Among the most significant challenges of GxE research is the lack of robust genetic main associations (Duncan & Keller, 2011); most studies in the literature focused on genetic variants whose relationships to alcohol phenotypes were inconsistent. Additionally, power to detect interactions is typically lower than power to detect main effects. Since continuous phenotypes offer more information than binary phenotypes by providing a range of values, and typically have more statistical power to detect genetic associations (Kuo et al., 2008; Waldman, Robinson, & Rowe, 1999), evaluating alcohol consumption and AUD with graded variables should increase power for detecting interactions, and is consistent with a general movement in psychiatry to address dimensional rather than binary traits (Ehlke, Hagman, & Cohn, 2012; Hasin et al., 2012). Studying GxE interaction using robust genetic and environmental risk factors and informative, graded phenotypes therefore offers a promising strategy.
