Statistical methods for genetic association studies
Over the past 10 years, massive genetic studies such as UK Biobank have revealed that the genetics of common diseases like diabetes and heart attack are vastly complex. Instead of being driven by one gene or several, these diseases are driven by a combination of thousands of genes and mutations with effects that are very small individually - usually increasing or decreasing risk by much less than 1% - but important in aggregate. This project aims to develop sophisticated statistical and computational methods to learn about the overall distribution of these small effects, and to aggregate them in ways that help us learn disease biology. For example, by modeling the correlations among the all of the genetic variants that are associated with increased disease risk, we aim to identify which of those variants actually cause increased risk, and to improve our ability to predict which individuals are at increased risk of disease. This research will aid in therapeutic development and clinical decision making.