Title: Bayesian analysis of GLMMs with nonlocal priors for genome-wide association studies
Authors: Shuangshuang Xu - Virginia Tech (United States)
Jacob Williams - Virginia Tech (United States)
Marco Ferreira - Virginia Tech (United States) [presenting]
Abstract: A novel Bayesian method is presented to find single nucleotide polymorphisms (SNPs) associated with particular phenotypes measured as discrete data from genome-wide association studies (GWAS). This is a regression problem with p two to three orders of magnitude larger than $n$, the subjects are correlated, and the SNPs regressors are highly correlated. To deal with these challenges, we propose nonlocal priors specifically tailored to GLMMs and develop related fast approximate computations for Bayesian model selection. To search through hundreds of thousands of possible SNPs, we use a two-step procedure: first, we screen for candidate SNPs; second, we perform model search that considers all screened candidate SNPs as possible regressors. A simulation study shows favorable performance of our Bayesian method when compared to other methods widely used in the GWAS literature. We illustrate our method with applications to the analysis of real GWAS datasets from plant science and human health.