Large-scale next-generation sequencing (NGS) datasets have been transformative for informing clinical variant interpretation and as reference panels for statistical and population genetic efforts. While such resources are often treated as ground truth, we find that in widely used reference datasets such as the Genome Aggregation Database (gnomAD), some variants pass gold standard filters yet are systematically different in their genotype calls across sequencing technologies. The inclusion of such discordant sites in study designs involving multiple sequencing platforms (e.g. whole genome and/or different whole-exome captures) could bias results and lead to false-positive hits in association studies due to technological artifacts rather than a true relationship to the phenotype. Here, we describe this phenomenon of discordant genotype calls across sequencing technologies, characterize the error mode of wrong calls, provide a blacklist of discordant sites identified in gnomAD that should be treated with caution in analyses, and present a metric and machine learning classifier trained on gnomAD data to identify likely discordant variants in other datasets. We find that different NGS technologies have different sets of variants at which this problem occurs but that there are characteristic variant features that can be used to predict discordant behavior. Discordant sites are largely shared across ancestry groups, though different populations are powered for discovery of different variants. We find that the most common error mode is that of a variant being heterozygous for one platform and homozygous for the other, with heterozygous in the genomes and homozygous reference in the exomes making up the majority of miscalls.

Download data

• Downloaded 328 times
• Download rankings, all-time:
  • Site-wide: 151,656
  • In genomics: 8,051
• Year to date:
  • Site-wide: 23,180
• Since beginning of last month:
  • Site-wide: 47,704

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide