No document available.
Abstract :
[en] Background/Purpose: Immune-mediated inflammatory disorders (IMIDs) share many genetic risk factors. Pleiotropy may exist at different levels and most of the underlying mechanisms are still to be uncovered. GWAS have identified hundreds of risk loci for IMIDs but causative genes have been identified in only a handful of cases. Recent fine-mapping efforts indicate that only a minority of risk variants are coding. This suggests that most risk variants will be regulatory hence affecting disease risk via eQTL effects.
Methods: To aid in the identification of causative genes for IMIDs, we generated transcriptome information (HT12 arrays) for six blood cell types (CD4, CD8, CD19, CD14, CD15 and platelets) and intestinal biopsies at three anatomical locations (ileum, colon, rectum) for 350 healthy Caucasians. The same individuals were genotyped with SNP arrays interrogating > 700K variants, augmented by imputation from the 1KG project. To detect cis-eQTL we tested variants within 0.5 megabase windows centered on the tested probe. The nominal p-value of the best SNP within a cis-window was Sidak-corrected for the window-specific number of independent tests. The corresponding best, Sidak-corrected p-values for each probe were jointly used to estimate their respective false discovery rate.To identify likely causative genes in GWAS identified risk loci variants and also better understand pleiotropic effects, we (i) developed a method that quantifies the correlation between “disease association pattern” (DAP) and “eQTL association pattern” (EAP) and provides an empirical estimate of its significance, and (ii) evaluated the effect of fitting known risk variants as covariates in the eQTL analysis following Nica et al. (2010). We applied both approaches to celiac disease (CE) and rheumatoid arthritis (RA) and the second one to type one diabetes (T1D), multiple sclerosis (MS), systemic lupus erythematosus (SLE), ankylosing spondylitis (AS) and psoriasis (PSO).
Results: We detected > 16000 significant cis-eQTL, with a degree of sharing between cell types ranging from 38 to 90% highlighting the utility of our multi-tissue panel. GWAS variants were drivers of ciseQTL effects across the different tissues in 399 tests (23.6%), mostly in CD4 cells, and pinpointing 64 new gene-disease associations (3.7%). The number of shared loci and shared eQTL were highly correlated (rho=0.66).RA and SLE showed the highest degree of sharing.
Conclusions: We identified new potential candidate genes for IMIDs and characterized pleiotropic effects through ciseQTL mapping in GWAS loci. These findings could shed a light on IMIDs pathogenesis and co-occurrence. Latest results will be presented.