Despite these limitations, this study has several strengths. First, the analysis focused on a genetic variant with strong statistical and biological evidence for alcohol-related measures, which addresses common criticisms of GxE studies (Duncan and Keller, 2011; Joober et al., 2007; Risch et al., 2009). Second, focusing on a youth population and employing a longitudinal study design reduced recall bias, enabling more accurate assessment of drinking behaviors during the critical period of adolescence. Third, the robust environment of respondent report of best friends drinking from ages 12–17 coincided with the timing of the primary outcomes under study. This analysis focused on drinking behaviors that are common in adolescence and therefore are more likely to be directly influenced by peer drinking during this period. Finally, studying adolescent drinking milestones facilitated the characterization of the unfolding of genetic and environmental risks across development. Recent studies further support the discovery potential of examining genetic variants during important behavioral transitions in at-risk youth (Belsky et al., 2013; Dick et al., 2013). Future research on alcohol use disorders may benefit from similar hypothesis-driven study designs that examine well-established genes and environments during critical developmental periods.
