Using singleton densities to detect recent selection in Bos taurus

Abstract

Many quantitative traits are subject to polygenic selection, where several genomic regions undergo small, simultaneous changes in allele frequency that collectively alter a phenotype. The widespread availability of genome data, along with novel statistical techniques, has made it easier to detect these changes. We apply one such method, the Singleton Density Score, to the Holstein breed of Bos taurus to detect recent selection (arising up to around 740 years ago). We identify several genes as candidates for targets of recent selection, including some relating to cell regulation, catabolic processes, neural-cell adhesion and immunity. We do not find strong evidence that three traits that are important to humans - milk protein content, milk fat content, and stature - have been subject to directional selection. Simulations demonstrate that since B. taurus recently experienced a population bottleneck, singletons are depleted so the power of SDS methods are reduced. These results inform on which genes underlie recent genetic change in B. taurus, while providing information on how polygenic selection can be best investigated in future studies. Impact statementMany traits of ecological or economic importance (including height, disease propensity, climatic adaptation) are polygenic. That is, they are affected by a large number of genetic variants, with each one only making a small contribution to a trait, but collectively influence variation. As selection acts on all of these variants simultaneously, it only changes the frequency of each one by a small amount, making it hard to detect such selection from genome data. This situation has changed in recent years, with the proliferation of whole-genome data from many individuals, along with the development of methods to detect the subtle effects of polygenic selection. Here, we use data from 102 genomes from domesticated cattle (Bos taurus) that has experienced intense artificial selection since domestication, and test whether we can detect signatures of recent selection (arising up to 740 years ago). Domesticated species are appealing for this kind of study, as they are subject to extensive genome sequencing studies, and genetic variants can be related to traits under selection. We carried out our analysis in two parts. We first performed a genome-wide scan to find individual genetic regions that show signatures of recent selection. We identify some relating to cell regulation, catabolic processes, neural-cell adhesion and immunity. In the second part, we then analysed genetic regions associated with three key traits: milk protein content, milk fat content, and stature. We tested whether these regions collectively showed a signature of selection, but did not find a significant result in either case. Simulations suggest that the domestication history of cattle affected the power of these methods. We end with a discussion on how to best detect polygenic selection in future studies.