The disinhibitory hypothesis, that disinhibition represents a substantial source of general and genetically-based risk for substance use, was strongly supported by these results. First, that behavioral disinhibition shares genetic etiology with substance use disorders is supported by the biometric twin results given in Figure 1, in that the measure of behavioral disinhibition was highly genetically correlated with the other traits. What is more, the genome-wide scores generated on behavioral disinhibition were predictive of all the substance use traits (Table 3), indicating the existence of a polygenetic SNP-based relationship between disinhibition and substance use. Despite statistical significance, the predictive validity of the genome-wide scores was modest, indicating that the ratio of signal to noise is very small for a brute-force genome-wide approach. Clearly more samples are required to have sufficient precision in estimating weights at a genome-wide level. While the GCTA SNP-based estimates account for considerably more—21–45% of the twin-estimated heritable variance—there remains a majority of that heritable variance to be explained. There is much conjecture about the source of remaining additive genetic variance, including non-additive or rare SNP effects, additive
