Fig. 6: cTWAS analysis of IBD and other traits using all GTEx tissues. | Nature Genetics

Fig. 6: cTWAS analysis of IBD and other traits using all GTEx tissues.

From: Adjusting for genetic confounders in transcriptome-wide association studies improves discovery of risk genes of complex traits

Fig. 6

a, Estimated percentages of heritability mediated by eQTLs for IBD using MESC (x axis) and cTWAS (y axis). Each dot represents the result from one tissue. The black dotted line denotes equivalence between the methods, and the blue dashed line denotes the slope relating cTWAS and MESC-mediated heritability estimates. b, The number of cTWAS genes detected at PIP > 0.8 in the top two tissues for IBD, SBP and SCZ. Top tissues per trait are determined by the number of detected genes. c, The number of cTWAS genes detected for IBD at PIP > 0.8 across major tissue groups. A gene was detected in a tissue group if it was detected in any of the tissues in that group. A gene was considered new if it was not a silver standard gene (‘known’) or if it was not the nearest gene to a genome-wide significant locus for IBD (‘nearest’). d, The number of tissues with cTWAS PIP > 0.5 for 56 IBD genes detected at PIP > 0.8 in the ‘blood/immune’ or ‘digestive’ tissue groups. e, Nonredundant GO terms enriched among 56 detected IBD genes in the ‘blood/immune’ or ‘digestive’ tissue groups. These terms were found using the Weight Set Cover method from WebGestalt. Of 56, 29 genes were associated with at least one of the GO terms. f, cTWAS results for IBD at the UBE2W locus using the ‘colon transverse’ eQTL data. Description is the same as the previous locus plots. Note that the P value of UBE2W from TWAS is significant using the less stringent Benjamini–Hochberg procedure of multiple testing corrections.

Back to article page