Semiparametric Mixed-Model Analysis for Nonlinear Gene-Environment Interactions in Genome-Wide Association Studies

• Activity Number: 415
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Zijian Huang*+ and Shujie Ma
• Companies: and University of California, Riverside
• Keywords: G*E interactions ; semiparametric mixed models ; B-splines ; profile estimation ; score test ; variance component test

Abstract

The use of linear mixed models (LMMs) in genome-wide association studies (GWAS) is widely accepted because LMMs have been shown to be capable of correcting for several forms of confounding due to genetic relatedness. On the other hand, gene and environment (G×E) interactions play a pivotal role in determining the risk of human diseases. In this paper, we propose a semiparametric mixed model to capture possible nonlinear G×E interactions in GWAS. To check whether random effect should be included for each SNP, we develop linear score test for variance component. It turns out that for some SNPs, mixed model should be considered, then profile quasi restricted maximum likelihood estimation method is applied. While for other SNPs, phenotype can be treated independently, then apply profile quasi-log-likelihood estimation method. For these profile estimators, asymptotic consistency and normality are established. Moreover, we develop Rao-score-type test procedures based on the profile estimation for regression parameters and nonparametric coefficient functions.Both simulation studies and an empirical example are presented to illustrate the use of our proposed models and methods.