Identification of Rare Variants in Schizophrenia-relevant Pathways: A Computational Analysis

Schizophrenia (SCZ) is a complex neuropsychiatric disorder with a strong genetic component. A comprehensive understanding of the genes and pathways involved remains elusive. The aim of this study was to identify rare genetic variants contributing to SCZ by leveraging a gene set with higher expression in brain regions relevant to SCZ and focusing on rare damaging variants identified in patients to refine the search for SCZ-associated genes. Using RNA-Sequencing data from healthy donors obtained from the Genotype-Tissue Expression (GTEx) project, an interregional differential expression analysis was performed between two target regions (frontal cortex and hippocampus) robustly implicated in schizophrenia (SCZ) and two contrast regions (cerebellar hemisphere and substantia nigra) with less compelling evidence of involvement in SCZ based on existing literature. For each of the four regions, RNA-Seq data from 25 donors were analysed, totalling 100 datasets. This comprehensive approach identified 1076 genes with more specific expression in the frontal cortex and hippocampus compared to control regions. Variant analysis was performed using whole-exome sequencing data from 3500 SCZ and 4800 control exomes from the Swedish-Schizophrenia Exomes project. Ultra-rare damaging variants (dURVs) were identified by excluding variants found in 125,748 exomes in gnomAD v2.1.1, and filtering to loss-of-function, or variants predicted to be damaging by five in silico prediction tools. Focussing exclusively on the identified 1076 highly expressed genes, 4,398 dURVs were identified exclusive to 2,444 SCZ cases and 1,066 SCZ cases had no dURVs. These findings implicate an oligogenic model of SCZ at the level of the individual, while highly polygenic at the level of the population. Variant aggregation and pathway analysis found pathway heterogeneity between SCZ cases involving any of signal transduction, the complement system, histone modification, expression regulation, and WNT, RHO GTPase, or calcium signalling. The most common gene with dURVs was CSMD1 found in 1.2% of cases.