Phenotypic and genotypic associations across the psychosis spectrum

Current diagnostic categories for psychosis and affective disorders are at odds with the continuous nature of clinical phenotypes and the shared genetic architecture of these disorders. Phenotypic analyses within and across diagnoses are required to dissect the relationship between genetic risk and phenotypic features of psychosis-spectrum disorders. To examine the relationship between schizoaffective disorder depressive-type (SA-D) and schizophrenia, I investigated phenotypic and polygenic differences between individuals with these diagnoses. SA-D was associated with greater severity of depression and risk factors for depression, including elevated polygenic risk score (PRS) for depression, but not for schizophrenia or bipolar disorder. These findings are consistent with SA-D being a hybrid of schizophrenia and depression. Next, I used factor analysis to derive symptom dimensions in individuals with schizophrenia, schizoaffective disorder, and bipolar disorder and used latent class analysis to cluster individuals into phenotypically-homogenous groups. Symptom domains were able to capture an additional degree of polygenic risk that was not explained by diagnosis. I identified three clusters, characterised by relatively lower functioning, intermediate functioning, and higher functioning. The classes were also able to capture an additional level of polygenic risk not explained by diagnosis. These findings suggest that dimensional or alternative categorical approaches to conceptualising psychosis-spectrum disorders may reflect underlying genetic liability to a greater extent than current systems. Lastly, I examined physical activity as a dimensional phenotype of relevance across disorders. Levels of physical activity differed substantially between individuals with and without psychiatric disorders but were modestly associated with PRS in the general population, indicating that reduced activity in psychiatric disorders may be a consequence of the disorder, rather than due to genetic liability. This thesis identifies novel associations between dimensional phenotypes and PRS for psychiatric disorders and contributes evidence towards refining diagnostic systems in order to improve the validity of psychosis-spectrum disorders