Bayesian Hierarchical Models with Spatial Mixture Priors in Genome-Wide Association Studies

• Activity Number: 347
• Type: Contributed
• Date/Time: Tuesday, July 31, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Jie Shen*+ and Hal Stern
• Companies: University of California at Irvine and University of California at Irvine
• Keywords: GWAS ; Bayesian hierarchical models ; dependence ; spatial mixture models

Abstract

Genome-wide analyses are common approaches to identifying genetic risk factors for disease. The standard analysis is based on a hypothesis test for each SNP separately and ignores relationships among SNPs existing in biology. This motivates our work to explore methods that incorporate dependence among markers into probability models. In this study, we develop a Bayesian hierarchical model with a spatial mixture as prior to analyze a range of SNPs simultaneously. The model adopts a spatial Markov process for the allocation of SNPs to normally distributed mixture densities. We also show that our method can be easily extended to accommodate multiple sources of dependence, e.g., integrating both purely physical spatial dependence and gene based dependences among markers into the model. Our method is then illustrated on simulated data and on a real Alzheimer's disease dataset.