Analysis of Interaction Between Genetic Pathway and Environment Using Kernel Machine Varying Coefficient Model

• Activity Number: 29
• Type: Contributed
• Date/Time: Sunday, July 29, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Xiaoshan Li*+
• Companies: North Carolina State University
• Keywords: Gene-environment interaction ; semiparametric regression ; kernel machines

Abstract

Gene-environment interaction (GxE) is the phenotypic effect of interactions between genes and the environment. It has become one big interest to genetic statisticians. In this work, we propose a semiparametric regression model that relates an outcome of interest to a genetic pathway and environmental factor using a varying coefficient model, where the pathway effect of multiple genetic covariates is modeled nonparametrically using kernel machines. This flexible modeling allows us to quantify GxE interaction effect while also accounting for complex and possibly nonlinear gene-gene interactions within the pathway. We will develop estimation and testing procedures for main effects of environmental factors as well as GxE effect and investigate properties of our estimates via theory and simulation. We will compare the testing procedure with the classical F test in terms of type I error and power. We will apply our methodology to the Normative Aging Study data to analyze the interaction between DNA methylation and environmental exposures to heavy metal and air particles in relation to various cardiovascular outcomes.