Skip to main content

Identifying Inflammatory Bowel Disease causative genes through trans-eQTLs mapping within GWAS loci

Final Report Summary - TRANS_CEDAR (Identifying Inflammatory Bowel Disease causative genes through trans-eQTLs mapping within GWAS loci)

The objective of Trans-CEDAR was the identification of new inflammatory bowel disease candidate genes through trans-eQTL mapping in nine different cells and tissue types from healthy individuals. Optimisation of quality control (QC) of clinical, genotype and expression data has been performed. Trans-eQTL mapping has been conducted on a SNP-by-SNP basis using linear regression (additive model) pre correcting for significant cis-eQTL effects. To estimate the degree of trans-eQTL sharing between tissue/cell-types we have selected the probes with FDR ≤ 0.05 in a reference tissue, and have tested (i) what fraction of those would have a trans-eQTL with FDR ≤ 0.05 in the other tissues, and (ii) their π1 proportion in the other tissues. Disease association pattern (DAP) in the UC GWAS-identified risk loci (raw data from IIBDGC plus imputed data) and colocalizing “trans-expression association pattern (trans-EAP)” have been confronted. DAP and “multigene trans-EAP” resemblance has been quantified using Spearman’s rank correlation. Significance of the DAP-EAP correlation (empirical pvalue) has been estimated by comparison to DAP-EAP correlation of in silico eQTLs. A high correlation with a ulcerative colitis (UC) associated locus at the major histocompability complex and EAP affecting four genes (AOAH, ZN672, EXOC1 and DEF8) has been identified. Variants detected as UC causative from fine mapping analysis have been tested as covariates in the regression. Significance of the drop of trans-eQTL signal has been estimated by comparing trans- eQTL pvalue fitting as covariate causative SNPs versus all SNPs in the region. UC causative variants induced an important drop in trans-eQTL signal for AOAH suggesting a link between the UC association and the trans-eQTL effect. AOAH appears as a good UC candidate gene as it encodes a leucocyte enzyme that selectively removes the secondary fatty acyl chains of bacterial lippopolysaccharides (LPS). Altered inactivation of LPS due to AOAH deficiency in colonic dendritic cells impairs mucosal Th17 immunity (Janelsins BM et al, Proc Natl Acad Sci USA 2014). In order to confirm a role for AOAH in UC susceptibility, sequencing of the gene in a large cohort of cases and controls should be performed. RNA levels in UC patients also could help in confirming if this gene deregulation plays a role in UC.

Our eQTL dataset has also been used in assigning functional effects to causal variants detected in a fine mapping project for Crohn's disease and Ulcerative Colitis. We have found a slight enrichment in ciseQTLs effects in CD14, Illeum and rectum for SNPs detected by fine mapping. In particular, fine mapping variants that had significant eQTL effects were characterized in order to identify t the physiopahtological mechanism. This was addressed by looking for functional signatures in silico using publicly available datasets from Epigenome analysis initiatives (Wahsu epigenome browser, Blueprint and ). If the detected variants had epigenomic signatures in blood cell types or intestinal tissues, allellic imbanlance in available Chip seq data was analyzed with VGA tool. We have found an exmple which a fine mapping variant was located in a transcription binding site and this having a functional impact since there was an allelic imbalance in chip seq data in a lymphobalstoid cell line (62%; p=0,001, 100000 permutations)

Since samples come from healthy individuals, our dataset is equally useful for the study of other diseases and phenotypes. So far, CEDAR eQTL dataset utility has been expanded into three projects where the researcher has implemented similar statistical methods to identify SNPs responsible for eQTL effects in different scenarios. In particular, we tested the regulatory function of genome wide association variants for immune mediated disorders summarised in immunobase: We fitted GWAS variants as a covariate in ciseQTLs detected in the association region. In addition to that, we tested cis eQTLs enrichment in enhnacer regions detected through Hi-C. Finnally, cis eQTL analysis was used in order to evaluate the detected variants in the first genome wide study for renal rejection transplantation.