Directional genetic relationships between obsessive-compulsive disorder and bipolar disorder and schizophrenia

Abstract Objective Obsessive-compulsive disorder (OCD) frequently co-occurs with bipolar disorder (BD) or schizophrenia (SCZ), and, importantly, can often precede their onset. However, the genetic architecture and directionality underlying these relationships remain unclear. We leveraged large-scale genome-wide association study (GWAS) data to examine shared genetic architecture and directional relationships among OCD, BD and SCZ, and used major depressive disorder (MDD) as a comparator. Methods Using linkage disequilibrium score regression (LDSC), MiXeR, and Generalized Summary-data-based Mendelian Randomization (GSMR) as well as complementary Mendelian randomization approaches, we characterized genetic correlations, polygenic overlap (Dice coefficient), and effect direction concordance (ρ β ) across disorders. Results We observed substantial genetic correlations between OCD and BD (r g =0.37), BD type 2 (BD2) (r g =0.54), and SCZ (r g =0.39), with a large proportion of shared causal variants between OCD and both BD (Dice=0.85) and SCZ (Dice=0.84). MiXeR analyses indicated that OCD and BD2 share a smaller proportion of causal variants (Dice=0.57) but there is a high concordance of effect directions amongst these causal variants (ρ β =0.96), whereas OCD and MDD showed minimal overlap but strong concordance among shared variants (Dice=0.09, ρ β =1). Directional GSMR and complementary TwoSampleMR analyses supported a causal effect of genetic risk to OCD on liability to BD (b=0.20, p=1.5×10□□), SCZ (b=0.52, p=9.5×10□²¹), and MDD (b=0.24, p=1.06×10□□), with little evidence for reverse causal effects. Conclusions Together, these findings indicate that genetic liability to OCD can represent an early component of transdiagnostic psychiatric risk, with implications for understanding and potentially predicting the emergence of broader psychopathology across the life course.