Causal variants in autoimmune disease: a commentary on a recent published fine-mapping algorithm analysis in genome-wide association studies study
Genome-wide association studies (GWAS), have become the most powerful tool to search the numerous potential risk genetic loci for the susceptibility of many complicated diseases in recent years. However, despite of the more comprehensive analysis of GWAS, there are some limitations for the method. First, it is difficult to identify true causal variants due to the haplotype construction based on the linkage disequilibrium. Besides, most causal variants identified by GWAS were non-coding variants. Although it has been suggested non-coding causal variants may contribute to epigenetic regulation, such as histone acetylation, methylation or DNA methylation, mRNA splicing and the regulation of RNA transcription. The mechanisms of action, and the cellular states and processes in which they function were largely unknown. In a recent study, Farh et al. developed a fine-mapping algorithm to identify candidate causal genetic variants in 21 autoimmune diseases from 39 GWAS studies (1). Through integrated predictions with transcription and cis-regulatory map for several kinds of immune and non-immune cell types, including resting and stimulated CD4+ T cell, regulatory cell, B cell and monocytes, etc., they had provided the unique information about the distributions and features of causal variants in the susceptibility of autoimmune diseases. Accordingly, more than 90% susceptible variants are reside in non-coding and around 60% variants were located in immune-cell transcription factor binding sites (enhancer), which contribute to activating or modulating T or B cell immune response. However, only 10–20% risk variants appear to act directly classical recognizable transcription factor binding sites to regulate gene expression while the 80–90% of non-coding genetic variants functions directly by modifying the non-classical regulatory sequence. In addition, most non-coding risk variants, including those that alter gene expression, affect non-canonical sequence determinants not well-explained by current gene regulatory models.