Abstract

I will discuss statistical approaches to incorporate prior biologic knowledge in genome-wide association studies (GWAS) and the analysis of rare variants. To improve efficiency of GWAS results, my group has proposed a Bayesian hierarchical quantile regression model to incorporate external information with the aim of improving the ranking of causal SNPs. Simulation results show that the proposed model improves the ranking of causal SNPs if the external information is informative (associated with the causality of a SNP ) and does not decrease the causal SNP ’s ranking if the external information is non-informative. We compare this approach to several alternatives, including a filtering framework, and demonstrate that these approaches can worsen the ranking of causal SNPs if the external information is not informative. As an example, we apply this approach to the Colon Cancer Family Registry (CCFR) GWAS data. Additionally, I will discuss the analysis of rare variants in genetic association studies focusing on the incorporation of prior biologic information. For these analyses, we are interested in two goals: (1) to determine if regional rare variation in aggregate is associated with risk; and (2) conditional upon the region being associated, to identify specific genetic variants within the region that are driving the association. I will present an analytical strategy that uses a Bayesian approach to incorporate model uncertainty in the selection of variants included in the index as well as the direction of the associated effects. The approach allows for inference at both the group and variant-specific levels and has added power over other popular rare variant methods to detect global associations. We have recently extended this approach to integrate external information to help guide the selection of associated variants and regions. I will provide examples of this method applied to a single gene region investigating second primary breast cancer and to multiple regions within the CCFR exome study.