Genetic and Phenotypic Features of Schizophrenia in the UK Biobank

Sophie E. Legge,Antonio F. Pardiñas,Grace Woolway,Elliott Rees,Alastair G. Cardno,Valentina Escott-Price,Peter Holmans,George Kirov,Michael J. Owen,Michael C. O’Donovan,James T. R. Walters

DOI: https://doi.org/10.1001/jamapsychiatry.2024.0200

2024-03-27

JAMA Psychiatry

Abstract:Importance Large-scale biobanks provide important opportunities for mental health research, but selection biases raise questions regarding the comparability of individuals with those in clinical research settings. Objective To compare the genetic liability to psychiatric disorders in individuals with schizophrenia in the UK Biobank with individuals in the Psychiatric Genomics Consortium (PGC) and to compare genetic liability and phenotypic features with participants recruited from clinical settings. Design, Setting, and Participants This cross-sectional study included participants from the population-based UK Biobank and schizophrenia samples recruited from clinical settings (CLOZUK, CardiffCOGS, Cardiff F-Series, and Cardiff Affected Sib-Pairs). Data were collected between January 1993 and July 2021. Data analysis was conducted between July 2021 and June 2023. Main Outcomes and Measures A genome-wide association study of UK Biobank schizophrenia case-control status was conducted, and the results were compared with those from the PGC via genetic correlations. To test for differences with the clinical samples, polygenic risk scores (PRS) were calculated for schizophrenia, bipolar disorder, depression, and intelligence using PRS-CS. PRS and phenotypic comparisons were conducted using pairwise logistic regressions. The proportions of individuals with copy number variants associated with schizophrenia were compared using Firth logistic regression. Results The sample of 517 375 participants included 1438 UK Biobank participants with schizophrenia (550 [38.2%] female; mean [SD] age, 54.7 [8.3] years), 499 475 UK Biobank controls (271 884 [54.4%] female; mean [SD] age, 56.5 [8.1] years), and 4 schizophrenia research samples (4758 [28.9%] female; mean [SD] age, 38.2 [21.0] years). Liability to schizophrenia in UK Biobank was highly correlated with the latest genome-wide association study from the PGC (genetic correlation, 0.98; SE, 0.18) and showed the expected patterns of correlations with other psychiatric disorders. The schizophrenia PRS explained 6.8% of the variance in liability for schizophrenia case status in UK Biobank. UK Biobank participants with schizophrenia had significantly lower schizophrenia PRS than 3 of the clinically ascertained samples and significantly lower rates of schizophrenia-associated copy number variants than the CLOZUK sample. UK Biobank participants with schizophrenia had higher educational attainment and employment rates than the clinically ascertained schizophrenia samples, lower rates of smoking, and a later age of onset of psychosis. Conclusions and Relevance Individuals with schizophrenia in the UK Biobank, and likely other volunteer-based biobanks, represent those less severely affected. Their inclusion in wider studies should enhance the representation of the full spectrum of illness severity.

psychiatry