to detect genetic association (Waldman et al., 1999, Kuo et al., 2010). However, it is of note that the largest of the aforementioned studies (Schumann et al., 2011), a meta-analyses of alcohol consumption GWAS on over 21,000 individuals, did not produce a genome wide significant variant in either DRD2 or ANKK1. The association with DRD2/ANKK1 appears to be contingent upon the specific measure of the phenotype, specific SNPs, and specific population used in a study. This is consistent with the implications of our twin studies that indicate that different genetic factors may contribute to risk for different measures of the “same” outcome (Dick et al., 2011). Moreover, while two measures of alcohol problems can both be valid and widely used, they are not necessarily genetically homogenous.
