Detecting Signatures of Selection within the Dog Genome

Deciphering the genetic basis of phenotypic diversity is one of the central aims of biological research. Domestic animals provide a unique opportunity for making substantial progress towards this goal. Intense positive selection has lead to a rich reservoir of phenotypes and underlying genotypes that can be interrogated using genetic tools to gain insight into the genetic basis of phenotypic diversity. The dog is the most phenotypically diverse mammal. It was domesticated from the grey wolf 11-30,000 years ago. After domestication, a period of intense breeding has lead to the massive phenotypic diversity seen amongst dog breeds today. These two phases of strong positive selection at domestication and at breed creation are likely to have left their signature on the genome. In this thesis, we have analysed genome-wide patterns to detect genomic regions involved in selection in both of these phases. We used whole genome sequences from 60 dogs and 12 wolves, to detect dog domestication selective sweeps. We find evidence for genes involved in memory formation, neurotransmission and starch digestion. To decipher the genetic signals underlying breed diversity, we used genome-wide genotype data from >170,000 SNPs in 509 dogs from 46 different breeds. We find evidence for genes under selection in many breeds, and only a few breeds. In addition, we identify novel sweeps underlying morphology and behavior. Recombination can influence the configuration of alleles present on a haplotype, and can thus increase or decrease the efficiency of selection. The PRDM9 protein has been shown to be important for determining recombination hotspot locations in humans and other mammals, but of all the mammals studied so far the dog is the only one to have a non-functional PRDM9. We used the genome-wide genotype data described above to characterise the fine scale recombination map in dogs. We find that recombination hotspots exist in dogs despite the absence of PRDM9. Moreover, we show that these hotspots are enriched for GC rich peaks and that these peaks are getting stronger over time. Our results show that the absence of PRDM9 has lead to the stabilisation of the recombination landscape in dogs.