Contradictory to the extant twin literature positing 50% heritable variation in cannabis use disorders, the aggregate effects of SNPs on the array captured 21% of genetic variation; however, this estimate was not statistically significant. The lack of significance is primarily due to our sample size. For instance, with cigarette smoking, a sample size of 4181 yielded a heritability estimate of 19% at p=0.024. It is, however, worth noting that similar to other major psychiatric disorders (Lee et al., 2013), common variation on commercial arrays does not capture all the postulated heritability in complex traits. This may be attributable to imperfect linkage disequilibrium between these SNPs and rarer causal variants or due to other factors, such as epistasis, gene-environment interplay and other variation (e.g., copy number variants).
