Evolution-guided association mapping using long-read genomes from a self-identified Black or African American cohort in All of Us

BackgroundIdentifying functional variants that influence human health remains a challenge in human genetics. Conventional genotype-phenotype association studies test millions of variants, imposing a substantial multiple-testing burden as genomic and biobank datasets expand in size and variant diversity. Additionally, the contribution of structural variation, which is a substantial driver of genetic diversity, remains underexplored in these studies. Signatures of recent positive selection offer a powerful biological basis for prioritizing potentially functional variation, although demographic history can confound their detection and interpretation in these analyses. ResultsWe performed evolution-guided association mapping using long-read genomes from 1,027 self-identified Black or African American participants in the All of Us Research Program and short-read genomes from an additional 38,552 matched participants of African-ancestry. Using demography-calibrated inferences, we identified 228 genomic regions with local ancestry enrichment or depletion and 1,616 candidate loci showing putative recent positive selection signals. These loci were enriched for genic and regulatory sequences, as well as previously reported eQTL, sQTL, and GWAS associations. Leveraging the resolution of the long-read data, we mapped 15 structural variants to these candidate loci, including 10 located within 1,000 base pairs of putative cis-regulatory enhancers. Admixture- and selection-based phenome-wide association testing identified 13 and 446 nominal associations, respectively, 17 of which remained significant following replication and meta-analysis. ConclusionsOur results support the use of evolutionary signatures as a robust framework for uncovering clinically relevant variation, including structural variants, in underrepresented populations.