Ultra-rare coding variants in schizophrenia risk genes are associated with longitudinal phenotypic differences across psychotic disorders

Background: The Schizophrenia Exome Meta-Analysis (SCHEMA) Consortium has recently identified 40 genes at a false discovery rate <0.05 in which ultra-rare damaging variants substantially increase schizophrenia (SZ) risk. However, the impact of these variants on clinical trajectories across the psychosis spectrum is not yet known. Methods: We screened 9,821 sequenced participants with psychotic disorders who passed genomic quality control from the SUPER-Finland psychosis study for ultra-rare variants predicted to be damaging in these 40 SZ-associated genes, identifying 163 variant carriers. To assess the impact of these variants on clinical course and functioning across psychotic disorders, we used five decades of nationwide, full-coverage health register data together with interview, questionnaire, and cognitive assessments collected at study enrollment. We compared the clinical trajectories and functional outcomes of these variant carriers with those of non-carriers, including comparisons across different strata of the schizophrenia polygenic risk score (SZ-PGS) and with participants harboring deleterious variants in other genes associated with neurodevelopmental disorders (NDDs). Results: Overall, the variant carriers had a similar distribution of major psychotic disorder diagnoses to that of non-carriers. Despite this diagnostic similarity, variant carriers had earlier hospital admissions than non-carriers for both psychiatric (HR=2.30 [1.48-3.57], p=2.05e-04) and non-psychiatric diagnoses (HR=1.40 [1.10-1.79], p=6.88e-03), indicating increased medical needs early in life. They also showed greater exposure to childhood adversity, lower attainment of adult social milestones, and poorer neuropsychological test performance at study inclusion. Across longitudinal measures, variant carriers showed less favorable clinical and functional trajectories than non-carriers and clustered more closely with individuals with high SZ-PGS (≥90th percentile). Consistent with a liability-threshold model, the variant carriers had lower SZ-PGS than non-carriers among participants with a diagnosis of SZ (β = -0.40, p = 5.32e-05) but higher SZ-PGS than population controls (β = 0.55, p = 9.36e-08), suggesting that the variant carriers developed psychosis at lower levels of common-variant liability. Gene-level analyses identified associations of variants in AKAP11, SETD1A, and XPO7 with lower SZ-PGS and/or poorer neuropsychological test performance, while carriers of damaging variants in SZ-risk genes also associated with bipolar disorder showed high cumulative burdens of affective disorder diagnoses. Conclusion: The disease trajectories of carriers of ultra-rare damaging variants in SZ-risk genes broadly overlapped with those of non-carriers, but carriers showed several differences: greater childhood adversity, earlier hospitalizations, and poorer neuropsychological test performance. These carrier-associated differences were detectable across psychotic disorder diagnoses and suggest a broader phenotypic impact of rare SZ-risk variants than is captured by SZ diagnosis alone, with potential clinical relevance for future risk stratification and genetic counseling.