Connecting myelin-related and synaptic dysfunction in schizophrenia with SNP-rich gene expression hubs

Abstract

The recent availability of several genome-wide data sets such as genome-wide mapping of SNP-rich regions and differentially methylated genes in schizophrenic individuals and gene expression data in all brain compartments across the span of human life prompted us to integrate these datasets to gain a better insight into the underlying gene networks driving this enigmatic disease. We summed up the differentially methylated 'expression neighbors' (i.e. genes with positively or negatively correlating expression values) of genes that fall into one of 108 distinct schizophrenia-associated genetic loci with high number of SNPs in schizophrenic patients derived from a large cohort of pooled sequencing experiments. Surprisingly, the number of expression neighbors (with a Pearson correlation of R>=0.8 or R<=-0.7) of the genes falling into the 108 genomic regions were about 35 times higher for the positively correlating genes and 32 times higher for the negatively correlating ones than for the rest of the ~16000 genes outside these loci. While the genes in the 108 loci have relatively little known impact in schizophrenia, using this approach we identified many more known schizophrenia-related important genes with a high degree of connectedness to other genes and high scores of differentially methylated probes for their expression neighbors (such as MBP, MOBP, GRIA1, COMT, SYNGR1, MAP2 and DGCR6), validating our approach. The analysis revealed that the most positively correlating as well as the most negatively correlating genes affect synapse-related genes the most, offering an explanation and a unified view into the root cause of schizophrenia.