Even though twin studies suggest thatapproximately 50-60% of variance in cannabis use disorders (abuse and dependence) is attributable to additive genetic influences (Verweij et al., 2010), molecular genetic studies have had limited success identifying variants implicated in the emergence of problematic cannabis involvement. For instance, a genome-wide association study (GWAS) of cannabis dependence, performed in the replication sample used in the current study, failed to find any genome-wide significant (i.e., p < 5E-8) associations (Agrawal et al., 2011). While GWAS is a powerful tool for SNP discovery, it is fairly agnostic with regard to the putative mechanisms underlying the behavior under study, equally weighting, arguably unduly, variants in all gene systems. Those interested in hypothesis-driven data analyses have frequently reverted to candidate gene approaches, which, with rare exception, have yielded null or nonreplicable results (Agrawal & Lynskey, 2009; Flint & Munafò, 2013).
