen for HSV-1 and HSV-2 genome sequences,which also show high degrees of interspecies recom-bination [25, 26, 33]. The limited diversity of HHV-6Bas measured by average pairwise nucleotide diversity iscomparable to that found in HSV-2 in contrast to thatidentified in HSV-1 strains [25, 26]. Of note, the diver-sity seen in HHV-6B is substantially less than thatseen for the phylogenetically related human betaher-pesvirus CMV (HHV-5) [27]. No comparative genom-ics have been performed to date on the other humanbetaherpesvirus HHV-7.Limitations of our approach include the limited world-wide sampling of HHV-6B strains, which included theUganda, Japan, and the United States (with samples inthe iciHHV-6 Fred Hutchinson cohort coming fromseveral northern European individuals and only oneAustralian individual). Of note, our North AmericaniciHHV-6 sequencing included individuals from at least25 different states. More strains from both acutely in-fected and iciHHV-6 individuals are needed from Asia,the Middle East, South America, and Africa. Given thediversity seen in a limited subset of Ugandan strains andthe limited diversity seen in iciHHV-6 in our study, itwould be worthwhile to sequence iciHHV-6 from Africanpopulations to test hypotheses on the contribution offounder effect and strain sequence effects on HHV-6 inte-gration. Sequencing of the U90 gene from reactivatedHHV-6B strains from our clinical lab revealed additionallineages of HHV-6, which were subsequently confirmedby sequencing Ugandan HHV-6 isolates. Our clinical U90sequences indicate even more lineages exist that we havenot sampled on a genome-wide basis.We also were not able to sequence through everyrepeat in the virus and thus our estimates of diversitywould be biased to the null given that the repetitive ele-ments may be one of the first sites of genome evolution.We also were not able to recover near-complete genomesof HHV-6A due to the use of a HHV-6B capture panel forsequencing. Future studies should be focused on continu-ing to probe the global diversity of HHV-6 sequences,understanding the degree of admixture between acuteFig. 9 Antisense transcription and novel splicing of HHV-6B U83 gene. Nearly all of the strand-specific RNA-seq reads from Sup-T1 cells at theannotated HHV-6B Z29 U83 gene were antisense to the existing annotation and included a novel splice site. The same splice site in the contextof antisense transcript predominance was recovered from virus cultured in MOLT3 cells. No high-confidence peptides were recovered to thisalternatively spliced antisense transcript by shotgun proteomicsGreninger et al. BMC Genomics  (2018) 19:204 Page 11 of 17infections and iciHHV-6 strains, and whether genotypesidentified here are associated with different clinical out-comes. Based on the results presented here, there was noclear association between viral sequence and clinical phe-notypes such as CNS symptoms, although our power todetect such differences was limited. Future studies willalso be required to test the contribution of human SNPsand genetic diversity to any associations found betweeniciHHV-6 sequences and clinical phenotypes.Our RNA-sequencing data found novel spliceoforms andantisense transcripts in 10% of the genes currently anno-tated in HHV-6B Z29. These data were limited by the useof a single transcriptome replicate for MOLT3 cells, al-though we note biological replicates were highly correlatedin SupT1 cells. Shotgun proteomic analysis recovered pep-tides for three changes in HHV-6B coding sequences andconfirmed expression of 39 existing proteins in lytic HHV-6B infection. We also discovered differential splicing ofU79 in SupT1 versus MOLT3 cells. These data allow forthe most comprehensive annotation of an HHV-6 genometo date and will allow for confident study of HHV-6protein-protein interactions [22, 34]. Certainly, more workis also required to characterize how the novel spliceoforms,extensions, and transcripts discovered here affect viralreplication and gene function, and whether they arepresent in the many strains sequenced here.ConclusionsThe sequences recovered here represent by far the lar-gest HHV-6 sequencing effort conducted to date andsignificantly increases the number of available genomesfor HHV-6B. Using these data, we propose a model ofintermittent de novo integration of HHV-6B into hostgermline cells during active infection with a large contri-bution of founder effect in iciHHV-6B. Our data providea significant advance in the genomic annotation ofHHV-6B, which will contribute to the detection, diver-sity, and control of this virus. By building consensusgene and protein annotations, immediate outcomes in-formed by the experiments detailed here have includedthe development of a HHV-6B ORFeome that will en-able downstream studies in gene function and T-cell epi-tope and antigen discovery and the design of RT-PCRprimers and RNA-ISH probes to target highly expressedgene to test clinical samples for HHV-6 reactivation insitu. These data also underscore the continual need forgenome sequences to achieve consensus annotation forunderstanding microbial biology [35].MethodsCollection of specimensNew York cohortThirty five HHV-6B viral isolates were obtained fromperipheral blood samples from children under 3 years ofage with acute febrile illnesses or seizures presenting tothe University of Rochester Medical Center EmergencyDepartment or ambulatory settings in Rochester, NY, aspreviously described [4, 18, 36, 37]. Samples from chil-dren with a known abnormality of immune functionwere excluded.Peripheral blood mononuclear cells (PBMCs) wereseparated from EDTA anticoagulated blood samples viadensity gradient centrifugation (Histopaque 1077;SigmaDiagnostics, St. Louis, Mo.), and co-cultivated with stim-ulated cord blood mononuclear cells. Positive cultureswere identified by characteristic cytopathic effect (CPE),confirmed by indirect immunofluorescent staining withmonoclonal antibodies directed against HHV-6A andHHV-6B, and polymerase chain reaction, as previouslydescribed [4, 38].Japanese cohortHHV-6B was isolated from PBMCs obtained from 10 ESpatients and 10 HSCT recipients by co-cultivation withstimulated cord blood mononuclear cells. Infected cul-tures were identified on the basis of cytopathic effect(i.e., characteristics of pleomorphic, balloon-like largecells). The presence of virus was confirmed by immuno-fluorescence staining of the co-cultures with a specificHHV-6B monoclonal antibody (OHV-3; provided by T.Okuno, Department of Microbiology, Hyogo College ofMedicine, Hyogo, Japan). Co-cultivated cord blood mono-nuclear cells infected with the clinical isolates were storedafter several passages at − 80 °C until assayed.Uganda cohortSaliva samples were obtained from infants in a previouslydescribed birth cohort study of primary herpesvirus infec-tion [39]. Acute HHV-6B infection determined by weeklyPCR testing of oral swabs. Whole saliva was collectedevery 4 months using the Salivette® collection system(Sarstedt), transferred to cryovials, and frozen at − 80 °Cuntil assayed. The samples used for this study were from 2infants (both 3 months old at the time of sampling), 3older children (ages 2.1 years, 2.8 years, and 4.2 years),and 1 adult (age unknown).IciHHV-6 cohortSeventy four individuals with iciHHV-6A or -6B were iden-tified as part of a continuation of a previously describedstudy [40]. DNA was extracted from beta lymphoblastoidcell lines (LCLs) generated from Epstein-Barr virus in-fected peripheral blood mononuclear cells (PBMCs) ob-tained from hematopoietic cell transplant recipients anddonors. Patients received HSCTs at Fred HutchinsonCancer Research Center (FHCRC) in Seattle, WA. Donorswere sourced from patient relatives and international bonemarrow donor registries. We then used a pooling testingGreninger et al. BMC Genomics  (2018) 19:204 Page 12 of 17strategy as previously described using quantitative PCR[41] and droplet digital PCR [42] to identify individualswith iciHHV-6. A conserved region of the U94 genewas amplified to distinguish between species HHV-6Aand HHV-6B.University of Washington Virology patient cohortSamples from 21 different individuals previously foundto be HHV-6 PCR positive were randomly selectedfrom plasma submitted for testing in the ClinicalMolecular Virology Laboratory at the University ofWashington in 2014-2015. The majority of sampleswere from post-transplant-associated testing for suspectedHHV-6 systemic infections. Eleven samples were fromchildren < 16 years of age (3-16 years old), and 10 werefrom adults (17-51 years old). Samples were from 13males and 8 females. Five samples had viral loads < 1000copies/mL (910, 740, 720, 550, and 480) while theremaining viral loads ranged from 1000 to 53,000 c/mL.Of these 11 gave sufficient sequence after nested PCR tobe included in downstream analyses.DNA extraction and quantitative PCR and U90 sequencingApproximately 5 μg of DNA were extracted from B-LCLs with iciHHV-6 and aliquoted at concentrationsof ~ 200 ηg/μL. DNA from the Japan and New Yorkstrains was extracted from 200 ul of viral cultureusing QIAamp 96 DNA kit (Qiagen) and eluted into 100 ulof AE buffer (Qiagen). To quantify the amount of HHV-6and human DNA, 10ul of purified DNA was used to per-form real-time quantitative PCR as described previously[40]. Plasma samples from the University of Washingtonpatient cohort were extracted using a MagnaPure LC(Roche) and MagnaPure LC DNA Isolate Kit with a start-ing volume of 200uL and elution volume of 100uL. TheU90 locus was amplified using a nested PCR protocol asdescribed previously [43]. Amplicons from the same pa-tient were pooled, diluted, and next-generation sequencinglibraries were created using the Nextera XT kit.Sequencing of U91 RNA transcriptSeven million HHV6B (Z29)-infected SupT1 cells(from NIH AIDS Reagent Program) were used as start-ing material to create an RNA library with the QiagenRNeasy Mini Kit according to manufacturer’s instruc-tions. Total RNA was treated with TURBO DNase I(Thermo Fisher Scientific) and then used to create acDNA library with SuperScript II Reverse Transcript-ase (Thermo Fisher Scientific) according to manufac-turer’s instructions. Using this cDNA as template andPlatinum Taq DNA Polymerase High Fidelity (ThermoFisher Scientific), the U91 transcript was amplified byPCR with annealing temperature of 55.5 °C for 30 cycleswith primers that included cloning recognition sequencesas follows: U91 sense, 5’-GGGGACAAGTTTGTACAAAAAAGCAGGCTTCTCTGTAACACTGATCATGATGGGATATGAGGA-3′; U91 antisense, 5’-GGGGACCACTTTGTACAAGAAAGCTGGGTCTTACACATTCATTTCAGTTTTCGGTATAATAGCCTC-3′. This PCR productwas inserted into the pDONR221 cloning vector (ThermoFisher Scientific) and Sanger sequenced using the M13F(− 21) and M13R primers.Capture sequencingSequencing libraries for the New York, Japan, andiciHHV-6 cohorts were prepared using 100 ng of gen-omic DNA using either NEB fragmentase, end repair/dAtailing, Y-adapter ligation, and dual-index Truseq PCRbased or via the Kapa HyperPlus kit, following manufac-turer’s protocol [44]. Approximately 60 ng of cleaned,amplified DNA library was pooled into sets of seven oreight samples based on relative viral qPCR to humanbeta-globin qPCR ratio, so that samples with similarrelative concentrations of virus were pooled together[45]. Capture sequencing was performed following theIDT xGen protocol with the use of half the amount ofblocking adapter and at least 4 h of 65C hybridization witha tiling biotinylated oligo capture library based on thereference HHV6-B genome (NC_000898). Post-capturelibraries were sequenced to achieve at least 200,000 readsper sample library (at least 100X coverage based on atleast 50% on-target) on a 1x180bp single-end run or on a300x300bp paired-end run on an Illumina MiSeq.Capture sequencing for the Uganda cohort (n = 6 sam-ples) was performed using a custom-designed SureSelectXToligonucleotide panel covering HHV-6 and HHV-7 ge-nomes and sequenced using an Illumina NextSeq using av2 300 cycle mid-output kit (2x150bp paired end) [46, 47].Libraries were prepared as outlined in the SureSelectXTAutomated Target Enrichment protocol version J0(December 2016) with two minor modifications. 20 ngof total DNA was sheared prior to end-repair, A-tailingand adapter ligation (1:100 dilution). Two extra cyclesof PCR were performed during library amplificationprior to hybridization while four extra cycles of PCRwere added to the post-hybridization amplification /indexing step.RNA-Seq of HHV-6B Z29 strainTotal RNA was extracted from MOLT3 and Sup-T1 cellsasynchronously infected with HHV-6B Z29 strain with> 106 copies/mL of virus in the supernatant. 3μg of totalRNA was used as input for polyA-purification andstrand-specific RNA-Seq libraries were prepared fromusing the NEBNext Ultra Directional RNA Library PrepKit. Two libraries were prepared from infected SupT1cells and one from infected MOLT3 cells. Transcriptomelibraries were sequenced on an Illumina MiSeq usingGreninger et al. BMC Genomics  (2018) 19:204 Page 13 of 17multiple runs types (2x94bp, 1x188bp). RPKM values forHHV-6B genes in both SupT1 and MOLT3 cell lines areavailable in Additional file 6: Table S2.Shotgun proteomicsProteomic samples were prepared from soluble cell ly-sates or serum-free conditioned media from HHV6-infected Sup-T1 cells. HHV-6B quantitation in lysateswas 23,683,766 copies per PCR reaction with a corre-sponding beta-globin copy number of 12,900 copies perreaction; HHV-6B quantitation in the serum-free mediawas 3,122,307 copies per reaction with a correspondingbeta-globin copy number of 10,115 copies per reaction.Approximately 2-20 micrograms of protein were sepa-rated on two 10–20% Criterion Tris-HCl run in MOPSor one 4-12% Criterion Tris-HCl run in MES SDS-PAGEgels (Bio-Rad), silver stained, and gel bands were excisedfor mass spectrometry-based peptide sequencing as de-scribed previously [48, 49] (Additional file 4: Figure S3).Samples were digested with sequencing grade trypsin(Promega) only, or with trypsin followed by AspN (Roche)following the standard UCSF MS facility protocol (http://msf.ucsf.edu/protocols.html) [50].Peptide sequencing was performed using an LTQ-Orbitrap Velos (Thermo) mass spectrometer, equippedwith a 10,000 psi nanoACUITY (Waters) UPLC. Reversedphase liquid chromatography was performed using anEasySpray C18 column (Thermo, ES800, PepMap, 3 μmbead size, 75 μm× 15 cm). The LC was operated at600 nL/min flow rate for loading and 300 nL/min for pep-tide separation over a linear gradient over 60 min from 2%to 30% acetonitrile in 0.1% formic acid. For MS/MSanalysis on the LTQ Orbitrap Velos, survey scans were re-corded over 350-1400 m/z range, and MS/MS HCD scanswere performed on the six most intense precursor ions,with a minimum of 2000 counts. For HCD scans, isola-tion width was 3.0 amu, with 30% normalized collisionenergy. Internal recalibration to a polydimethylcyclosi-loxane (PCM) ion with m/z = 445.120025 was used forboth MS and MS/MS scans [51].Mass spectrometry centroid peak lists were generatedusing in-house software called PAVA, and data weresearched using Protein Prospector software v. 5.19.1[52]. Data were searched with carbamidomethylation ofCys as a fixed modification, and as variable modifica-tions, oxidation of methionine, N-terminal pyrogluta-mate from glutamine, start methionine processing, andprotein N-terminal acetylation. Trypsin, or trypsin plusAspN specificity was chosen as appropriate for each ex-periment. Mass accuracy tolerance was set to 20 ppmfor parent and 30 ppm for fragment masses. For pro-tein identification, searches were performed against a9874 entry database containing all protein sequenceslonger than or equal to 8 amino acids derived fromHHV-6 Z29 strain genomic sequence translated in allsix reading frames combined with translated splicejunctions derived from RNA-Seq data. Searches werealso performed with the SwissProt human database(downloaded September 6, 2016) containing 20,198 en-tries, and fetal bovine serum (P02769) as a cell culturesupplement. Databases were concatenated withmatched, fully randomized versions of each databaseto estimate false discovery rate (FDR) [53].The HHV-6B protein database was searched initiallyallowing for two missed and one non-specific cleavageto allow for peptides with alternative splicing or unpre-dicted start/stop sites. Standard Protein Prospectorscores (minimum protein score 22, minimum peptidescore 15, maximum protein expectation value 0.01 andmaximum peptide expectation value 0.001) produced a5% FDR for protein identifications. All matched HHV6peptide spectra were manually de novo sequenced,and may be viewed with the freely available softwareMS-Viewer, accessible through the Protein Prospectorsuite of software at the following URL: http://prospec-tor2.ucsf.edu/prospector/cgi-bin/msform.cgi?form=ms-viewer, with the search key: 7awn6ehwzd. Raw massspectrometry data files and peak list files have beendeposited at ProteoSAFE (http://massive.ucsd.edu)with accession number MSV000081332 (Additionalfile 7: Table S3 Additional file 8: Table S4).Sequence analysisDNA Sequencing reads were quality and adapter-trimmedusing Trimmomatic v0.36 and Cutadapt, de novo assem-bled using SPAdes v3.7 and mapped to reference genomesNC_000898 and NC_001664 using Bowtie2 [54–56]. Con-tigs were aligned to reference genomes using the multiplealignment program Mugsy v1.2.3 and resolved againstconsensus sequences from mapped reads using customscripts in R/Bioconductor [57–59]. Final assemblies weregenerated after discarding any contigs with mapq <= 5.Assembled genomes were annotated using Prokka anddeposited to Genbank (accession numbers in Additionalfile 1: Table S1).As the sequencing length was not sufficient to regu-larly discern sequence in the direct repeats and acrossseveral of the smaller repeats present in the HHV-6Bgenome, analysis was performed on aligned sequencesthat were pruned to keep four non-repeat-containing re-gions: between R0 and R1 repeats (U), between R1 andR2A repeats (upstream and N-terminal U86 region), be-tween R2B and R3 repeats (containing U90/91 genes),and between U94-U100 genes (Fig. 1). Population gen-omics analyses including nucleotide diversity estimates,Tajima’s D, Achaz’s Y, and Hudson-Kaplan recombin-ation estimates were executed using the PopGenome Rpackage [28, 29, 60]. Recombination detection analysesGreninger et al. BMC Genomics  (2018) 19:204 Page 14 of 17were performed using the DualBrothers package using awindow length of 800 bp and a step size of 100 bp [61].RNA sequencing reads were trimmed using cutadapt andmapped to the HHV-6B Z29 reference genome usingGeneious v9.1 read aligner with structural variant discovery(decreased gap penalty) [62]. RPKM values were calculatedbased on HHV-6B Z29 reference genome annotations anddisplayed using custom scripts in R/Bioconductor.Additional filesAdditional file 1: Table S1. List of samples sequenced in this study andassociated accession numbers. (DOCX 72 kb)Additional file 2: Figure S1. Resequencing of select iciHHV-6Bspecimens confirms identical sequences among unrelated patients.Samples from select iciHHV-6B specimens with identical sequences werere-extracted, re-prepared and re-sequenced from original patient materialto rule out contamination or a sample specimen switch during thesequencing process. 11/12 of specimens gave identical sequencethroughout the unique long region directly from de novo assembly. Onespecimen (iciHHV-6B-30E3) had one nucleotide change (G77564 T) uponresequencing at a base that had a G/T variant allele frequency ofapproximately 50% each time the sample was sequenced. (PDF 145 kb)Additional file 3: Figure S2. Phylogenetic tree of HHV-6B completeU90/91 and U94/100 loci. HHV-6B genomes were aligned using MAFFT,curated for sequence outside of repeat regions, and phylogenetic treeswere constructed using MrBayes along the 6 kb U90/91 (A), and10 kb U94-100 (B) regions. HHV6-6B NY310 was used as an outgroup.Samples are colored and labeled for origin based on New York (green),Japan (blue), or iciHHV6-B from HSCT recipients or their donors in Seattle(black), as well as whether two genomes were recovered from first-degree relatives (red). Location images purchased from Adobe Stock.(ZIP 656 kb)Additional file 4: Figure S3. Non-contiguous gel images of silver stainof HHV-6B Z29 lysates in SupT1 cells or serum-free supernatant run on10-20% TrisHCl gels in MOPS buffer. (PDF 3011 kb)Additional file 5: Figure S4. Gel image of silver stain of HHV-6B Z29lysate in SupT1 cells or serum-free supernatant run on 4-12% TrisHCl gelin MES buffer. (PDF 1335 kb)Additional file 6: Table S2. RPKM values for RNA-Seq data. (XLSX 51 kb)Additional file 7: Table S3. HHV-6 Proteins Identified by ShotgunProteomics. Mass spectrometry database search results are shown forHHV6 proteins identified using Protein Prospector v 5.19.1 as described inMethods. Data were scored at the 5% FDR with Protein and Peptideminimum scores of 22 and 15, and maximum expectation values forproteins and peptides of 0.01 and 0.001, respectively. The number ofunique peptides, the peptide (or spectral) count, the percent sequencecoverage and the best peptide expectation value are given for eachprotein identification, merged from all samples. (XLSX 53 kb)Additional file 8: Table S4. HHV-6 Peptides Identified by ShotgunProteomics. Mass spectrometry database search results are shown forHHV6 peptides identified using Protein Prospector v 5.19.1 described inMaterials and Methods. The table reports the best matched peptidespectra. Provided are the mass to charge ratio (m/z), charge (z), masserror in ppm, the peptide sequence with previous and next amino acidsin the sequence, variable modification, the fraction and retention time asspectrum identifiers. The start and end sequence numbers are given,along with Protein Prospector peptide score and peptide expectationvalue. (XLSX 88 kb)AbbreviationsBMT: Bone marrow transplant; CMV: Cytomegalovirus; CPE: Cytopathic effect;DNA: Deoxyribonucleic acid; HCT: Hematopoietic cell transplant; HHV-6: Human herpesvirus 6; HSCT: Hematopoietic stem cell transplant; iciHHV-6: Inherited chromosomally integrated HHV-6; IRB: Institutional review board;PBMC: Peripheral blood mononuclear cells; PCR: Polymerase chain reaction;RNA: Ribonucleic acid; RPKM: Reads per kilobase of transcript per millionreads; U: UniqueAcknowledgementsMass spectrometry analysis was provided by the UCSF Mass SpectrometryFacility directed by Al Burlingame, supported by the Adelson MedicalResearch Foundation. We appreciate assistance from Alex Yamana, GabbyDolgonos and Krithika Nathamuni for careful inspection of peptide massspectral assignments. We thank Samia Naccache, Nicole Lieberman, JesseBloom for helpful comments on the manuscript.FundingNo specific funding was obtained for this study.Availability of data and materialsAll genomic is publicly available in Genbank at the accessions listed inAdditional file 1: Table S1 and proteomic peak list files have been deposited atProteoSAFE (http://massive.ucsd.edu) with accession number MSV000081332.Authors’ contributionsALG and KRJ designed experiments. ALG, GMK, DJH, RHS, MH, DPD, HX, JG,TN, VP acquired the data. ALG, GMK, PR analyzed data. JAH, MC, TY, SG, MB,DMK, DMK, LC provided samples. DMK, DZ helped interpret the data andprovided critical feedback on manuscript. All authors helped draft the finalmanuscript and approved its submission.Ethics approval and consent to participateIsolates from New York were originally obtained as part of IRB approvedepidemiology and pathogenesis studies [4, 18, 36, 37]. For this study,de-identified isolates and accompanying clinical information were shippedto the University of Washington and the protocol was approved by theUniversity of Rochester Institutional Review Board with a waiver of consent.The Japanese specimens were collected during routine pediatric visits.Informed oral consent was obtained from parent or guardian of all childparticipants on their behalf and documented in the medical record. The useof oral consent and the samples was approved by the Institutional ReviewBoard of Fujita Health University (No. 14-096). Use of the saliva samples frominfants in a birth cohort study [39] collected in Kampala, Uganda and obtainedfrom Dr. Soren Gantt was approved by Institutional Review Boards at theUniversity of Washington, the Fred Hutchinson Cancer Research Center, theUniversity of British Columbia, Makerere University, and the Ugandan NationalCouncil for Science and Technology. The University of WashingtonInstitutional Review Board approved use of the iciHHV-6 specimens fromthe Fred Hutchinson Cancer Research Center and use of anonymized ex-cess HHV-6-positive samples submitted for testing at the University ofWashingtonVirology lab. All samples were anonymized prior to analysis.Consent for publicationNot applicable.Competing interestsMichael Boeckh declares competing interests from Chimerix (personal feesand research funding), Vir (personal fees) and Microbiotix (personal fees).Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1Department of Laboratory Medicine, University of Washington, Seattle, WA,USA. 2Fred Hutchinson Cancer Research Center, Seattle, WA, USA.3Department of Pharmaceutical Chemistry, University of California, SanFrancisco, CA, USA. 4Division of Infection and Immunity, University CollegeLondon, London, UK. 5Department of Pediatrics, University of Washington,Seattle, WA, USA. 6University of British Columbia, BC Children’s HospitalResearch Institute, Vancouver, Canada. 7Department of Pediatrics, FujitaHealth University, Fujita, Toyoake, Japan. 8University of Rochester MedicalCenter School of Medicine, Rochester, New York, USA.Greninger et al. BMC Genomics  (2018) 19:204 Page 15 of 17Received: 27 November 2017 Accepted: 13 March 2018References1. Ablashi D, Agut H, Alvarez-Lafuente R, Clark DA, Dewhurst S, DiLuca D,Flamand L, Frenkel N, Gallo R, Gompels UA, Höllsberg P, Jacobson S, LuppiM, Lusso P, Malnati M, Medveczky P, Mori Y, Pellett PE, Pritchett JC,Yamanishi K, Yoshikawa T. Classification of HHV-6A and HHV-6B as distinctviruses. Arch Virol. 2014;159:863–70.2. Braun DK, Dominguez G, Pellett PE. Human herpesvirus 6. Clin MicrobiolRev. 1997;10:521–67.3. Millichap JG, Millichap JJ. Role of viral infections in the etiology of febrileseizures. Pediatr Neurol. 2006;35:165–72.4. Hall CB, Long CE, Schnabel KC, Caserta MT, McIntyre KM, Costanzo MA,Knott A, Dewhurst S, Insel RA, Epstein LG. Human herpesvirus-6 infection inchildren. A prospective study of complications and reactivation. N Engl JMed. 1994;331:432–8.5. Zerr DM, Meier AS, Selke SS, Frenkel LM, Huang M-L, Wald A, Rhoads MP,Nguy L, Bornemann R, Morrow RA, Corey L. A population-based study ofprimary human herpesvirus 6 infection. N Engl J Med. 2005;352:768–76.6. Clark DA. Clinical and laboratory features of human herpesvirus 6chromosomal integration. Clin Microbiol infect off Publ Eur soc ClinMicrobiol. Infect Dis. 2016;22:333–9.7. Arbuckle JH, Medveczky MM, Luka J, Hadley SH, Luegmayr A, Ablashi D,Lund TC, Tolar J, De Meirleir K, Montoya JG, Komaroff AL, Ambros PF,Medveczky PG. The latent human herpesvirus-6A genome specificallyintegrates in telomeres of human chromosomes in vivo and in vitro. ProcNatl Acad Sci U S A. 2010;107:5563–8.8. Arbuckle JH, Pantry SN, Medveczky MM, Prichett J, Loomis KS, Ablashi D,Medveczky PG. Mapping the telomere integrated genome of humanherpesvirus 6A and 6B. Virology. 2013;442:3–11.9. Hill JA, Magaret AS, Hall-Sedlak R, Mikhaylova A, Huang M-L, Sandmaier BM,Hansen JA, Jerome KR, Zerr DM, Boeckh M. Outcomes of hematopoietic celltransplantation using donors or recipients with inherited chromosomallyintegrated HHV-6. Blood. 2017;130:1062–9.10. Sedlak RH, Hill JA, Nguyen T, Cho M, Levin G, Cook L, Huang M-L, FlamandL, Zerr DM, Boeckh M, Jerome KR. Detection of human Herpesvirus 6B(HHV-6B) reactivation in hematopoietic cell transplant recipients withinherited chromosomally integrated HHV-6A by droplet digital PCR. J ClinMicrobiol. 2016;54:1223–7.11. Wallaschek N, Gravel A, Flamand L, Kaufer BB. The putative U94 integrase isdispensable for human herpesvirus 6 (HHV-6) chromosomal integration. JGen Virol. 2016;97:1899–903.12. Seo S, Renaud C, Kuypers JM, Chiu CY, Huang M-L, Samayoa E, Xie H, Yu G,Fisher CE, Gooley TA, Miller S, Hackman RC, Myerson D, Sedlak RH, Kim Y-J,Fukuda T, Fredricks DN, Madtes DK, Jerome KR, Boeckh M. Idiopathicpneumonia syndrome after hematopoietic cell transplantation: evidence ofoccult infectious etiologies. Blood. 2015;125:3789–97.13. Yozwiak NL, Skewes-Cox P, Stenglein MD, Balmaseda A, Harris E, DeRisi JL.Virus identification in unknown tropical febrile illness cases using deepsequencing. PLoS Negl Trop Dis. 2012;6:e1485.14. Kawada J-I, Okuno Y, Torii Y, Okada R, Hayano S, Ando S, Kamiya Y, KojimaS, Ito Y. Identification of viruses in cases of pediatric acute encephalitis andencephalopathy using next-generation sequencing. Sci Rep. 2016;6:33452.15. Leber AL, Everhart K, Balada-Llasat J-M, Cullison J, Daly J, Holt S, Lephart P,Salimnia H, Schreckenberger PC, DesJarlais S, Reed SL, Chapin KC, LeBlanc L,Johnson JK, Soliven NL, Carroll KC, Miller J-A, Dien Bard J, Mestas J, BankowskiM, Enomoto T, Hemmert AC, Bourzac KM. Multicenter evaluation of BioFireFilmArray meningitis/encephalitis panel for detection of Bacteria, viruses, andyeast in cerebrospinal fluid specimens. J Clin Microbiol. 2016;54:2251–61.16. Salimnia H, Fairfax MR, Lephart PR, Schreckenberger P, DesJarlais SM,Johnson JK, Robinson G, Carroll KC, Greer A, Morgan M, Chan R, LoeffelholzM, Valencia-Shelton F, Jenkins S, Schuetz AN, Daly JA, Barney T, Hemmert A,Kanack KJ. Evaluation of the FilmArray blood culture identification panel:results of a multicenter controlled trial. J Clin Microbiol. 2016;54:687–98.17. Green DA, Hitoaliaj L, Kotansky B, Campbell SM, Peaper DR. Clinical utility ofon-demand multiplex respiratory pathogen testing among adultoutpatients. J Clin Microbiol. 2016;54:2950–5.18. Caserta MT, Hall CB, Schnabel K, McIntyre K, Long C, Costanzo M, DewhurstS, Insel R, Epstein LG. Neuroinvasion and persistence of human herpesvirus6 in children. J Infect Dis. 1994;170:1586–9.19. Gompels UA, Nicholas J, Lawrence G, Jones M, Thomson BJ, Martin MED,Efstathiou S, Craxton M, Macaulay HA. The DNA sequence of humanHerpesvirus-6: structure, coding content, and genome evolution. Virology.1995;209:29–51.20. Isegawa Y, Mukai T, Nakano K, Kagawa M, Chen J, Mori Y, Sunagawa T,Kawanishi K, Sashihara J, Hata A, Zou P, Kosuge H, Yamanishi K. Comparisonof the complete DNA sequences of human Herpesvirus 6 variants a and B. JVirol. 1999;73:8053–63.21. Dominguez G, Dambaugh TR, Stamey FR, Dewhurst S, Inoue N, Pellett PE.Human Herpesvirus 6B genome sequence: coding content and comparisonwith human Herpesvirus 6A. J Virol. 1999;73:8040–52.22. Nakatsu F, Baskin JM, Chung J, Tanner LB, Shui G, Lee SY, Pirruccello M, HaoM, Ingolia NT, Wenk MR, De Camilli P. PtdIns4P synthesis by PI4KIIIα at theplasma membrane and its impact on plasma membrane identity. J Cell Biol.2012;199:1003–16.23. Arias C, Weisburd B, Stern-Ginossar N, Mercier A, Madrid AS, Bellare P,Holdorf M, Weissman JS, Ganem D. KSHV 2.0: a comprehensive annotationof the Kaposi’s sarcoma-associated herpesvirus genome using next-generation sequencing reveals novel genomic and functional features. PLoSPathog. 2014;10:e1003847.24. Stern-Ginossar N, Weisburd B, Michalski A, Le VTK, Hein MY, Huang S-X, MaM, Shen B, Qian S-B, Hengel H, Mann M, Ingolia NT, Weissman JS. Decodinghuman cytomegalovirus. Science. 2012;338:1088–93.25. Szpara ML, Gatherer D, Ochoa A, Greenbaum B, Dolan A, Bowden RJ,Enquist LW, Legendre M, Davison AJ. Evolution and diversity in humanherpes simplex virus genomes. J Virol. 2014;88:1209–27.26. Newman RM, Lamers SL, Weiner B, Ray SC, Colgrove RC, Diaz F, Jing L,Wang K, Saif S, Young S, Henn M, Laeyendecker O, Tobian AAR, Cohen JI,Koelle DM, Quinn TC, Knipe DM. Genome sequencing and analysis ofgeographically diverse clinical isolates of herpes simplex virus 2. J Virol.2015;89:8219–32.27. Renzette N, Bhattacharjee B, Jensen JD, Gibson L, Kowalik TF. Extensivegenome-wide variability of human cytomegalovirus in congenitally infectedinfants. PLoS Pathog. 2011;7:e1001344.28. Achaz G. Frequency spectrum neutrality tests: one for all and all for one.Genetics. 2009;183:249–58.29. Tajima F. Statistical method for testing the neutral mutation hypothesis byDNA polymorphism. Genetics. 1989;123:585–95.30. Greninger A, Roychoudhury P, Makhsous N, Hanson D, Chase J, Krueger G,Xie H, Huang M-L, Saunders L, Ablashi D, Koelle DM, Cook L, Jerome KR.Copy number heterogeneity, large origin tandem repeats, and interspeciesrecombination in HHV-6A and HHV-6B reference strains. bioRxiv. 2017:193805. https://www.ncbi.nlm.nih.gov/pubmed/29491155.31. Zhang E, Bell AJ, Wilkie GS, Suárez NM, Batini C, Veal CD, Armendáriz-CastilloI, Neumann R, Cotton VE, Huang Y, Porteous DJ, Jarrett RF, Davison AJ,Royle NJ. Inherited chromosomally integrated human herpesvirus 6genomes are ancient, intact and potentially able to reactivate fromtelomeres. J Virol JVI. 2017;91:01137–17.32. Tesini BL, Epstein LG, Caserta MT. Clinical impact of primary infection withRoseoloviruses. Curr Opin Virol. 2014;9:91–6.33. Koelle DM, Norberg P, Fitzgibbon MP, Russell RM, Greninger AL, Huang M-L,Stensland L, Jing L, Magaret AS, Diem K, Selke S, Xie H, Celum C, LingappaJR, Jerome KR, Wald A, Johnston C. Worldwide circulation of HSV-2 × HSV-1recombinant strains. Sci Rep. 2017;7:44084.34. Morris JH, Knudsen GM, Verschueren E, Johnson JR, Cimermancic P,Greninger AL, Pico AR. Affinity purification-mass spectrometry andnetwork analysis to understand protein-protein interactions. Nat Protoc.2014;9:2539–54.35. Greninger AL, Messacar K, Dunnebacke T, Naccache SN, Federman S,Bouquet J, Mirsky D, Nomura Y, Yagi S, Glaser C, Vollmer M, Press CA,Kleinschmidt-DeMasters BK, Klenschmidt-DeMasters BK, Dominguez SR, ChiuCY. Clinical metagenomic identification of Balamuthia mandrillarisencephalitis and assembly of the draft genome: the continuing case forreference genome sequencing. Genome Med. 2015;7:113.36. Hall CB, Caserta MT, Schnabel KC, Long C, Epstein LG, Insel RA, Dewhurst S.Persistence of human herpesvirus 6 according to site and variant: possiblegreater neurotropism of variant a. Clin Infect Dis Off Publ Infect Dis Soc Am.1998;26:132–7.37. Norton RA, Caserta MT, Hall CB, Schnabel K, Hocknell P, Dewhurst S.Detection of human herpesvirus 6 by reverse transcription-PCR. J ClinMicrobiol. 1999;37:3672–5.Greninger et al. BMC Genomics  (2018) 19:204 Page 16 of 1738. Pruksananonda P, Hall CB, Insel RA, McIntyre K, Pellett PE, Long CE, SchnabelKC, Pincus PH, Stamey FR, Dambaugh TR. Primary human herpesvirus 6infection in young children. N Engl J Med. 1992;326:1445–50.39. Gantt S, Orem J, Krantz EM, Morrow RA, Selke S, Huang M-L, Schiffer JT, JeromeKR, Nakaganda A, Wald A, Casper C, Corey L. Prospective characterization ofthe risk factors for transmission and symptoms of primary human Herpesvirusinfections among Ugandan infants. J Infect Dis. 2016;214:36–44.40. Hill JA, HallSedlak R, Magaret A, Huang M-L, Zerr DM, Jerome KR, Boeckh M.Efficient identification of inherited chromosomally integrated humanherpesvirus 6 using specimen pooling. J Clin Virol Off Publ Pan Am Soc ClinVirol. 2016;77:71–6.41. Zerr DM, Gupta D, Huang M-L, Carter R, Corey L. Effect of antivirals onhuman herpesvirus 6 replication in hematopoietic stem cell transplantrecipients. Clin Infect Dis Off Publ Infect Dis Soc Am. 2002;34:309–17.42. Sedlak RH, Cook L, Huang M-L, Magaret A, Zerr DM, Boeckh M, Jerome KR.Identification of chromosomally integrated human herpesvirus 6 by dropletdigital PCR. Clin Chem. 2014;60:765–72.43. Stanton R, Wilkinson GWG, Fox JD. Analysis of human herpesvirus-6 IE1sequence variation in clinical samples. J Med Virol. 2003;71:578–84.44. Salipante SJ, SenGupta DJ, Cummings LA, Land TA, Hoogestraat DR,Cookson BT. Application of whole-genome sequencing for bacterial straintyping in molecular epidemiology. J Clin Microbiol. 2015;53:1072–9.45. Johnston C, Magaret A, Roychoudhury P, Greninger AL, Reeves D, Schiffer J,Jerome KR, Sather C, Diem K, Lingappa JR, Celum C, Koelle DM, Wald A.Dual-strain genital herpes simplex virus type 2 (HSV-2) infection in the US,Peru, and 8 countries in sub-Saharan Africa: a nested cross-sectional viralgenotyping study. PLoS Med. 2017;14:e1002475.46. Depledge DP, Palser AL, Watson SJ, Lai IY-C, Gray ER, Grant P, Kanda RK,Leproust E, Kellam P, Breuer J. Specific capture and whole-genomesequencing of viruses from clinical samples. PLoS One. 2011;6:e27805.47. Tweedy J, Spyrou MA, Donaldson CD, Depledge D, Breuer J, Gompels UA.Complete genome sequence of the human Herpesvirus 6A strain AJ fromAfrica resembles strain GS from North America. Genome Announc. 2015;3.https://doi.org/10.1128/genomeA.01498-14.48. Greninger AL, Knudsen GM, Betegon M, Burlingame AL, Derisi JL. The 3Aprotein from multiple picornaviruses utilizes the golgi adaptor proteinACBD3 to recruit PI4KIIIβ. J Virol. 2012;86:3605–16.49. Greninger AL, Knudsen GM, Betegon M, Burlingame AL, DeRisi JL. ACBD3interaction with TBC1 domain 22 protein is differentially affected byenteroviral and kobuviral 3A protein binding. MBio. 2013;4:e00098–13.50. Hellman U, Wernstedt C, Góñez J, Heldin CH. Improvement of an “in-gel”digestion procedure for the micropreparation of internal protein fragmentsfor amino acid sequencing. Anal Biochem. 1995;224:451–5.51. Olsen JV, de LMF G, Li G, Macek B, Mortensen P, Pesch R, Makarov A, LangeO, Horning S, Mann M. Parts per million mass accuracy on an Orbitrap massspectrometer via lock mass injection into a C-trap. Mol Cell ProteomicsMCP. 2005;4:2010–21.52. Chalkley RJ, Baker PR, Medzihradszky KF, Lynn AJ, Burlingame AL. In-depthanalysis of tandem mass spectrometry data from disparate instrumenttypes. Mol Cell Proteomics MCP. 2008;7:2386–98.53. Elias JE, Gygi SP. Target-decoy search strategy for increased confidence inlarge-scale protein identifications by mass spectrometry. Nat Methods. 2007;4:207–14.54. Martin M. Cutadapt removes adapter sequences from high-throughputsequencing reads. EMBnet. Journal. 2011;17:10–2.55. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS, LesinVM, Nikolenko SI, Pham S, Prjibelski AD, Pyshkin AV, Sirotkin AV, Vyahhi N,Tesler G, Alekseyev MA, Pevzner PA. SPAdes: a new genome assemblyalgorithm and its applications to single-cell sequencing. J Comput Biol.2012;19:455–77.56. Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illuminasequence data. Bioinforma Oxf Engl. 2014;30:2114–20.57. Seemann T. Prokka: rapid prokaryotic genome annotation. Bioinforma OxfEngl. 2014;30:2068–9.58. Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Dudoit S, Ellis B,Gautier L, Ge Y, Gentry J, Hornik K, Hothorn T, Huber W, Iacus S, Irizarry R,Leisch F, Li C, Maechler M, Rossini AJ, Sawitzki G, Smith C, Smyth G, TierneyL, Yang JYH, Zhang J. Bioconductor: open software development forcomputational biology and bioinformatics. Genome Biol. 2004;5:R80.59. Johnston C, Magaret A, Roychoudhury P, Greninger AL, Cheng A, Diem K,Fitzgibbon MP, Huang M-L, Selke S, Lingappa JR, Celum C, Jerome KR, WaldA, Koelle DM. Highly conserved intragenic HSV-2 sequences: results fromnext-generation sequencing of HSV-2 UL and US regions from genitalswabs collected from 3 continents. Virology. 2017;510:90–8.60. Hudson RR, Kaplan NL. Statistical properties of the number ofrecombination events in the history of a sample of DNA sequences.Genetics. 1985;111:147–64.61. Minin VN, Dorman KS, Fang F, Suchard MA. Dual multiple change-pointmodel leads to more accurate recombination detection. Bioinforma OxfEngl. 2005;21:3034–42.62. Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, Buxton S,Cooper A, Markowitz S, Duran C, Thierer T, Ashton B, Meintjes P,Drummond A. Geneious basic: an integrated and extendable desktopsoftware platform for the organization and analysis of sequence data.Bioinforma Oxf Engl. 2012;28:1647–9.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:Greninger et al. BMC Genomics  (2018) 19:204 Page 17 of 17


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items