Indeed, environmental and developmental moderation of genetic effects are two possible reasons (of many) why genetic association studies have failed thusfar in identifying more variants associated with behavioral traits. Substance use development, for example, shows significant change in structure and heritability during adolescence Vrieze, Hicks, et al. (2012), which suggests a possible limitation of the present study, as we examined middle-aged parents along with their 17-year-old children. If genetic effects are substantially different between these age groups it may minimize the effects observed in the present study. Unfortunately, determining whether a SNP or other genetic variant is moderated by development (or environment) is greatly facilitated by a priori knowledge of SNPs known to be associated with the phenotype (Vrieze, Iacono & McGue, 2012), and there are very few such SNPs known at present. There has been preliminary work in evaluating developmental moderation for height and smoking (Vrieze, McGue, et al. [2011]; Vrieze, McGue, & Iacono [2012]), two phenotypes where there are strong SNP associations found through consortia with large meta-analytic GWAS results. Increased data sharing and the resulting larger samples will provide more hits for future work, which will allow powerful investigation of interaction effects for behavioral traits.
