The aim of the study was to explore the contribution of GVs to initiation of cannabis use and age at onset. Using GCTA and a sample of distantly related individuals from the NTR, we estimated that the genomewide SNPs collectively explain 25 % (SE = 0.088; P = 0.0016) of the variance in initiation. Although lower than the twin-based heritability estimate (i.e., of about 44 % (95 % CI [16 %,74 %], Vink et al. 2010), our estimate provides an indication of the total signal in the currently measured (and tagged) SNPs, confirming that initiation of cannabis use is a heritable trait. The remaining variance (up to 44 %) may, in part, be attributable to rare variants, weakly correlated with the measured SNPs (Visscher et al. 2010). Our estimate is larger than that reported by Verweij and colleagues, namely 6 % (95 % CI [0 %, 26 %], P-value = ns). A possible reason for this difference is that we use more densely distributed SNPs. In addition to the common SNPs overlapping with the HapMap SNPs used by Verweij and
