Human populations across the globe have been shaped by admixture--gene flow between previously diverging groups. The sudden combination of previously distinct genotypes through admixture can rapidly change allele frequencies, heterozygosity, and patterns of linkage-disequilibrium. These processes create new material for both positive and negative selection to act upon, but also depend on the independent adaptive histories of the source populations. Admixed populations also provide powerful test cases for understanding how selection shapes evolution in general, since changes in ancestry patterns in admixed populations are much easier to observe on short timescales compared to changes in allele frequencies in source populations. Despite the ubiquity of admixture, current methods for inferring selection do not consider how admixture changes the action of selection and the genetic signatures that it leaves. Standard methods to detect selection do not work in admixed populations; since selection post-admixture is often on a very short timescale, and admixture- induced shifts in allele frequencies and haplotype structure can obscure classic signals of selection. The lack of appropriate methods constrains our understanding of disease risk and human evolution. Further, few studies have addressed how recombination modulates selection in admixed populations by shuffling haplotypes from distinct source populations and influencing the exposure of deleterious variation. To address these gaps, this proposal tests how two important evolutionary forces--positive and negative selection--shape the genetics of admixed populations. This proposal combines methods development and empirical analyses to provide insight into how admixture shapes fundamental evolutionary processes in multiple admixed African diaspora populations.
Specific Aim 1 will develop statistics to detect positive selection in admixed populations by leveraging local ancestry information to incorporate the effects of admixture on haplotype structure. These new statistics will be integrated into open-access software and applied to infer selection in both simulated and empirical data representing diverse demographic scenarios.
Specific Aim 2 will test how admixture combines the distinct distributions of deleterious variation found in source populations. Tracking the frequencies of segregating deleterious alleles and their membership in runs-of-homozygosity will determine how admixture and the landscape of recombination modulate the exposure of deleterious variation. Characterizing the dynamics of deleterious variation in admixed populations and their source populations will provide a window into how admixture changes genetic load. This proposal will advance methodology for the study of natural selection in admixed populations and elucidate how both positive and negative selection shape patterns of genetic variation and disease risk in understudied admixed populations.
Previously isolated populations often exchange genes, leading to mixtures of ancestry from the contributing populations. Most methods for detecting and understanding natural selection are not designed to account for this mixed ancestry, so admixed populations are often neglected in the context of medical genetics and understanding how selection influences disease risk. This proposal develops methods to understand natural selection in admixed populations and applies these techniques to understand how genetic variation shapes disease risk in understudied admixed populations.
