Detection of Gene Environment Interaction in High-Volume Genetic Studies

• Activity Number: 110
• Type: Topic Contributed
• Date/Time: Monday, July 30, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): James Gauderman*+ and Pingye Zhang and Juan-Pablo Lewinger
• Companies: University of Southern California and University of Southern California and University of Southern California
• Keywords: environment ; power ; discovery ; gwas ; GxE

Abstract

A common paradigm in genomewide association studies (GWAS) is to scan the genome for SNPs with a direct (main) effect on the trait. Many genes have been identified by this approach, but for most traits there still remains a significant amount of variation left unexplained. It is likely that there are additional SNPs with a modest main effect but a large effect in a subgroup defined by an environmental factor. Analysis of gene-environment (GxE) interaction has the potential to identify these SNPs. It is well known that standard analysis of a case-control sample by logistic regression has poor power to detect a GxE interaction. Substantially greater power can be obtained by a case-only analysis, but at the cost of potential Type I error inflation. We describe an efficient 2-step approach that first screens all available SNPs and then formally tests GxE interaction for only the subset of SNPs that pass the screen. We show that this 2-step approach preserved the Type I error rate while providing substantially greater power to detect interactions in a GWAS compared to a standard case-control analysis, and in many models, greater power than a case-only analysis.