The interaction between rs604300 genotype and CA was significantly associated with cannabis dependence symptoms in SAGE (Table S6; Figure 2; bGxE = -0.75, 95% CI [-1.42, -0.09], ΔR2 = .005, ΔF(1,837) = 4.95, p = 0.026). As in CATS, CA was associated with greater endorsement of DSM-IV cannabis dependence criteria (bE = 0.98, 95% CI [0.72, 1.24], p < .001), but this effect was observed only in G homozygotes (bE = 1.13, 95% CI [0.84, 1.42], p < .001), with no association between CA and cannabis dependence symptoms among carriers of the minor A allele (bE = 0.38, 95% CI [-0.22, 0.98], p = .215). Genotype did not predict CA (bG = -0.02, 95% CI [-0.10, 0.61], p = .640), but, unlike in CATS, there was a significant main effect of genotype on cannabis dependence symptoms, such that A allele carriers had relatively greater symptoms (bG = 0.32, 95% CI [0.02, 0.63], p = .039), though post-hoc tests indicated this effect was significant only among participants reporting no CA (bG = 0.59, 95% CI [0.20, 0.98], p = .003)5.
