Genetic risk effects on psychiatric disorders act in sets

Abstract

Genetic studies of psychiatric disorders have typically assumed that all genetic effects contribute additively to disease liability. However, it is likely that psychiatric disorders have unrecognized subtypes, where synergistic sets of risk variants co-occur within certain cases more than expected under additivity. The existence of synergistic sets induces a structured form of statistical interactions called coordinated epistasis. We test for these interactions in five psychiatric disorders and find evidence for synergistic sets, and by extension, disorder subtypes. We further find that synergistic sets contributing to comorbidities are mostly disorder-specific, despite high genetic correlations between disorders, supporting current diagnostic distinctions between disorders. Finally, we find that genetic risk factors shared across disorders identify a cross-disorder subtype that is likely the result of heritable confounders, rather than disorder-specific etiology. Our results show that genetic risk effects for psychiatric disorders act in sets, implying the existence of subtypes, and re-interpret the importance of shared genetic effects in understanding disease biology and classification.