Human Migration, Population Divergence, and the Accumulation of Deleterious Alleles: Insights from Private Genetic Variation and Whole-exome Sequencing.

Abstract

Private genetic variants---genetic variants found only in a single population in a sample---have been highly informative for studies of evolutionary history. This dissertation uses theoretical modeling and empirical analysis of private genetic variation to understand human demography. I derive theory for analyzing the size distribution of private microsatellite alleles and develop a rarefaction approach to analyze private-allele sharing. Furthermore, I investigate genome-wide patterns of deleterious variants, which are often private. Through the analysis of whole-exome sequences, I describe how potentially deleterious coding variants accumulate and concentrate in inbred genomes.

First, I introduce the concept of generalized private alleles and develop a method to count them while correcting for differences in the number of individuals sampled. I use this method to analyze worldwide human populations and observe an excess of alleles shared between Africa and Oceania. The results support the theory of a coastal migration out of Africa into Oceania separate from the migrations responsible for the majority of the ancestry of the modern populations of Asia.

Next, I explore how population-genetic parameters affect the size distribution of private microsatellite alleles under a two-population coalescent model, assuming the symmetric stepwise mutation model. Using this framework, I theoretically predict that private microsatellites occur in the tails of the allele size distribution more frequently as genetic differentiation between populations increases. Empirically observing this phenomenon in human populations, I conclude that the model accurately describes patterns of private microsatellite alleles in diverged populations.

Finally, I analyze how the genome-wide distribution of runs of homozygosity (ROH) underlies patterns of deleterious variation. Whereas short and intermediate ROH are generated by isolation or bottlenecks, long ROH are likely the result of recent inbreeding. I find that long ROH harbor disproportionately more deleterious homozygotes than is predicted solely by genomic ROH coverage, indicating that inbreeding contributes an abundance of deleterious variation to ROH.

This dissertation expands our knowledge of human population genetics and develops novel theoretical and methodological frameworks to study human migration and population divergence from private genetic variation. Furthermore, it provides insights into the accumulation and concentration of deleterious variation.