The Role of Covariate Heterogeneity in Meta-Analysis of Gene-Environment Interactions with Quantitative Traits

• Activity Number: 183
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 10:30 AM to 12:20 PM
• Sponsor: ENAR
• Author(s): Shi Li*+ and Bhramar Mukherjee
• Companies: University of Michigan and University of Michigan
• Keywords: gene-environment interaction ; gene-environment independence ; individual patient data ; meta-analysis ; meta-regression ; power calculation

Abstract

In this paper, we study the effect of environmental covariate heterogeneity on two approaches for fixed-effects meta-analysis: the standard inverse variance weighted meta-analysis and a meta-regression approach. Akin to the results obtained in Simmonds and Higgins (2007), we obtain analytical efficiency/power expressions for both methods under the assumption of gene-environment independence. The relative efficiency/power of the two methods depend on the ratio of within versus between cohort variance of the environmental covariate. Instead of discretely choosing meta-analysis versus meta-regression, we propose to use an adaptive combination of meta-analysis and meta-regression estimates that can be used as a default choice, retaining full efficiency for the interaction parameter using individual patient level data under certain natural assumptions. The adaptive approach bypasses issues with sharing of individual data across studies without sacrificing efficiency. The results are illustrated through meta-analysis of interaction between SNPs on FTO gene and body mass index on high-density lipoprotein cholesterol data from a large consortium of Type 2 diabetes.