Two-stage testing procedures with independent filtering for genome-wide gene-environment interaction

Abstract

Several two-stage multiple testing procedures have been proposed to detect gene-environment interaction in genome-wide association studies. In this article, we elucidate general conditions that are required for validity and power of these procedures, and we propose extensions of two-stage procedures using the case-only estimator of gene-treatment interaction in randomized clinical trials. We develop a unified estimating equation approach to proving asymptotic independence between a filtering statistic and an interaction test statistic in a range of situations, including marginal association and interaction in a generalized linear model with a canonical link. We assess the performance of various two-stage procedures in simulations and in genetic studies from Women's Health Initiative clinical trials.

Keywords: Case-only estimator; Filtering; Gene-treatment interaction; Multiple testing; Pharmacogenetics; Randomization.

Fig. 1

Power comparison of the two-stage procedures for detecting gene-treatment interaction. The four panels represent the scenarios (a) rare disease and quantitative interaction, (b) rare disease and qualitative interaction, (c) common disease and quantitative interaction, and (d) common disease and qualitative interaction. Five testing methods are plotted: the one-stage case-control interaction (solid), the one-stage case-only interaction (short dashes), the two-stage procedure with marginal association filtering and standard interaction testing (long dashes), the two-stage procedure with marginal association filtering and case-only interaction testing (dot-dash), and the two-stage procedure with gene-environment association filtering and standard interaction testing (dots).