Genetic risk prediction and variance in diverse populations
Principal Investigator: Professor Iain Mathieson
Approved Research ID: 33923
Approval date: April 6th 2018
Most genetic studies are carried out in populations of European ancestry. We aim to to develop statistical methods that allow these findings to be transferred to populations with different ancestries. These methods are not specific to any particular trait and, indeed, part of the analysis will involve comparing the results for different traits in Biobank. We also aim to test the hypothesis that differences in ancestry affect phenotypic variance, through either i) increasing genetic variance ii) local epistasis, and iii) dominance. This research aims to improve genetic risk prediction of complex traits and disease, and therefore directly impact public health. Further, understanding the genetic basis of complex traits has an indirect impact, by extending our knowledge of the biological basis of both disease and non-disease states. We will infer the genetic factors affecting complex traits (for example, height, BMI) using the European-ancestry individuals in the dataset. Then, we will develop statistical methods to predict those phenotypes in the non- European ancestry individuals in the dataset. In particular, these methods will take into account local ancestry - that is, differences in ancestry along the genome of each individual. Finally, we will test whether differences in genetic ancestry affect phenotypic variance and test for factors that might moderate those effects. Full cohort.