One species, two genomes: A critical assessment of inter-isolate variation and identification of assembly incongruence inHaemonchus contortus

Abstract

BackgroundNumerous quality issues may compromise genomic data's representation of its underlying organism. In this study, we compared two genomes published by different research groups for the parasitic nematode Haemonchus contortus, corresponding to divergent isolates. We analyzed differences between the genomes, attempting to ascertain which were attributable to legitimate biological differences, and which to technical error in one or both genomes. ResultsWe found discrepancies between the H. contortus genomes in both assembly and annotation. The genomes differed in representation of genes that are highly conserved across eukaryotes, with clear evidence of misassembly underlying conserved genes missing from one genome or the other. Only 45% of genes in one genome were orthologous to genes in the other genome, with one genome exhibiting almost as much orthology to C. elegans as its counterpart H. contortus strain. The two genomes differed substantially in probable causes underlying this unexpectedly low orthology. One genome included many more inparalogues than the other, and more frequently assembled inparalogues together on the same portions of contiguous sequence. It also exhibited cases of better-conserved gene position relative to C. elegans. ConclusionThe discrepancies between the two genomes far exceeded those expected as a consequence of biological differences between the two H. contortus isolates. This implies substantial quality issues in one or both genomes, suggesting that researchers must exercise caution when using genomic data for newly sequenced species. BackgroundIn the decade following publication of the human genome, the use of genomic sequence data has increased dramatically [1], with a continued exponential drop in sequencing costs pushing researchers to utilize sequencing data in sundry studies spanning biology and medicine.Genomic data production, however, is complicated, suffering from numerous issues that compromise the resulting data's reflection of its underlying organism. The precision of genomic data can easily be mistaken for accuracy-though every nucleotide, gene boundary, and splice site is denoted exactly, uncertainty in the processes producing them is poorly represented in the resulting genome [2]. Thus, researchers who draw on genomic data without understanding their potential pitfalls will place undue faith in the validity of results built atop genomic foundations.Though sequencing costs have dropped precipitously, assembly, annotation, and analysis remain complex processes [3] subject to problems affecting the quality of the resulting genome [4]. The shortread sequencing technology prevalent today struggles to resolve repetitive regions [5], resulting in repeat sequences being collapsed, or physically distant regions being mistakenly adjoined because they flank highly similar sequences. With sequencing platforms being 85.0% to 99.9% accurate in each called base [4], distinguishing misread bases from legitimate variation stemming from rare alleles is difficult. This problem is...

Extensive genomic and transcriptomic variation defines the chromosome-scale assembly ofHaemonchus contortus, a model gastrointestinal worm

“…Draft genome assemblies were published in 2013 for two anthelmintic-sensitive isolates, MHco3(ISE).N1 [21] and McMaster [22] ; both assemblies were produced with a combination of short-read, high-throughput sequencing technologies (including both short-and long-insert libraries), resulting in estimated genome sizes of 370 Mbp and 320 Mbp, respectively. However, direct comparison of these two assemblies shows discordance, revealing clear differences in gene family composition [23] and variation in assembly quality and gene content [24] . Although some of these differences reflect true biological variation in this highly polymorphic species [25][26][27] , many differences are technical artefacts, as revealed by comparing fragmented genome assemblies from sequencing DNA derived from pools of a genetically diverse organism.…”

Genomic and transcriptomic variation defines the chromosome-scale assembly of Haemonchus contortus, a model gastrointestinal worm

“…Draft genome assemblies were published in 2013 for two anthelmintic-sensitive isolates, MHco3 (ISE).N1 23 and McMaster 24 ; both assemblies were produced using short-read, high-throughput sequencing technologies (including both short-and long-insert libraries), resulting in assembly sizes of 370 Mbp and 320 Mbp, respectively. However, direct comparison of these assemblies shows discordance, revealing clear differences in gene family composition 25 and variation in assembly quality and gene content 26 . Although some of these differences reflect biological variation in this highly polymorphic species [27][28][29] , many differences are technical artefacts, as revealed by comparing fragmented genome assemblies from sequencing DNA derived from pools of a genetically diverse organism.…”