Identification and prioritization of gene sets associated with schizophrenia risk by network analysis

Genome-wide association studies (GWASs) are used to identify genetic variants for association with schizophrenia (SCZ) risk; however, each GWAS can only reveal a small fraction of this association.

This study systematically analyzed multiple GWAS data sets to identify gene subnetwork and pathways associated with SCZ.

We identified gene subnetwork using dmGWAS program by combining SCZ GWASs and a human interaction network, performed gene-set analysis to test the association of gene subnetwork with clinical symptom scores and disease state, meanwhile, conducted spatiotemporal and tissue-specific expression patterns and cell-type-specific analysis of genes in the subnetwork.

We identified a gene subnetwork comprising 48 genes was associated with SCZ, and confirmed gene subnetwork's gene-set association with SCZ (Case vs. Control) in two independent cohorts. Gene prioritization identified CALM1 and TCF4 as hub genes in the subnetwork. Meanwhile, using gene-set analysis, it was determined that the gene subnetwork was also linked to generality symptoms and Positive and Negative Syndrome Scale (PANSS) total score in SCZ. 12 out of 48 genes were higher expression in early prenatal brain. In addition, expressions of CALM1, NCAM1, and TCF4 were dysregulated in cerebral organoids of SCZ patients compared with healthy controls. CALM1 and NCAM1 were mainly expressed on the surface of glutamatergic neurons.

Our findings identified CALM1, NCAM1, and TCF4 as SCZ risk genes and provided important clues to support the neurodevelopmental hypothesis and new therapeutic targets of SCZ.