Detection of genomic signatures of selection in roe deer and reindeer populations

Abstract

In this thesis I present the outcomes of genetic analyses of several reindeer and roe deer datasets, using two types of data: single nucleotide polymorphism (SNP) data and whole genome sequencing data. I assess the population structure, genetic diversity and demographic history of the study populations and study species, but the main focus is on selection analyses: the detection of genetic signals of selection.

In Chapter 2 I present SNP data analysis outcomes which are suggestive of a shared positive selection event in two reindeer founder populations on the South Atlantic island South Georgia. This finding therefore possibly provides empirical evidence that positive selection can overcome drift in heavily bottlenecked founder populations, and can be detected despite elevated background neutral variation. In addition, I report a new selection scan called Genome Wide Differentiation Scan (GWDS).

In Chapter 3 I infer from a SNP dataset that the effective population size of the native UK roe deer population has numbered several thousand individuals throughout the Holocene. The dataset suggests that neither drift nor positive selection has caused fixed differences between the UK population and the European mainland population, despite a split time of ~1500 generations.

In Chapter 4 I investigate the demographic and evolutionary history of the extant roe deer sister species: the European roe deer (C. capreolus) and the Siberian roe deer (C. pygargus). Whole genome sequences analyses suggest that the two species split maximum 1.6Mya and show pronounced differences in terms of genetic diversity and effective population sizes (Ne). In the species with lower genetic diversity and lower historical Ne, C. capreolus, I find higher proportions of lineage specific amino acid substitutions. This negative relationship between Ne and number of non-synonymous substitutions is suggestive of relaxation of purifying selection, but alternative explanations (such as episodes of positive selection and data artifacts resulting from differences in genome quality) can not be excluded.

In Chapter 5 I discuss the results presented in this thesis in the light of the neutral theory of molecular evolution.