Genome-wide association study of mania and psychosis reveals complex relationships between psychiatric disorders and their cardinal symptoms

Recent advances in psychiatric genetics have provided novel insight into the etiology of serious mental illnesses, such as bipolar disorder and schizophrenia. However, few studies have investigated the genetic architecture of... [ view full abstract ]

Recent advances in psychiatric genetics have provided novel insight into the etiology of serious mental illnesses, such as bipolar disorder and schizophrenia. However, few studies have investigated the genetic architecture of specific symptoms of serious mental illness. To address this dearth of research, we conducted a series of genome-wide association studies (GWAS) of continuous indices of mania (N ~ 220k) and psychosis (N ~ 140k) in the general population, constructed with Item Response Theory (IRT) scaling. We found complex patterns of genetic overlap between specific symptoms, psychiatric disorders, and related social, behavioral, and cognitive traits. Most notably, we observed high genetic correlations among many pairs of psychiatric symptoms (r_g ~ .60 to .90), but only moderate genetic correlations between bipolar disorder and schizophrenia and their cardinal symptoms (r_g ~ .30 to .40). To further probe the relationship between continuous symptom variation and psychiatric diagnoses, we applied a new method, Genomic SEM, to explore the multivariate genetic architecture of psychiatric symptoms (mania, psychosis, irritability, and depression) and their corresponding disorders (bipolar disorder, schizophrenia, and major depressive disorder). Finally, we conducted a series of polygenic prediction and bioinformatics analyses to interrogate shared and dissociable genetic risk for various forms of psychopathology. In doing so, we use a variety of methods to investigate sources of heterogeneity within psychiatric genetics.