Challenges imposed by minor reference alleles on the identification and reporting of clinical variants from exome data

Abstract

BackgroundThe conventional variant calling of pathogenic alleles in exome and genome sequencing requires the presence of the non-pathogenic alleles as genome references. This hinders the correct identification of variants with minor and/or pathogenic reference alleles warranting additional approaches for variant calling.ResultsMore than 26,000 Exome Aggregation Consortium (ExAC) variants have a minor reference allele including variants with known ClinVar disease alleles. For instance, in a number of variants related to clotting disorders, the phenotype-associated allele is a human genome reference allele (rs6025, rs6003, rs1799983, and rs2227564 using the assembly hg19). We highlighted how the current variant calling standards miss homozygous reference disease variants in these sites and provided a bioinformatic panel that can be used to screen these variants using commonly available variant callers. We present exome sequencing results from an individual with venous thrombosis to emphasize how pathogenic alleles in clinically relevant variants escape variant calling while non-pathogenic alleles are detected.ConclusionsThis article highlights the importance of specialized variant calling strategies in clinical variants with minor reference alleles especially in the context of personal genomes and exomes. We provide here a simple strategy to screen potential disease-causing variants when present in homozygous reference state.Electronic supplementary materialThe online version of this article (10.1186/s12864-018-4433-3) contains supplementary material, which is available to authorized users.

Is it time to change the reference genome?

“…Alongside personal genomes, major alleles have been useful in improving disease analysis and alignment [45], especially in regions of high variation (such as the human leukocyte antigen (HLA) locus) or for clinically relevant analyses where variant pathogenicity was misattributed (see examples in [48, 71]). In the same way that the consensus sequences of transcription-factor-binding motifs represent the most common version of the motif, a consensus genome represents the most common alleles and variants within a population.…”

“…For example, Dewey et al [45] used major alleles in the sequence to study the HLA. Minor alleles (assessed in [71]) or those that are absent from certain ethnically distinct populations cause trouble in downstream clinical assessments [73] and tools have been built to screen for them [48]. The Locus Reference Genomic Project (LRG) is working to improve on gene sequences, primarily to correct for minor and disease alleles in variant significance assessments.…”

Construction and integration of three de novo Japanese human genome assemblies toward a population-specific reference

“…As such, the reference genome inevitably harbors rare or even private variants. Over 90,000 rare variants were used as a reference allele including disease-susceptibility variants for thrombophilia and type 2 diabetes 8,9 . Inclusion of such variants in the reference can lead to erroneous and confusing results of short-read mapping or variant calling 9 .…”

Quality control of large genome datasets using genome fingerprints

“…Minor reference alleles are a problem for identifying and reporting clinical variants 22, 23 , and discrepant ALT alleles could produce similar issues. Cross-comparison of TGP versions identifies many discrepancies of alternate alleles observed.…”