To test the robustness of our effects against potential confounders, we performed supplementary analyses that included all covariate x environment and covariate x genotype interaction terms in addition to the G × E interaction, per the recommendations of Keller (2014). Finally, in view of Perry et al.’s (2013) recent findings that the interaction between positive life events and GABRA2 in predicting alcohol dependence is more pronounced in males compared to females, we also examined whether the anticipated G × E effect was further modified by sex.
