Shirin Nataneli

Runs of homozygosity (ROH) are genomic regions that arise when identical haplotypes are inherited from a shared ancestor. In this study, we explored ROH across 556 whole-genome sequences from domesticated and non-domesticated dogs. Then, we leveraged ROH from 466 breed dogs, representing 13 breed groups and 13 phenotypic traits, to investigate associations between genetic diversity and non-disease phenotypes. We identified significant associations between the ROH-based inbreeding coefficient (FROH) and multiple phenotypes. These include three quantitative traits (height, weight, lifespan) and ten morphological and coat-related traits. After correcting for population structure, we identified more than 45 genes associated with quantitative traits that exceeded suggestive or genome-wide significance (GWS) thresholds. We also observed distinct patterns of inbreeding across dog populations, including elevated levels of long ROH in modern breed dogs relative to more ancient breeds, consistent with intensive breeding practices during Victorian-era breed formation. Together, our results demonstrate how domestication, demographic bottlenecks, and selective breeding have shaped patterns of homozygosity and contributed to the genetic architecture of complex traits in dogs, highlighting an important role for non-additive genetic variation and polygenicity.

Complex demographic events have shaped human history, leaving signatures of genetic variation across the genome. Here, we investigate the recent evolutionary dynamics of admixed populations that descend from distinct ancestral sources. We present a discrete, generalizable model of admixture that leverages ancestry switches, which are recombination breakpoints that mark changes in ancestral origin along a chromosome. We derive analytical expectations for the number of ancestry switches within a genomic segment as functions of recombination rate, ancestry heterozygosity, and effective population size. We then extend these expectations to incorporate population-specific recombination maps. Our theoretical predictions are in close agreement with forward-in-time simulations that trace ancestry junction accumulation following an initial admixture event under both constant and variable recombination models. We observe minimal variability in switch counts across ten simulation replicates, underscoring the robustness of the theoretical expectation. Furthermore, model-based switch counts, parameterized using literature-informed demographic values, agree with empirical observations from African-American individuals in the 1000 Genomes Project. For example, when modeling human chromosome 1, we found a mean of approximately six switches per haplotype, which aligns with the theoretical expectation under an initial African ancestry proportion of 0.85 and agrees with published estimates from other African-American cohorts. Overall, the model provides a new route for using ancestry switches to understand how recombination and demography jointly shape ancestry patterns in admixed populations without requiring separation into parental sources.

Complex demographic events have shaped human history and genetic variation across the genome. Here, we investigate the recent evolutionary history of admixed populations that descend from distinct ancestral sources. We present a discrete, generalizable model of admixture that leverages ancestry switches, which are recombination breakpoints that mark changes in ancestral origin along a chromosome. We derive analytical expectations for the number of ancestry switches within a genomic segment as functions of recombination rate, ancestry heterozygosity, and effective population size. We then extend these expectations to incorporate population-specific recombination maps. Our theoretical predictions are in close agreement with forward-in-time simulations that we use to trace ancestry junction accumulation since an initial admixture event with both constant and variable recombination models. We observe minimal variability in switch counts across ten simulation replicates, underscoring the robustness of the theoretical expectation. Furthermore, model-based switch counts, parameterized using literature-informed demographic values, agree with empirical observations from African American individuals in the 1000 Genomes Project. For example, when modeling human chromosome 1, we found a mean of approximately six switches per haplotype, which aligns with the theoretical expectation under an initial African ancestry proportion of 0.85, and agrees with published estimates from other African-American cohorts. Overall, the model provides a new route for using ancestry switches to understand how recombination and demography jointly shape ancestry patterns in admixed populations without requiring separation into parental sources.