Ryan Kele Waples:
|
This thesis covers work on three basic concepts in population genetics: relatedness, admixture, and linkage disequilibrium, and contains both new statistical methods as well as application of recently published methods to better understand the recent human past in Greenland.
The first manuscript included in this thesis addresses a gap in the availability of methods to infer rela-tedness with limited data. Relatedness matters for all aspects of population genetics, but most methods to infer relatedness rely on the availability of population allele frequencies and accurate genotype data. We present a method that can infer close familial relationships without relying on estimates of popula-tion allele frequencies, and directly from low-depth (4x) sequencing data without genotype calling. It requires genetic data from only two individuals and has the potential to expand the number of studies able to infer relatedness despite limited data.
In the second manuscript we examine the history of contact between the Greenlandic Inuit and Eu-ropeans from a genetic perspective. The current population of Greenland has experienced substantial gene flow from Europe, but the European source countries of this ancestry was genetically unknown. There is a tight historical relationship between Greenland and Denmark, but there is also a history of DanoNorwegian and German missionaries, Dutch whalers, as well as other European contact with Greenland. Using dense SNP array data from Greenlanders and Europeans we identify Denmark as the primary source of European ancestry in Greenland, and quantify the ancestry contribution from 14 different European countries. We discuss in detail how these results reflect the history of Green-land/European contact.
In the third and last manuscript we present a software tool for estimating linkage disequilibrium (LD) in admixed populations. The LD in a population is affected by many aspects of the populations history, including effective population size and past admixture. LDadmix estimates the two-locus haplotype frequencies within the source ancestries of a recently admixed population. These two-locus haplotype frequencies reflect the LD within each ancestry source prior to the admixture event. Through simulations and application to real data, we show that LDadmix can recover LD patterns in different admixture scenarios and also infer an elevated LD decay curve for the ancestral American ancestry, a signal that was previously masked by recent African and European admixture.
Together, these manuscripts highlight the continuing need for genetic methods that can be applied in the challenging and data-limited scenarios that will continue to be frequent in biology despite the recent expansion of available genetic data. They also highlight how new insights can be gained about different populations, like the Greenlandic, when such methods are developed and applied.