Significant impact of miRNA–target gene networks on genetics of human complex traits

MicroRNA (miRNA), a small non-coding RNA molecule of approximately 22 nucleotides, regulates degradation and translational repression of a specific gene through its binding to the 3′ UTR of target mRNA.1 MiRNA has essential impacts on the pathogenesis of human complex traits, including cancers, cardiovascular diseases, and autoimmune diseases; thus, can act as a disease biomarker as well as a therapeutic target1,2. To date, approximately 2,000 human miRNAs have been annotated in the miRNA registry (miRBase), targeting and regulating majority of the coding genes3. Recent technological development has enabled the identification of additional functional miRNA4, thereby increasing the impact of miRNA in the field of bioscience. The regulatory effect of miRNA is a heritable genetic trait5. Previous studies investigated the contributions of human genetic polymorphisms to miRNA functions, by surveying single nucleotide polymorphisms (SNPs) that alter miRNA seed or target sites6 or by conducting expression quantitative trait (eQTL) analyses of miRNAs7. These approaches have identified several empirical examples that could link SNPs to human disorders; for example, a synonymous variant in IRGM confers a risk for Crohn’s disease by altering a miR-196 binding site8. However, in comparison with the progress achieved in the field of mRNA epigenomics, the comprehensive landscape regarding the impact of miRNA on genetics of human complex traits has not been fully elucidated. A challenge in miRNA epigenomics is the complexity of miRNA–target gene networks. Given the vast amount of potential combinations of miRNAs and target genes, systematic computational predictions of miRNA–target genes are necessary. However, current target gene prediction algorithms include uncertainty in their accuracy, which is represented by the output of quantitative prediction scores that are inconsistent among algorithms9. Integration of this high-dimensional network information with existing genetic or other epigenetic resources will require novel bioinformatics approaches. Here, we report a novel analytical method to comprehensively evaluate the enrichment of genome-wide association study (GWAS) signals in miRNA–target gene networks (miRNA–target gene enrichment analysis in GWAS; MIGWAS). The application of our method in large-scale GWAS results of human complex traits could provide an empirical and quantitative estimation of the impact of miRNA–target gene networks on the genetics of these human complex traits. Our method also provides a list of miRNA and target gene pairs with excess genetic association signals, which may contribute to the discovery of therapeutic miRNAs and drug target genes.