Bayesian Test for Colocalisation between Pairs of Genetic Association Studies Using Summary Statistics

Abstract

Genetic association studies, in particular the genome-wide association study (GWAS) design, have provided a wealth of novel insights into the aetiology of a wide range of human diseases and traits, in particular cardiovascular diseases and lipid biomarkers. The next challenge consists of understanding the molecular basis of these associations. The integration of multiple association datasets, including gene expression datasets, can contribute to this goal. We have developed a novel statistical methodology to assess whether two association signals are consistent with a shared causal variant. An application is the integration of disease scans with expression quantitative trait locus (eQTL) studies, but any pair of GWAS datasets can be integrated in this framework. We demonstrate the value of the approach by re-analysing a gene expression dataset in 966 liver samples with a published meta-analysis of lipid traits including >100,000 individuals of European ancestry. Combining all lipid biomarkers, our re-analysis supported 26 out of 38 reported colocalisation results with eQTLs and identified 14 new colocalisation results, hence highlighting the value of a formal statistical test. In three cases of reported eQTL-lipid pairs (SYPL2, IFT172, TBKBP1) for which our analysis suggests that the eQTL pattern is not consistent with the lipid association, we identify alternative colocalisation results with SORT1, GCKR, and KPNB1, indicating that these genes are more likely to be causal in these genomic intervals. A key feature of the method is the ability to derive the output statistics from single SNP summary statistics, hence making it possible to perform systematic meta-analysis type comparisons across multiple GWAS datasets (implemented online at http://coloc.cs.ucl.ac.uk/coloc/). Our methodology provides information about candidate causal genes in associated intervals and has direct implications for the understanding of complex diseases as well as the design of drugs to target disease pathways.

Genome‐wide association studies of multiple sclerosis

“…Two distinct approaches have emerged, with overlapping goals: the first is to identify genes with an eQTL driven by an MS risk variant in a locus and the second is to identify specific regulatory elements driving disease risk, and through these, the genes were affected, which must by definition be pathogenic. In attempts to overlap GWAS and eQTL data, the key issue is not just to identify eQTLs in a GWAS locus, but to identify those that appear to be driven by the same underlying genetic variant driving disease risk 59. This has proven a difficult challenge, as a result of linkage disequilibrium60.…”

“…Practically, because eQTL are very common, and many variants show association to disease in a locus, it is likely that at least some variants associated with disease will also have eQTL evidence for a nearby gene 61. Several methods have been proposed to address this colocalisation issue,59, 62 each of which aims to compare GWAS and eQTL data to identify pleiotropic effects between them. Recently, we developed a joint likelihood approach to this problem and used it to compare MS risk associations from the IMSGC ImmunoChip study to eQTL in CD4 + T cells, CD14 + monocytes and lymphoblastoid cell lines 63.…”

Genetic effects on gene expression across human tissues

“…Moreover, the cis -eQTL signal for IRF1 co-localized with the trans -eQTL signals for both trans -eGenes (Fig. 6d; posterior probability >0.99) 43 . Together, these results suggest that cis -regulatory loci affecting IRF1 are regulators of interferon-responsive inflammatory processes involving genes including PSME1 and PARP10 , with implications for complex traits specific to muscle tissue.…”

“…In order to assess the probability that molecular traits as estimated by cis - and trans -eQTLs and physiological traits as estimated by GWAS share the same causal variant, we applied the coloc R package 43 . For each GWAS, we approximated the number of independent loci by extracting variants with at least genome-wide significance ( P <5 ×10 −8 ) and farther than 1 Mb away from all other variants of higher statistical significance.…”

“…However, the mechanism of any cis -effects of these trans -eVariants remains uncertain from the GTEx data; a post hoc analysis demonstrated that PEER correction removed broad regulatory signals from the 9q22 locus, and particularly from cis - and trans -eQTL signals for FOXE1 (Supplementary Information 17). In PEER-corrected data, cis - and trans -eQTL signals co-localized for another cis -eGene in 9q22, TRMO , for both trans -eGenes 43 (posterior probability >0.99). Mendelian randomization analysis of the PEER-corrected data supported the idea that TRMO regulates TMEM253 ( P ≤1.3×10 −9 ) and ARFGEF3 ( P ≤2.1×10 −11 ) based on trans -eVariant rs1867277.…”