UBC Faculty Research and Publications

Widespread, focal copy number variations (CNV) and whole chromosome aneuploidies in Trypanosoma cruzi… Minning, Todd A; Weatherly, D B; Flibotte, Stephane; Tarleton, Rick L Mar 7, 2011

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

Item Metadata


52383-12864_2010_Article_3274.pdf [ 1.45MB ]
JSON: 52383-1.0220563.json
JSON-LD: 52383-1.0220563-ld.json
RDF/XML (Pretty): 52383-1.0220563-rdf.xml
RDF/JSON: 52383-1.0220563-rdf.json
Turtle: 52383-1.0220563-turtle.txt
N-Triples: 52383-1.0220563-rdf-ntriples.txt
Original Record: 52383-1.0220563-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessWidespread, focal copy number variations (CNV)and whole chromosome aneuploidies inTrypanosoma cruzi strains revealed by arraycomparative genomic hybridizationTodd A Minning1, D Brent Weatherly1, Stephane Flibotte2, Rick L Tarleton1*AbstractBackground: Trypanosoma cruzi is a protozoan parasite and the etiologic agent of Chagas disease, an importantpublic health problem in Latin America. T. cruzi is diploid, almost exclusively asexual, and displays an extraordinarilydiverse population structure both genetically and phenotypically. Yet, to date the genotypic diversity of T. cruzi andits relationship, if any, to biological diversity have not been studied at the whole genome level.Results: In this study, we used whole genome oligonucleotide tiling arrays to compare gene content inbiologically disparate T. cruzi strains by comparative genomic hybridization (CGH). We observed that T. cruzi strainsdisplay widespread and focal copy number variations (CNV) and a substantially greater level of diversity than canbe adequately defined by the current genetic typing methods. As expected, CNV were particularly frequent ingene family-rich regions containing mucins and trans-sialidases but were also evident in core genes. Gene groupsthat showed little variation in copy numbers among the strains tested included those encoding protein kinasesand ribosomal proteins, suggesting these loci were less permissive to CNV. Moreover, frequent variation inchromosome copy numbers were observed, and chromosome-specific CNV signatures were shared by geneticallydivergent T. cruzi strains.Conclusions: The large number of CNV, over 4,000, reported here uphold at a whole genome level the long heldparadigm of extraordinary genome plasticity among T. cruzi strains. Moreover, the fact that these heritable markersdo not parse T. cruzi strains along the same lines as traditional typing methods is strongly suggestive of geneticexchange playing a major role in T. cruzi population structure and biology.BackgroundHuman infection with T. cruzi, a vector-borne proto-zoan parasite, is the cause of Chagas disease, which is apotentially fatal malady endemic to much of LatinAmerica. T. cruzi is a member of the order Trypanoso-matida, and, like other members of this early divergedgroup of eukaryotes, is diploid and has a primarily clo-nal population structure which is extremely diverse [1].The broad host range of T. cruzi, which includes over100 species of both wild and domestic mammals(reviewed in [2]) may also contribute to its remarkablephenotypic and genetic diversity.The population structure of T. cruzi has been exam-ined by sub-genomic methods, such as multi-locussequence typing (MLST) and microsatellite analyses,resulting in the classification of T. cruzi strains into sixdiscrete typing units (DTU), type I and types IIa-e(reviewed in [3]). A change in nomenclature from theseDTU names to TCI-TCVI has recently been suggested[4]. However, in the interest of clarity we will heretoforeuse the more familiar nomenclature of type I and typesIIa-IIe. Although an asexual replication mechanism bestexplains the population structure of T. cruzi as it is cur-rently understood, molecular analyses support theoccurrence of at least two hybridization events in the* Correspondence: tarleton@cb.uga.edu1Center for Tropical and Emerging Global Diseases, University of Georgia,Athens, Georgia, 30602, USAFull list of author information is available at the end of the articleMinning et al. BMC Genomics 2011, 12:139http://www.biomedcentral.com/1471-2164/12/139© 2011 Minning et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative CommonsAttribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction inany medium, provided the original work is properly cited.past resulting in mosaic genomes in two of the DTU(types IIa and IIc) and hybrid genomes in another two(types IId and IIe), one of which (CL-Brener-type IIe)was used as the type strain for the T. cruzi genomesequencing effort [5].The genotypic diversity evidenced by sub-genomicmethods of analysis suggests that whole genome ana-lyses of T. cruzi strain diversity would be fruitful, espe-cially for elucidating the underpinnings of straindiversity in biological characteristics, such as variabilityin complements of large gene family members. Nearlyone half of the T. cruzi genome contains repeatsequences largely comprised of thousands of membersof large gene families, including trans-sialidases, mucinassociated surface proteins (MASP), mucins, retrotran-sposon hotspot (RHS) proteins, dispersed gene family1 proteins (DGF), and surface protease gp63 [6]. Manyof these have been shown to be important targets ofimmune responses in infected hosts [7,8]. Isolates of T.cruzi (generally referred to as “strains”) appear to showa near limitless range of variation in important biologi-cal characteristics, among these, the numbers of para-sites in the blood and tissues of various hosts, thefocus and location of inflammation and thus the mor-bidity and mortality in these hosts, and susceptibilityof these isolates to anti-T. cruzi drugs. The currentgenetic classification cannot fully account for this var-iation (e.g. all type I strains are not equally virulent)despite the fact that a substantial proportion of thisvariation is almost certainly based upon genetic differ-ences among isolates.To date only the reference strain, CL-Brener, hasbeen fully sequenced [6]. Moreover, the genomesequence of the CL-Brener strain has only recentlybeen assembled into chromosome-sized pieces that willfacilitate genome-wide strain comparisons [9]. Tofurther explore the degree of genetic variabilitybetween T. cruzi isolates, and to examine the relation-ship between genotypic and biological diversity on agenomic scale, we used whole genome oligonucleotidetiling arrays to determine copy number variations(CNV) in 16 T. cruzi strains by competitive hybridiza-tions using the CL-Brener strain as reference. Ourresults partially support the proposed type I - type IIdichotomy in the T. cruzi population structure but alsoreveal similarities and differences between T. cruzistrains that cannot be explained by the current DTUscheme. These findings suggest that either coevolutionof distinct, chromosome-specific CNVs frequentlyoccurs in different T. cruzi strains or that chromosomeexchange between T. cruzi strains is much more com-mon than currently thought. The results also suggestthat T. cruzi is remarkably permissive to substantialCNV, including whole chromosome CNV.ResultsWhole genome oligonucleotide tiling arrays (~290,000spots) were designed as described in the Methods sec-tion, using the only fully sequenced T. cruzi strain, thehybrid CL Brener, as a template. Because of the hybridnature of the CL Brener strain, with its “Esmeraldo-like”(Esm) and “non-Esmeraldo-like” (non-Esm) alleles, itwas possible to design allele-specific probes for manyregions of the genome. Regions of the genome that arerich in genes that are members of large gene familieshave relatively few probes because the sequence similari-ties among family members did not permit the design ofgene-specific probes.As a validation of the accuracy of the arrays we per-formed hybridizations comparing wt T. cruzi strains withgene knockout parasite lines generated in our laboratory[10]. Additional file 1 shows the CGH data for one ofthese hybridizations comparing a knockout strain forenoyl-CoA hydratase/isomerase family protein (tandemgenes each singly replaced - [Tc00.1047053511529.160and Tc00.1047053511529.150]) vs wt T. cruzi. The CNVresulting from single knockout of one copy each of thesetwo tandemly arrayed genes is obvious at the whole chro-mosome view of chromosome 35, panel A. Moreover, theclose-up view in panel B reveals overlap of the Esm andnon-Esm probes (green and blue spots, respectively)within the coding portions of the genes, but divergenceof Esm and non-Esm ratios for the intergenic region.This reflects the fact that the Esm and non-Esm allelesfor these two genes are nearly identical, but the Esm andnon-Esm sequences for the intergenic region are not.Genomic DNA samples from 16 T. cruzi strains werecompared to genomic DNA from the CL-Brener strainby competitive hybridizations on this array and thecomplete set of figures for all 41 chromosomes foreach of the 16 strain comparisons are shown in Addi-tional file 2 and the full dataset can be viewed in Addi-tional files 3 and 4. The data have also been depositedin the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under the accession GSE23576. Themost striking feature of these results is the very largenumber of both segmental (ranging in size from 500bp to over 500 kbp) and whole chromosome aneuploi-dies, representative examples of which are shown inthe hybridization results for chromosome 39 for 12strains (Figure 1). Probes with log2ratios approximat-ing zero are present in equal copy numbers in boththe test and reference (CL Brener) strains and probeswith positive log2ratios have higher copy number inthe test strains versus the reference strain, and thusrepresent amplifications in the test strain. Probes withnegative log2ratios indicate deletions in the test strainand/or sufficient sequence divergence to result indecreased hybridization in the test channel. SegmentalMinning et al. BMC Genomics 2011, 12:139http://www.biomedcentral.com/1471-2164/12/139Page 2 of 11aneuploidies (e.g. a 500 kbp segment in the Brazilstrain [arrow]) and shared, focal deletions in multiplestrains are readily evident (Figure 1 and Additional file2). Sequence divergence resulting in decreased hybridi-zation intensity can be seen in the Esmeraldo panelwhere the green dots (representing Esmeraldo allele-specific probes) have positive ratios and blue dots(non-Esmeraldo probes) have negative log2ratios(Figure 1). However, note that the average log2ratiofor all probes is near zero in the Esmeraldo panel, sug-gesting that Esmeraldo is homozygous for the Esmer-aldo haplotype as would be expected and that the totalnumber copies of chromosome 39 in this strain is 2(assuming, of course, that CL-Brener has 2 copies ofchromosome 39). The presence of apparent deletionsin the test strains that are shared between manygenetically divergent test strains (as seen in the redboxed regions) suggests that such instances are bonafide deletions rather than sequence divergence.To identify the regions of the T. cruzi genome that aremost prone to CNV, we annotated segments of the gen-ome that had CNVs relative to the overall average of allstrains (Figure 2). ‘Hotspots’ of CNV were readily evi-dent and were present on every chromosome, althoughmuch more prevalent on some (e.g. 18, 38 and 41) thanothers (e.g. 11, 34, 36, 37, 39). The CNV were also focaland widespread, frequently associated with gene familyrich regions of the genome (note that chromosomes 18,38, and 41 are the most gene family rich chromosomes),but also in core regions. Due to the scarcity of probeswithin the gene family rich regions a statistical test ofthe significance of the bias of hotspot regions beinglocated in gene family rich regions of the genome wasnot possible.The classes of genes mapping to these hotspot regionsare ranked in order of frequency of CNV in Table 1(for all gene groups comprised of 5 or more genes, eachwith a minimum of 5 probes per 500bp). Perhaps notFigure 1 Representative plots from 12 hybridizations for T. cruzi chromosome 39. Each dot represents an oligonucleotide probe. The CL-Brener strain, which was used as the reference strain for genome sequencing, is hybrid, thus probes were designed to non-Esmeraldo (non-Esm)sequences (blue spots), Esmeraldo-like (Esm) sequences (green spots), non-Esm gene family sequences (black spots), and Esm gene familysequences (gray spots). In each panel positive log2 ratios of signal intensities (test strain/reference strain) represent amplification in the test strainand negative log2 ratios represent deletion in the test strain, relative to CL-Brener, which was the reference strain in all hybridizations. Shown areexamples of segmental aneuploidies of different sizes boxed in gray (>500 kbp amplification in Brazil strain relative to CL-Brener -arrow) and red(~40 kbp deletion in several strains). The Esmeraldo hybridization (boxed in orange) shows decreased log2ratios for the non-Esm probes (blue)and increased log2ratios for the Esm probes (green), as expected because Esmeraldo is homozygous for the Esm haplotype.Minning et al. BMC Genomics 2011, 12:139http://www.biomedcentral.com/1471-2164/12/139Page 3 of 11Figure 2 Graphical view of proportion of strains displaying CNV for each sequence feature on each chromosome. A sequence feature isdefined as either an annotated gene or the un-translated region (UTR) between two genes. CNV criteria were as follows: minimum log2 ratio ofsignal intensities (test strain/reference) +/- 0.6, minimum number of probes 5, tolerance for extending a CNV 0.1. At the top of each panel is aline drawing representing a chromosome. Vertical bars represent CNV that met the criteria for significance. Vertical bars above the line areamplifications and those below the line are deletions relative to CL-Brener. The height (depth) of the vertical bars is proportional to the numberof strains showing that CNV. The vertical bars are colored to indicate strain type (type 1 and type 2) as follows: green, type 1 strainsamplification; blue, type 1 strains deletion; maroon, type 2 strains amplification; yellow, type 2 strains deletion. Below each chromosome linedrawing is a diagram representing the annotated chromosome. Genes are color coded based on annotation; named genes not belonging tolarge gene families (dark blue), hypothetical genes (light blue), trans-sialidases (red), mucin associated surface proteins (MASP) (burnt orange),mucins (orange), retrotransposon hotspot (RHS) proteins (green), dispersed gene family 1 (DGF) (gray), surface protease gp63 (purple).Minning et al. BMC Genomics 2011, 12:139http://www.biomedcentral.com/1471-2164/12/139Page 4 of 11surprisingly, genes that are members of large families ofsurface proteins, such as the mucins, trans-sialidases,mucin-associated surface proteins (MASPs), and surfaceproteases are among those with the highest number ofCNV among the strains tested, with type II mucins show-ing CNV over 70% of the time. Interestingly, proteinkinases, a large and diverse family of related genes, as wellas hypothetical proteins were the gene groups with theleast association with hotspot regions, suggesting that allor nearly all of them may be under strict copy numbercontrol. Other genes that are relatively stable in terms ofcopy number included genes for ribosomal proteins andDNA repair, consistent with their expected essential roles.CNVs in gene groups represented by less than 5 genes areshown in Additional file 5.The typing of T. cruzi into 6 discrete typing units(DTUs) is based upon a relatively small number of loci,including small subunit rRNA, elongation factor 1a,actin, dihydrofolate reductase-thymidylate sythase, andtrypanothione reductase among others [11,12]. To deter-mine if this DTU classification system matched to pat-terns of CNV among typed strains, we ‘typed’chromosomes based on their CNV patterns, creatingCNV signature types (Additional file 4) and thengrouped these CNV-typed T. cruzi strains by similarityin the number of shared chromosome types (Figure 3).The typing results suggest that there is substantiallymore genetic diversity among the type I strains thanthere is among the type II strains. This is consistentwith recent microsatellite analyses of type I strain diver-sity [13]. However, the apparently greater diversityamong type I versus type II strains observed in ourstudy could be a result of the feature selection. Featuresfor typing were selected by an unbiased computationalTable 1 Distribution of genes associated with hotspot regions, minimum of 5 candidates on the arraysGene Name # Genome Occurrences1 #Array Occurrences2 # Array Candidates3 # Sig CNV4 % Sig of Candidatesmucin TcMUCII 728 506 164 121 73.7serine carboxypeptidase 7 7 6 4 66.6tryptophanyl-tRNA synthetase 23 8 6 4 66.6mucin-associated surface 1377 948 437 243 55.6retrotransposon hot 753 419 82 41 50ATP-dependent chaperone 5 5 5 2 40beta galactofuranosyl 66 47 10 4 40mitogen-activated protein kinase 11 11 10 4 40Sialidase 3209 789 135 51 37.7clathrin coat 7 7 7 2 28.5surface protease 425 258 88 25 28.4myosin heavy 19 19 17 3 17.6PIF1 helicase-like 12 12 12 2 16.6mismatch repair 13 13 13 2 15.3ATP-dependent DEAD/H 106 88 66 10 15.1leucine-rich repeat 17 16 16 2 12.5DNA topoisomerase 9 9 9 1 11.1protein tyrosine 10 10 9 1 11.1cytochrome c 20 18 13 1 7.6protein transport 16 16 14 1 7.1elongation factor 203 65 15 1 6.6ABC transporter 33 33 33 2 6ubiquitin-conjugating enzyme 24 24 23 1 4.3serine/threonine protein 80 78 78 3 3.860S ribosomal 106 106 93 3 3.2DNA repair 37 36 35 1 2.8RNA-binding protein 80 72 71 2 2.8hypothetical protein 11812 10489 9735 122 1.2protein kinase 311 277 270 1 0.31. # genome occurrences = number of instances for the annotation in the genome.2. # array occurrences = number of annotated genes with probes on the arrays.3. # array candidates = number of annotated genes represented by probes on the arrays and which had probe density of 5 unique probes per 500 bp. Genescould have insufficient probe density due to repeat regions or they could have too few probes due to length of the gene (minimum of 5 probes).4. # sig CNV = CNV with a minimum log2 ratio difference of +/- 0.5, for a minimum of 5 probes over a segment size of 500 bp in at least 5 test strains.Minning et al. BMC Genomics 2011, 12:139http://www.biomedcentral.com/1471-2164/12/139Page 5 of 11method, but not all of the nearly limitless possible com-binations of available CNV features were used in thetyping. Although type I and type II strains groupedtogether in the classification, on an individual chromo-some basis some type I strains clearly had chromosomalCNV patterns matching those in type II strains and viceversa. For example, on chromosome 11 Esmeraldo,M5631, and Tu18 (types IIb, IIc, and IIb respectively)each have a CNV that is shared among all of the type Istrains (Additional file 4, slide 11). Yet the other type IIstrains, including Y strain which is also type IIb, do nothave this CNV. This CNV, a deletion, corresponds tothe locus for proline racemase that may be a target ofimmune selection [14]. Taken together, these data sug-gest that either there is extensive homoplasy among T.cruzi strains (i.e. the same amplification/deletion eventsoccurring multiple times in strains that are evolvingindependently) or that chromosome exchange hasoccurred more frequently in T. cruzi than the currenttwo-hybridization event theory would suggest [5].DiscussionIn this study, we used whole-genome tiling arrays tocatalog the diversity in gene copy number among 17 T.cruzi strains with widely varying biological characteris-tics. The overwhelming conclusion from this work isthat there is substantial and widespread variation ingene copy numbers among T. cruzi isolates. The obser-vation of CNVs in T. cruzi is not surprising. The relativeinfrequency of sexual recombination, coupled with thenear absence of transcriptional promoters in T. cruzimeans that both the generation of genetic diversity andthe regulation of the level of proteins is determined inpart by within-isolate gene recombination and amplifica-tion. Nevertheless, the extent of variation in a relativelysmall set of isolates is quite dramatic. This variation isevident from the level of entire chromosomes down toindividual genes, and CNVs tend to cluster in regionsthat are heavily populated with repeat sequences, includ-ing those encoding the members of gene families of thesurface proteins, trans-sialidases, mucins, mucin asso-ciated surface proteins (MASP’s), as well as retrotran-sposon hot spot (RHS) proteins, the dispersed genefamily 1 (DGF-1), and protease gp63. Multicopy genesare frequently sources of mitotic recombination result-ing in amplification and deletion of duplicate genes [15]so their relationship to regions of CNV is also not unex-pected. The surface protein families in particular areknown to be the major targets of protective immuneresponses and this selective pressure presumably drivestheir expansion and variation among strains.In addition to the expected association of CNV withgenes that are under immunological pressure, the pre-sence and absence of CNV among other gene groupsmay be informative as to their functions and uniqueactivities in T. cruzi. A relatively high fraction of serinecarboxypeptidases, tryptophanyl aminoacyl tRNAsynthetases, beta galactofuranosyl transferases and mito-gen-activated protein kinases show CNVs among the T.cruzi isolates studied. The tryptophanyl aminoacyl tRNAsynthetases are expanded to 10 distinct genes (notincluding pseudogenes) in T. cruzi whereas T. bruceiand Leishmania each have only two, a cytoplasmic anda mitochondrial version. Interestingly, aminoacyl-tRNAFigure 3 Chromosomes were typed based on CNV signatures. Diagnostic CNV were selected based on the criteria that they were at least500 bp in length, represented by at least 5 oligonucleotide probes, and had log2ratios of at least +/- 1 (tolerance of 0.2) in at least 5 strains. Upto two CNV meeting the criteria were selected for each chromosome based on their ability to differentiate between the strains; i.e. the CNV thatseparated the strains into the highest number of groups compared to other candidate diagnostic CNV were selected. Thus, for eachchromosome there were a finite number of signature types to which strains were assigned based upon the presence/absence of the diagnosticCNV’s the direction of change (log2ratio +/- 0.8). Each chromosome type is represented by a different color, with CL-Brener always assigned totype ‘A’ (blue). The strains were then sorted based on the overall similarity of their chromosome types, taken together, relative to CL-Brener.Minning et al. BMC Genomics 2011, 12:139http://www.biomedcentral.com/1471-2164/12/139Page 6 of 11synthetase genes in other eukaryotes have been reportedto be increased in number and to acquire a diverse setnon-enzymatic functions, including as inflammatory andangiogenic cytokines [16]. The location of this expandedset of the tryptophanyl tRNA sythetases genes in hot-spots for recombination in T. cruzi also suggests thatthey may be modified and selected for as yet undeter-mined secondary functions. Similarly, T. cruzi has 37beta-galactofuranosyl transferase genes.(compared to 2for L. braziliensis, 1 each for L. major and L. infantum,and none for T. brucei) located in regions of highbetween-strain CNV. These enzymes are involved in thesynthesis of glycans on mucins in T. cruzi. Mucin glycanstructure is complex and heterogeneous between T.cruzi strains, again consistent with the expansion andvariation in transferases among T. cruzi strains notedherein [17]. Conversely, only 1 of 270 protein kinasesthat passed the data filters for this analysis and <2% ofthe nearly 10,000 hypothetical proteins were associatedwith hotspot regions, suggesting that variation, amplifi-cation and recombination of these genes is not toleratedor provides no selective advantage.Although this analysis suggests extensive occurrenceof whole chromosome aneuploidy in T. cruzi, the rangeof chromosome numbers as well as the absolute numberof individual chromosomes in each strain cannot be esti-mated based upon our analysis. The decreased signalintensities evident across whole chromosomes in thisCGH analysis would appear to be deletions but couldalso be in part due to sequence divergence in the teststrains relative to the CL Brener reference sequences.Since individual isolates or strains of T. cruzi arethought to evolve without significant between-straingenetic exchange due to the relative absence of sexualrecombination [1], the accumulation of sequence var-iants would be expected. Normally, traditional cytologi-cal methods such as metaphase spreads would be usedto resolve karyotypic differences between isolates. How-ever, this approach is not possible in the case of T. cruzibecause T. cruzi replicates via endodyogeny, where thenuclear membrane does not break down and chromo-somes do not fully condense during replication [18].Nevertheless it seems highly unlikely that sequencedivergence alone could explain our observations of chro-mosomes with significantly decreased signal intensitiesin many of the test strains relative to the CL-Brenerreference. And certainly whole chromosome increases insignal relative to the reference strain cannot beaccounted for by sequence divergence, since the targetprobes are based upon the reference strain sequenceand diverged sequences are unlikely to out-compete thehomologous sequences. For example, Y, Colombiana,M5631, wtCL, and TeDa2 each appear to have an extrachromosome 3 (Additional file 4, slide 3). It is mostreasonable to interpret significant deviation from a log2ratio of zero for test strain vs. reference strain over thelength of a chromosome as being indicative of a real dif-ference in the number of copies of that chromosome inthe test and reference strains, supporting the conclusionthat T. cruzi karyotypes are highly variable betweenstrains and suggesting that there is perhaps no ‘euploid’state for T. cruzi. This karyotypic plasticity may beanother mechanism T. cruzi uses to generate diversity inspite of apparently being asexual or nearly so and mayto some extent parallel what has been observed for thepathogenic yeast, Candida glabrata. Array CGH andpulsed field gel electrophoresis analyses of this haploid,asexual organism revealed extensive variability in thekaryotypes of C. glabrata strains and suggest that thisvariability may be linked to drug resistance [19,20].The biological characteristics of T. cruzi strains havebeen used to classify isolates into biodemes [21,22] andas well the DTU classification has been associated tocertain transmission and virulence characteristics of iso-lates [23-26]. Nevertheless, the link between DTU typeand the biological characteristics of a strain are notstrong [27]. Likewise, we find that the DTU organizationof strains is a poor predictor of patterns of CNV in indi-vidual chromosomes. Thus, clear and specific patterns ofamplifications and deletions in chromosomes areobserved among strains, but these specific patterns arealmost never restricted to or predictive of DTU type.This result has two important implications. First, thesubstantial and even continuous variation in biologicalcharacteristics of T. cruzi isolates will be difficult toaccount for using a limited number of genetic markers.Patterns of CNV provides another tool, but this too isnot sufficient to predict behaviors that likely have acomplex genetic basis. Nevertheless, these typingapproaches do provide insights into the populationstructure of the species and the evolution of individual“strains” that are thought to be genetically isolated fromeach other. The results of typing of chromosomes byCNV patterns that we present herein provides newinsights but also new questions related to geneticexchange in T. cruzi. The observation of shared patternsof CNV that do not track with the DTU type of a strainand are not consistent between chromosomes within thesame strains can be explained in two ways: either suchsimilar patterns of CNV are arising independently inisolates, perhaps due to common selective forces, orchromosomes of T. cruzi are being resorted orexchanged between isolates at a much higher frequencythan is currently appreciated. The similarity and com-plexity of CNV would seem to favor the latter of thesepossibilities but a mechanism by which such exchangewould occur is not clear. The full sequencing of addi-tional T. cruzi isolates, as is currently underway, shouldMinning et al. BMC Genomics 2011, 12:139http://www.biomedcentral.com/1471-2164/12/139Page 7 of 11help discriminate between these possibilities. Moreover,CGH arrays designed from the sequences of multiple T.cruzi strains, analogous to multi-species taxonomicarrays for Saccharomyces cerevisiae may help resolve theextent and nature of genetic exchange between T. cruzistrains and reveal heretofore undiscovered instances ofintrogression [28].In addition to these insights into the biology and evo-lution of T. cruzi, the identification of “hot spots” forCNV within the T. cruzi genome should also provideguidance in the selection of candidates for vaccines andfor targets for drug development. Although selectingcandidates that are encoded outside these hot spotsdoes not guarantee the absence of variation or thedevelopment of variation between isolates, avoidinggenes that are in such hotspots would seem prudent.ConclusionsCNV among T. cruzi strains are substantial in number,widespread throughout the T. cruzi genome, range insize from a few hundred base pairs to whole chromo-somes (based upon the assembled CL-Brener genome),and are discordant with traditional DTU assignmentsfor the strains tested. Taken together, these results sug-gest that there is much more genotypic diversity amongT. cruzi strains than can be fathomed using traditionaltyping methods and that genetic exchange, possibly bythe proposed “fusion then loss” mechanism [29], occursbetween T. cruzi strains much more frequently thancurrent dogma supposes. The large number of CNVshared between divergent DTU supports the geneticexchange hypothesis over the hypothesis that the sharedCNV arose due to homoplasy.CNV in T. cruzi tended to be concentrated in genefamily rich regions of the genome, suggesting that thehighly repetitive nature of the sequences in theseregions is a strong driver of mitotic recombination in T.cruzi whether occurring within diploid strains or instrains resulting from hybridization events.Neither the traditional DTU assignments for the T.cruzi strains used in this study nor the strain groupingsbased upon CNV patterns are currently good predictorsof the biological characteristics of these strains. Thus,further CGH and sequencing studies of a larger panel ofwell-characterized T. cruzi strains is warranted to deter-mine appropriate genotypic markers of T. cruzi biologi-cal diversity. Moreover, it would seem prudent to usethe CNV data reported herein when selecting targets fordrug and vaccine studies in order to avoid genes in hot-spots of CNV occurrence.MethodsParasites and DNA isolationT. cruzi strains were selected to represent a biologicallydiverse sample set from multiple DTU (Table 2). Epi-mastigotes were cultured in Liver Infusion Tryptose(LIT) medium and harvested as previously described[30]. Genomic DNA was obtained as previouslydescribed [31]. Briefly, epimastigotes were washed threetimes with ice-cold PBS, resuspended at a final densityof 2E + 08 cells per ml in lysis buffer (150 mM NaCl,100 mM EDTA, 100 ug/ml proteinase K, 10 ug/mlRNase A, and 0.5% sodium sarcosinate, pH 8.0) andincubated at 50°C for 30 m. Lysates were then extractedtwice with phenol/chloroform and the DNA was ethanolprecipitated, then resuspended in 5 mM Tris.Cl, pH 7.5.Table 2 Trypanosoma cruzi strains used in this studystrain origin host DTU virulence drug resistance referenceBrazil Brazil Human I high low [37]Chinata Bolivia Triatoma infestans I high ND unpublishedColombiana Colombia Human I high moderate [21]Esmeraldo Brazil Human IIb moderate ND [38]M5631 Brazil Dasypus novemcinctus IIc ND ND [39]M78 Argentina Human I low ND [40]Montalvania Brazil Human I high moderate Andrade unpublishedPalDa1 (clone 9) Argentina Didelphis albiventris I low ND [41]Sylvio X10/4 Brazil Human I low ND [42]TCC unknown unknown I low ND [43]TEDa2 (clone 4) Argentina Didelphis albiventris I ND ND [41]TEP6 (clone 5) Argentina dog I ND ND [41]Tu18 (clone 1) Bolivia Triatoma infestans IIb ND ND [44]Tulahuen Chile Human IIe high low [45]wtCL Brazil Triatoma infestans IIe high low [46]Y Brazil Human IIb high low [47]CL-Brener Brazil Triatoma infestans IIe high ND [48]Minning et al. BMC Genomics 2011, 12:139http://www.biomedcentral.com/1471-2164/12/139Page 8 of 11DNA sample integrity was checked by agarose gel elec-trophoresis. Genomic DNA for CL-Brener strain andEsmeraldo strains were kindly provided by JM Kelly andB Zingales, respectively.Array DesignWhole genome tiling arrays for comparative genomeanalysis were designed as previously described [32] withthe notable exception that all the criteria related tosequence similarity with other regions of the genomehad to be relaxed in T. cruzi in order to ensure propercoverage. We used 50-mer oligonucleotide probes andthe microarray had a total capacity of approximately380,000 probes, approximately 25% of which were usedfor preliminary quality assessment in this first genera-tion design. The ~290,000 probes relevant for the cur-rent work were selected with the following procedure:1) only the contigs including annotated genes were tar-geted, 2) the 50-mers occurring more than 4 times inthe genome were eliminated, 3) the homopolymerslonger than 5 nucleotides were eliminated, 4) only the50-mer oligonucleotides with a melting temperature Tmwithin +- 5°C of the median melting temperature werekept (where Tm = 0.41GC + constant), 5) the oligonu-cleotides requiring more than 150 cycles [33] to synthe-size were eliminated, 6) the oligonucleotides with a self-folding energy smaller than -1 kcal/mol according to ahybrid-ss-min calculation [34] were eliminated, and 7)from the probes passing all the above filters ~290,000were selected with a strategy designed to maximize theuniformity of coverage of the entire genome. However,for some regions of the genome, gene family richregions in particular, the DNA sequences were highlyrepetitive resulting in low tiling densities for thoseregions. For the same reason probes were not designedto sequencing scaffolds that were not used in the gen-ome assembly. Construction of the arrays was per-formed by Roche NimbleGen Inc. (see reference [35] foradditional details).Hybridization, Scanning, and Image AnalysisGenomic DNA from T. cruzi strains was sonicated andlabeled per the Nimblegen protocol. Labeling was byrandom-primed synthesis using Cy3/5-labeled randomnonamers, and mixed probes were hybridized to thearrays per the manufacturer’s protocol. Microarrayswere scanned using a ScanArray 5000 (Perkin Elmer)and signal intensity files were generated using Nimble-Scan software. Ratios of fluorescence intensities werecalculated without applying any background subtractionand the log2ratio values were normalized following aLOESS regression and visualized using the R statisticalsoftware [36]. The data from these experiments havebeen deposited in the Gene Expression Omnibus(http://www.ncbi.nlm.nih.gov/geo/) under the accessionGSE23576.Additional materialAdditional file 1: Microsoft PowerPoint file of theCGHViewer viewof T. cruzi chromosome 35 showing the CNV generated byknockout of one copy each of ECH1 and ECH2 (enoyl-CoAhydratase/isomerase family protein; Tc00.1047053511529.160,Tc00.1047053511529.150). Each dot represents an oligonucleotideprobe. The CL-Brener strain, which was used as the reference strain forgenome sequencing, is hybrid, thus probes were designed to non-Esmeraldo (non-Esm) sequences (blue), Esmeraldo-like (Esm) sequences(green), non-Esm gene family sequences (black), and Esm gene familysequences (gray). Positive log2 ratios of signal intensities (wild typestrain/knockout strain) represent deletion in the knockout strain andnegative log2 ratios represent amplification in the knockout strain,relative to wt T. cruzi. Units for the X axis (Position) are base pairs. Insetin panel A is the GBrowse view of the locus (ECH genes purple circle).Panel B is a close-up view of the locus on chromosome 35.Additional File 2: PowerPoint file of all of the array data ordered bychromosome. Representative plots from 16 hybridizations for each T.cruzi chromosome. Each dot represents an oligonucleotide probe. TheCL-Brener strain, which was used as the reference strain for genomesequencing, is hybrid, thus probes were designed to non-Esmeraldo(non-Esm) sequences (blue), Esmeraldo-like (Esm) sequences (green),non-Esm gene family sequences (black), and Esm gene family sequences(gray). In each panel positive log2 ratios of signal intensities (test strain/reference) represent amplification in the test strain and negative log2ratios represent deletion in the test strain, relative to CL-Brener, whichwas the reference strain in all hybridizations. Boxed regions were thefeatures selected for chromosome typing as explained in Figure 3. Thedifferent patterns observed for each chromosome are displayed at thebottom with letters corresponding to the typing letters in Figure 3. Blueboxes denote lower copy number in the test strain versus the referencestrain, red boxes higher copy number in the test strain, and black boxesequal copy number in the test and reference strains. In each case foreach chromosome, the CL-Brener type was the default type “A.” Also,while two strains may have been assigned to the same CNV signaturetype for a given chromosome, they were not necessarily identical forthat chromosome, as not every single CNV for every single chromosomewas used in the typing (such an analysis would render everychromosome for every strain unique and make finding common patternsimpossible). Note that chromosomes 4, 5, 18, 28, and 29 did not presentsufficiently informative typing regions. Thus, all strains were type “A” forthese chromosomes. Also, the CNV were haplotype specific. Therefore, insome cases this made the up or down calls (box color) appear incorrect,especially if the log2ratio for the feature was off the scale making itappear as if the boxed region is referring to the other haplotype. Forexample see PalDa for chromosome 8. In the boxed region the green(Esm) probes appear to be up yet the box is blue, indicating lower copynumber, because the feature refers to non-Esm probes which are off thescale of the figure. Lastly, the boxes are guides to identifying the CNVused for typing, but they do not represent the exact bounds of thosetyping regions. In some cases due to the scale of the figure CNV that areclose to the typing CNV appear to be part of the typing CNV.Additional file 3: A java-based executable file for viewing all of theCGH array data from this study. The CGH data were visualized andexplored using ‘CGH_Viewer,’ which was written in the Javaprogramming language. Additional file 2 is a Windows executable (.exe)that will install the CGH_Viewer, along with Java if it is not detected, onthe target computer. The CGH_Viewer takes as input the mapping ofprobes from the microarray to the assembled chromosomes of T. cruzi aswell as the results of 1 or more experiments between 2 strains. Theprobe-to-chromosome mapping file and the 16 experimental files areprovided in the “Data” sub-folder. The results (dot plots of log2 ratios) ofmultiple chromosomes and multiple experiments may be viewedsimultaneously. Zooming on an area of interest on a chromosome willshow the same region across all experiments. The data can be filteredMinning et al. BMC Genomics 2011, 12:139http://www.biomedcentral.com/1471-2164/12/139Page 9 of 11based on 1) the haplotype of sequence from which the probes weredesigned as well, 2) the ID or name of the sequence from which theprobe was designed, 3) by raw intensity of the probe (a qualitymeasurement), or 4) a sub-region of a chromosome. For additionalinformation, see the provided documentation in the CGH_Viewerprogram folder (c:\Program Files\CGH_Viewer\).Additional file 4: Excel file containing the average normalized log2ratios of signal intensities (test strain/CL Brener) for each codingand non-coding region in the T. cruzi genome for each of thehybridizations performed in this study. Esmeraldo-like (Esm) and non-Esmeraldo (non-Esm) probes for the indicated regions are averagedseparately. Density is the genomic range (in base pairs) divided by thenumber of probes covering that range.Additional file 5: MSWord file containing a chart of the distributionof genes associated with hotspot regions having less than 5candidates on the arrays. Candidates were determined by having atleast 5 probes within a maximum sequence size of 500 bp.AcknowledgementsWe acknowledge Angel Padillia, Patricio Diosque, Miriam Postan, and RobertSabatini for supplying T. cruzi strains, and John Kelly and Bianca Zingales forsupplying CL-Brener and Esmeraldo DNA, respectively. We also acknowledgeCharles Rosenberg for helpful discussions. Lastly, we acknowledge Dan Xufor providing DNA for the initial validation experiments. RLT, TM, and DBWwere supported by for this work by P01AI044979.Author details1Center for Tropical and Emerging Global Diseases, University of Georgia,Athens, Georgia, 30602, USA. 2Department of Zoology, University of BritishColumbia, Vancouver, British Columbia V6T 1Z4, Canada.Authors’ contributionsTAM designed and performed the microarray experiments, analyzed andinterpreted the data, and wrote the manuscript. DBW assisted in thebioinformatics and data analysis and wrote the CGH_Viewer program. SFdesigned the microarrays, performed validation experiments, and wrote thewithin-hybridization normalization program. RLT initiated and guided theproject, analyzed and interpreted data, and wrote the manuscript. Allauthors have read and approved the final manuscript.Received: 31 August 2010 Accepted: 7 March 2011Published: 7 March 2011References1. Tibayrenc M, Ward P, Moya A, Ayala FJ: Natural populations ofTrypanosoma cruzi, the agent of Chagas disease, have a complexmulticlonal structure. Proc Natl Acad Sci USA 1986, 83(1):115-119.2. Woo PT, Soltys MA: Animals as reservoir hosts of human trypanosomes.Journal of wildlife diseases 1970, 6(4):313-322.3. Tibayrenc M: Genetic subdivisions within Trypanosoma cruzi (DiscreteTyping Units) and their relevance for molecular epidemiology andexperimental evolution. Kinetoplastid Biol Dis 2003, 2(1):12.4. Zingales B, Andrade SG, Briones MR, Campbell DA, Chiari E, Fernandes O,Guhl F, Lages-Silva E, Macedo AM, Machado CR, Miles MA, Romanha AJ,Sturm NR, Tibayrenc M, Schijman AG: A new consensus for Trypanosomacruzi intraspecific nomenclature: second revision meeting recommendsTcI to TcVI. Mem Inst Oswaldo Cruz 2009, 104(7):1051-1054.5. Westenberger SJ, Barnabe C, Campbell DA, Sturm NR: Two hybridizationevents define the population structure of Trypanosoma cruzi. Genetics2005, 171(2):527-543.6. El-Sayed NM, Myler PJ, Bartholomeu DC, Nilsson D, Aggarwal G, Tran AN,Ghedin E, Worthey EA, Delcher AL, Blandin G, et al: The genome sequenceof Trypanosoma cruzi, etiologic agent of Chagas disease. Science 2005,309(5733):409-415.7. Martin DL, Weatherly DB, Laucella SA, Cabinian MA, Crim MT, Sullivan S,Heiges M, Craven SH, Rosenberg CS, Collins MH, Sette A, Postan M,Tarleton RL: CD8+ T-Cell responses to Trypanosoma cruzi are highlyfocused on strain-variant trans-sialidase epitopes. PLoS pathogens 2006,2(8):e77.8. Buscaglia CA, Campo VA, Frasch AC, Di Noia JM: Trypanosoma cruzisurface mucins: host-dependent coat diversity. Nature reviews 2006,4(3):229-236.9. Weatherly DB, Boehlke C, Tarleton RL: Chromosome level assembly of thehybrid Trypanosoma cruzi genome. BMC genomics 2009, 10:255.10. Xu D, Brandan CP, Basombrio MA, Tarleton RL: Evaluation of highefficiency gene knockout strategies for Trypanosoma cruzi. BMCmicrobiology 2009, 9:90.11. Brisse S, Dujardin JC, Tibayrenc M: Identification of six Trypanosoma cruzilineages by sequence-characterised amplified region markers. MolBiochem Parasitol 2000, 111(1):95-105.12. Tomazi L, Kawashita SY, Pereira PM, Zingales B, Briones MR: Haplotypedistribution of five nuclear genes based on network genealogies andBayesian inference indicates that Trypanosoma cruzi hybrid strains arepolyphyletic. Genet Mol Res 2009, 8(2):458-476.13. Llewellyn MS, Miles MA, Carrasco HJ, Lewis MD, Yeo M, Vargas J, Torrico F,Diosque P, Valente V, Valente SA, Gaunt MW: Genome-scale multilocusmicrosatellite typing of Trypanosoma cruzi discrete typing unit I revealsphylogeographic structure and specific genotypes linked to humaninfection. PLoS pathogens 2009, 5(5):e1000410.14. Reina-San-Martin B, Degrave W, Rougeot C, Cosson A, Chamond N,Cordeiro-Da-Silva A, Arala-Chaves M, Coutinho A, Minoprio P: A B-cellmitogen from a pathogenic trypanosome is a eukaryotic prolineracemase. Nature medicine 2000, 6(8):890-897.15. Hurles M: Gene duplication: the genomic trade in spare parts. PLoSbiology 2004, 2(7):E206.16. Brown MV, Reader JS, Tzima E: Mammalian aminoacyl-tRNA synthetases:cell signaling functions of the protein translation machinery. Vascularpharmacology 52(1-2):21-26.17. Jones C, Todeschini AR, Agrellos OA, Previato JO, Mendonca-Previato L:Heterogeneity in the biosynthesis of mucin O-glycans fromTrypanosoma cruzi tulahuen strain with the expression of novelgalactofuranosyl-containing oligosaccharides. Biochemistry 2004,43(37):11889-11897.18. Spadiliero B, Nicolini C, Mascetti G, Henriquez D, Vergani L: Chromatin ofTrypanosoma cruzi: In situ analysis revealed its unusual structure andnuclear organization. Journal of cellular biochemistry 2002, 85(4):798-808.19. Polakova S, Blume C, Zarate JA, Mentel M, Jorck-Ramberg D, Stenderup J,Piskur J: Formation of new chromosomes as a virulence mechanism inyeast Candida glabrata. Proc Natl Acad Sci USA 2009, 106(8):2688-2693.20. Muller H, Thierry A, Coppee JY, Gouyette C, Hennequin C, Sismeiro O,Talla E, Dujon B, Fairhead C: Genomic polymorphism in the population ofCandida glabrata: gene copy-number variation and chromosomaltranslocations. Fungal Genet Biol 2009, 46(3):264-276.21. Andrade S: Caracterizacao de cepas do Trypanosoma cruzi isoladas noReconcavo Bahiano. Revista de Patologica Tropical 1974, 3:65-121.22. Andrade SG, Magalhaes JB: Biodemes and zymodemes of Trypanosomacruzi strains: correlations with clinical data and experimental pathology.Rev Soc Bras Med Trop 1996, 30(1):27-35.23. Zingales B, Stolf BS, Souto RP, Fernandes O, Briones MR: Epidemiology,biochemistry and evolution of Trypanosoma cruzi lineages based onribosomal RNA sequences. Mem Inst Oswaldo Cruz 1999, 94(Suppl1):159-164.24. Briones MR, Souto RP, Stolf BS, Zingales B: The evolution of twoTrypanosoma cruzi subgroups inferred from rRNA genes can becorrelated with the interchange of American mammalian faunas in theCenozoic and has implications to pathogenicity and host specificity. MolBiochem Parasitol 1999, 104(2):219-232.25. Hall CA, Polizzi C, Yabsley MJ, Norton TM: Trypanosoma cruzi prevalenceand epidemiologic trends in lemurs on St. Catherines Island, Georgia. JParasitol 2007, 93(1):93-96.26. Llewellyn MS, Lewis MD, Acosta N, Yeo M, Carrasco HJ, Segovia M, Vargas J,Torrico F, Miles MA, Gaunt MW: Trypanosoma cruzi IIc: phylogenetic andphylogeographic insights from sequence and microsatellite analysis andpotential impact on emergent Chagas disease. PLoS neglected tropicaldiseases 2009, 3(9):e510.27. del Puerto R, Nishizawa JE, Kikuchi M, Iihoshi N, Roca Y, Avilas C, Gianella A,Lora J, Velarde FU, Renjel LA, Miura S, Higo H, Komiya N, Maemura K,Hirayama K: Lineage analysis of circulating Trypanosoma cruzi parasitesMinning et al. BMC Genomics 2011, 12:139http://www.biomedcentral.com/1471-2164/12/139Page 10 of 11and their association with clinical forms of Chagas disease in Bolivia.PLoS neglected tropical diseases 4(5):e687.28. Muller LA, McCusker JH: A multispecies-based taxonomic microarrayreveals interspecies hybridization and introgression in Saccharomycescerevisiae. FEMS yeast research 2009, 9(1):143-152.29. Gaunt MW, Yeo M, Frame IA, Stothard JR, Carrasco HJ, Taylor MC, Mena SS,Veazey P, Miles GA, Acosta N, de Arias AR, Miles MA: Mechanism ofgenetic exchange in American trypanosomes. Nature 2003,421(6926):936-939.30. Minning TA, Weatherly DB, Atwood J, Orlando R, Tarleton RL: The steady-state transcriptome of the four major life-cycle stages of Trypanosomacruzi. BMC genomics 2009, 10:370.31. Ozaki LS, Cseko YMT: Genomic cloning and related techniques. In Genesand antigens of parasites: A laboratory manual. second edition. Edited by:Morel CM. Rio de Janeiro, Brazil: Dept. of Biochemistry and MolecularBiology, Instituto Oswaldo Cruz; 1984.32. Flibotte S, Moerman DG: Experimental analysis of oligonucleotidemicroarray design criteria to detect deletions by comparative genomichybridization. BMC genomics 2008, 9:497.33. Singh-Gasson S, Green RD, Yue Y, Nelson C, Blattner F, Sussman MR,Cerrina F: Maskless fabrication of light-directed oligonucleotidemicroarrays using a digital micromirror array. Nature biotechnology 1999,17(10):974-978.34. Markham N: Hybrid: A software system for nucleic acid folding,hybridizing, and melting conditions. Rensselaer Polytechnic Institute Troy,NY; 2003.35. Maydan JS, Flibotte S, Edgley ML, Lau J, Selzer RR, Richmond TA, Pofahl NJ,Thomas JH, Moerman DG: Efficient high-resolution deletion discovery inCaenorhabditis elegans by array comparative genomic hybridization.Genome research 2007, 17(3):337-347.36. R: A language and environment for statistical computing. [http://www.R-project.org].37. Pan SC: Establishment of clones of Trypanosoma cruzi and theircharacterization in vitro and in vivo. Bulletin of the World HealthOrganization 1982, 60(1):101-107.38. Miles MA, Toye PJ, Oswald SC, Godfrey DG: The identification byisoenzyme patterns of two distinct strain-groups of Trypanosoma cruzi,circulating independently in a rural area of Brazil. Trans R Soc Trop MedHyg 1977, 71(3):217-225.39. Miles MA, Povoa MM, de Souza AA, Lainson R, Shaw JJ, Ketteridge DS:Chagas’s disease in the Amazon Basin: Ii. The distribution ofTrypanosoma cruzi zymodemes 1 and 3 in Para State, north Brazil. TransR Soc Trop Med Hyg 1981, 75(5):667-674.40. Engel JC, Dvorak JA, Segura EL, Crane MS: Trypanosoma cruzi: biologicalcharacterization of 19 clones derived from two chronic chagasicpatients. I. Growth kinetics in liquid medium. The Journal of protozoology1982, 29(4):555-560.41. Diosque P, Barnabe C, Padilla AM, Marco JD, Cardozo RM, Cimino RO,Nasser JR, Tibayrenc M, Basombrio MA: Multilocus enzyme electrophoresisanalysis of Trypanosoma cruzi isolates from a geographically restrictedendemic area for Chagas’ disease in Argentina. Int J Parasitol 2003,33(10):997-1003.42. Postan M, Dvorak JA, McDaniel JP: Studies of Trypanosoma cruzi clones ininbred mice. I. A comparison of the course of infection of C3H/HEN-mice with two clones isolated from a common source. Am J Trop MedHyg 1983, 32(3):497-506.43. Basombrio MA, Besuschio S, Cossio PM: Side effects of immunization withliver attenuated Trypanosoma cruzi in mice and rabbits. Infect Immun1982, 36(1):342-350.44. Breniere SF, Bosseno MF, Telleria J, Bastrenta B, Yacsik N, Noireau F,Alcazar JL, Barnabe C, Wincker P, Tibayrenc M: Different behavior of twoTrypanosoma cruzi major clones: transmission and circulation in youngBolivian patients. Exp Parasitol 1998, 89(3):285-295.45. Cariola J, Prado R, Agosin M, Christen R: Susceptibility of the hamster andPeromyscus to experimental Trypanosoma cruzi infection (Tulahuenstrain). Boletin de informaciones parasitarias chilenas 1950, 5(4):44-45.46. Brener Z, Chiari E: Morphological Variations Observed in Different Strainsof Trypanosoma Cruzi. Rev Inst Med Trop Sao Paulo 1963, 19:220-224.47. Carvalheiro Jda R, Collares EF: Studies on the behavior, in mice, of ahighly virulent strain (Y strain) of Trypanosoma cruzi after passage inTriatoma, rats and cultures. Revista brasileira de biologia 1965,25(2):169-175.48. Brisse S, Barnabe C, Banuls AL, Sidibe I, Noel S, Tibayrenc M: Aphylogenetic analysis of the Trypanosoma cruzi genome project CLBrener reference strain by multilocus enzyme electrophoresis andmultiprimer random amplified polymorphic DNA fingerprinting. MolBiochem Parasitol 1998, 92(2):253-263.doi:10.1186/1471-2164-12-139Cite this article as: Minning et al.: Widespread, focal copy numbervariations (CNV) and whole chromosome aneuploidies in Trypanosomacruzi strains revealed by array comparative genomic hybridization. BMCGenomics 2011 12:139.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at www.biomedcentral.com/submitMinning et al. BMC Genomics 2011, 12:139http://www.biomedcentral.com/1471-2164/12/139Page 11 of 11


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