• Biases and caveats to implementing genomic medicine in diverse populations

      Kessler, Michael D.; O'Connor, Timothy D.; 0000-0003-1258-5221 (2019)
      Genetic science has traditionally focused on the study of European populations, which has resulted in the under-representation of other ancestral backgrounds across the vast majority of genetic resources. Since data derived from a limited number of ancestral backgrounds are unlikely to represent the biological variation inherent to diverse populations, genetic and genomic methods and models predicated on this Eurocentric data will have questionable accuracy and limited value when applied to diverse patient populations. This can drive scientific and/or clinical disparities, which can then seem opaque and be difficult to resolve. The work outlined in this thesis aims to increase the prevalence of ancestrally informed genetic science by characterizing bias deriving from a lack of ancestry awareness, evaluating ancestral representation in genomic resources, and identifying novel ancestry-informed biological signal. I first characterize bias by evaluating a typical clinical variant prioritization pipeline, and I demonstrate a significant positive correlation between African ancestry proportion and the identified number of clinically evaluable variants. I then more directly explore ancestral representation across translational resources by estimating the genetic ancestry for 1,018 common cancer cell lines. This analysis highlights the marked ancestral underrepresentation that exists among preclinical cancer cell line models, and identifies novel signals of ancestry-specific gene expression and somatic mutation. Finally, I evaluate how de novo mutation rates vary across diverse human populations from the NHLBI Trans-Omics for Precision Medicine (TOPMed) program. I find associations with heterozygosity, a reduced mutation rate in the Amish founder population, and near zero estimates of narrow-sense heritability. On the whole, these findings help to quantify the effects of ancestral diversity and under-representation on the application of genomic medicine.
    • Characterizing Genetic Drift and Migration between Populations

      Harris, Daniel Nathan; O'Connor, Timothy D. (2019)
      Genetic drift and migration are two of the main mechanisms of evolution at the genetic level. These mechanisms greatly impact the genetic relatedness of populations. Therefore, by studying how genetic drift and migration impact populations, it is possible to reconstruct elements of population history throughout a geographic region. This thesis examines how genetic drift and migration rapidly impact human population structure. My first aim studies the peopling of the Peruvian region and models migration patterns. Here, I suggest that the peopling of Peru was a rapid process that occurred by 12,000 years ago. Furthermore, migration modeling suggests that the majority of migration in Peru is from populations in the high altitude Andes Mountains to the low altitude coast and Amazon. My second aim, then analyzes genetic drift and migration on the fatty acid desaturase gene region, where there is a suggestion of a public health problem due to the apparent fixation of the low efficient ancestral haplogroup in Native Americans. I confirm the near fixation of the ancestral haplogroup in Native American ancestry, and suggest that understanding the frequency of the ancestral haplogroup in the Native American founding population is crucial for determining how this apparently detrimental haplogroup reached near fixation. Furthermore, I suggest that the global distribution of the haplogroup is not indicative of Neanderthal introgression and is instead likely an ancient polymorphism that arose during the divergence of modern humans and Neanderthals. For my final aim, I use rare variant and haplotype based approaches to discern the population structure and evolutionary history of Samoans. I suggest that the early population size was small and a strong population expansion occurred approximately 35 generations ago, consistent with the Samoan archaeological record. In addition, Samoan population structure is likely strongly driven by urbanization, and not biogeography. In sum, this thesis demonstrates that migration and genetic drift can rapidly impact the genetic variation in populations. In addition, I demonstrate that rare variant and haplotype based approaches are essential for examining the impact of migration and genetic drift between populations within an extremely recently founded region with minimal geographic barriers.