From Gene Expression To Disease Association

Abstract

Overall Abstract Association studies have achieved significant progress in mapping hundreds of genetic variants to major psychiatric disorders. But the underlying molecular mechanisms regarding specific genes, pathways involved, and interactions among the genes remain largely unknown. Gene expression and its regulation systems are critical for unraveling the causal relationships. This symposium presents cutting edge studies of brain gene expression regulation and its relevance to disease associations. Hermona Soreq from the Hebrew University of Jerusalem will present a study of microRNA (miR) and pseudogene expression and their roles in psychiatry. A selected group of non-coding pseudogenes (PSGs) carry micro RNA (miR) recognition elements (PSG+MRE) compete with brain-expressed genes for the available miR regulators. Furthermore, SNPs surrounding these coding genes tend to be more abundant in psychiatric patients compared to controls. They used transfection-mediated over-expression of PSG+MRE and GapmeR-directed suppression of selected cholinergic transcripts sharing MREs with them in cultured human cells, and demonstrated bi-directional modulation of the interaction of these transcripts with miRs. The study illustrated functional roles of PSGs+MRE in cholinergic functioning, with potential impact on mental disorders. Seth Ament from the Institute for Systems Biology will present a study of transcription factor (TF) binding targets in the brain. They reconstructed a model for the genomic binding sites and functional target genes of 778 TFs in the human brain by integrating large-scale epigenomics and transcriptomics datasets. Using this model, they predicted master regulator transcription factors and functional non-coding variants associated with risk for bipolar disorder and schizophrenia. They experimentally validated the prediction that a master regulator transcription factor, POU3F2, interacts with a risk-associated SNP to modulate the VRK2 promoter. Chunyu Liu from the University of Illinois at Chicago will present a study of co-expression networks involving noncoding RNAs. They retrieved lncRNAs mapped to the seven major CNV deletion regions known to increase risks of developing schizophrenia and carried out weighted gene co-expression network analysis (WGCNA) using data from Genotype-Tissue Expression (GTEx) and BrainSpan projects to look for co-expression modules that harbor CNV-lncRNAs. They identified coexpression modules that were associated with male reproduction in male individuals and associated with neuronal functions in both male and female individuals. The findings suggest that lncRNAs inside those rare CNVs might play significant temporal and spatial roles in regulating other protein-coding genes and subsequently contribute to schizophrenia risk. Hae Kyung Im from the University of Chicago will present the PrediXcan method, which was developed by him, Nancy Cox, and colleagues. By using genotype to predict expression (or other molecular traits) and correlating them with the trait of interest. They have developed prediction models for gene expression in 40 human tissues using the GTEx and Depression Genes Network data. They also extended the method so that only summary statistics are needed to infer PrediXcan results so that they can perform large meta analysis. Using this new method called MetaXcan they have generated results for 117 phenotypes with publicly available GWAS meta-analysis results.