Varying Index Coefficient Model for Dynamic Gene-Environment Interactions

• Activity Number: 431
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Jingyi Zhang* and Xu Liu and Yuehua Cui
• Companies: Michigan State University and Michigan State University and Michigan State University
• Keywords: Single-index model ; P-splines ; Longitudinal data ; Varying coefficient model

Abstract

Gene-environment interactions play key roles in human complex diseases. Existing literature has shown the power of integrative gene-environment interaction analysis by considering the joint effect of environmental mixtures. In this work, we propose a varying index coefficient model for multiple longitudinal measurements of environmental variables and assess how the genetic effects on a disease trait are nonlinearly modified by a mixture of environmental influences. We derive an estimation procedure for the nonparametric varying index coefficients based on the quadratic inference functions and penalized splines. Theoretical results such as consistency and asymptotic normality of the estimates are established. We also evaluate the performance of our estimation procedure through simulation studies. The proposed method is illustrated by applying to a real data set in which SNP effects are nonlinearly modulated by the combination of hormone mixtures to affect women's eating disorder.