Identification of relevant hub genes for early intervention at gene coexpression modules with altered predicted expression in schizophrenia.

Genetic risk for schizophrenia is due to the joint effect of multiple genes acting mainly at two different processes, prenatal/perinatal neurodevelopment and adolescence/early adulthood synapse maturation. Identification of important genes at the second process is of relevance for early intervention. The aim of this work was to identify gene co-expression modules with altered expression in schizophrenia during adolescence/early adulthood. To this goal, we predicted frontal cortex gene expression in one discovery sample, the largest GWAS of schizophrenia from the Psychiatric Genomics Consortium, using S-prediXcan, and in one target sample, consisting of 625 schizophrenic patients and 819 controls from Spain, using prediXcan. Prediction models were trained on GTEx frontal cortex expression dataset. In parallel, we identified brain co-expression modules from BrainSpan using WGCNA. Then, we estimated polygenic risk scores based on predicted expression (PE-PRS) for each co-expression module in the target sample, based on PE-PRS model from the discovery sample. This analysis led to the identification of a module with mainly adolescence/adulthood expression whose PE-PRS was significantly associated with schizophrenia. The module was significantly enriched in synaptic processes. Several hub genes at this module are drugabble, according to the drug-gene interaction database, and/or involved in synaptic transmission, such as the voltage-gated ion channels SCN2B and KCNAB2, the calcium calmodulin kinases CAMK2A and CAMK1G, or genes involved in synaptic vesicle cycle, such as DNM1, or SYNGR1. Therefore, identification of this module may be the first step in patient stratification based on biology, as well as in drug design and drug repurposing efforts.