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

Lifetime prevalence of inflammatory bowel disease (IBD) is reaching an alarming rate of >1/400 in industrialized societies. Improved understanding of disease pathogenesis is essential to develop more effective preventive, diagnostic and therapeutic measures. Genome-wide association studies (GWAS) have identified ~ 160 risk loci contributing to inherited predisposition to IBD, leading to the identification of new perturbed pathways and potential drug targets. Nevertheless, causative genes and variants remain unknown for the vast majority of risk loci. GWAS loci are likely to be regulatory and therefore alter expression levels of other genes. We hypothesize that if an IBD associated SNP is an expression quantitative loci (eQTL)-the " disease-association pattern " (DAP) should mirror the " eQTL association pattern " (EAP) of the causative gene if looking in the right target tissue(s). With this premise, our project aims to detect causative genes implicated in IBD's susceptibility through the evaluation of trans-eQTLs within GWAS loci. To this purpose, nine blood cell types and ileal, colonic and rectal biopsies have been collected for 330 healthy individuals of Northern European descent. All individuals have been genotyped with the OmniExpress Illumina array interrogating > 700K genetic variants. Transcriptome analysis has been conducted for all individuals and all cell/tissue types using Illumina HT12 arrays interrogating > 47,000 transcripts. Genotype and transcriptome data have undergone rigorous quality control. Transcriptome data have been pretreated variance stabilizing transformation, QQ normalization and correction for random and fixed effects in each cell type. Only expression probes mapped against Refseq have been considered. Genomic positions have been recovered and probes mapping to more than one genomic position (taking into account splice junctions with Tophat software) with a 96% identity have been discarded. Trans-eQTL mapping will be conducted on a SNP-by-SNP basis using linear regression (additive model) with PLINK software. In order to circumvent genome wide multiple testing penalty, we will test for a given SNP in the genome, any evidence for an excess of low p-values when testing its effect on the expression of genes located on other chromosomes or far away on the same chromosome. Confirmation of putative multigene transregulators will afterwards be performed by RNAseq experiments. We will then quantify the resemblance between DAP in the 160 GWAS-identified risk loci (raw data from IIBDGC plus imputed data) and " multigene trans-EAP " with Spearman's rank correlation. We will also evaluate the biological relevance of this list by performing a network analysis after adding the identified trans targets to the list of previously identified positional candidate genes (mapping to GWAS-identified IBD risk loci). Finally, as a the ultimate proof of causality, the selected genes will be resequenced in 3,000 IBD cases and 3,000 controls, using 600 DNA pools of 10 individuals with Illumina Truseq Amplicon. With this strategy, we expect to detect new causative variants that may constitute new drug targets for IBD. Latest results will be presented.