Enhancing genetic discovery through narrow phenotyping in schizophrenia

Abstract

Background: Schizophrenia varies greatly from person to person, mainly because of its polygenic nature. Consequently, schizophrenia patients form distinct subphenotypes of schizophrenia, with specific symptom patterns and outcomes.

Methods: This study included 4257 adults, with long-term schizophrenia (control - 8955 individuals) who were assessed for schizophrenia with potentially severe outcomes based on following criteria: disability in functional and/or physical domains before the age of 40; severe negative symptoms (present in infancy or shortly after onset); a continuous course of the disease. Additionally, the time of the onset and aggressive/antisocial tendencies were assessed as one the predictors of potentially severe outcomes. A total of 817 participants met at least three of these criteria, i.e., had disruptive schizophrenia. A genome-wide and transcriptome-wide association study was conducted using linear regression and the PrediXcan algorithm. The obtained data were used to develop a polygenic risk model for early risk prediction of schizophrenia with potentially severe outcomes.

Results: Significant associations were found between schizophrenia and variants in CAMTA1, TRHDE, NELFE, and others. The PRS model demonstrated high performance in training, internal and external validation (ROC AUC of 0.9, 0.89, and 0.68, respectively). The functional pathway analysis highlighted pathways involved in ATP metabolism, myeloid cell differentiation, and apoptotic processes.

Conclusion: Subphenotyping schizophrenia may enhance the discovery of genetic factors affecting its development and progression. The GWAS and TWAS findings revealed general mechanisms involved in the development of schizophrenia with potentially severe outcomes, such as synapse regulation, inflammation, and apoptosis.

Keywords: Genetic factors; Genome-wide association study; Narrow phenotyping; Polygenic risk scores; Potentially severe outcomes; Schizophrenia; Transcriptome-wide association study.