Real-world effects of using gnomAD 4.1.0 and All of Us population reference data sets on reporting of variants of uncertain significance.

Abstract

Clinical testing of rare genetic variants relies on population genomic data sets as a source of evidence. New data sets from the Genome Aggregation Database and All of Us Research Program are many-fold larger than previous data sets, and the latter includes highly detailed phenotype data. The effect of these data sets on variant classification and reporting for highly penetrant pediatric-onset disease is not well characterized; however, one likely effect is reduced reporting of variants of uncertain significance (VUS). We retrospectively identified VUS previously reported by our clinically licensed laboratory and evaluated whether they are still reportable based on new reference data sets. By examining allele counts in new data sets, we identified 24 variants that are likely no longer reportable. Additionally, the All of Us phenotype data suggest that an additional 10 VUS are no longer reportable. Overall, we found that nearly one-fifth of VUS (34/173, 19.6%) are no longer reportable. We conclude that the use of these new data sets is likely to reduce reported VUS for highly penetrant pediatric-onset disease. This may be further augmented through updated gene-specific thresholds and improved accessibility of All of Us phenotype data.