Next, we further characterized the residual SUDs by estimating the genetic correlations of each of the four univariate (original) SUD GWAS, and the GWAS of the residual SUDs (after removing genetic influences shared with other externalizing conditions in the Externalizing model) with external correlates (Figure 5a and Supplementary Tables 20-23). Two notable patterns emerged. First, residual SUD effects appeared to capture substance specific effects in an expected manner. For example, all four univariate SUDs were significantly genetically correlated with maximum alcoholic beverage (rGs ranged from .30-.76); however, after accounting for variance shared with other externalizing traits, only residual-PAU remained significantly correlated (rG = .41, p < .001). Second, many residual SUD genetic effects retained significant associations with other forms of psychopathology, highlighting the complex associations between SUDs and both internalizing and thought disorder psychopathology. In other words, the residual genetic effects reflect not only substance specific effects like drug metabolism, but also genetic influences that impact that SUD through other mechanisms not shared with externalizing.
