Fig. 4: Incorporation of gene expression traits to facilitate functional interpretation of disease trait correlations.

a Bar plot of the number of eQTL genes (as defined by their genic regions) tested in each LD block. The fill of the bars indicates whether eQTL genes were significantly correlated with at least one disease trait. b Bar plot of the number of eQTL genes that were significantly correlated with at least one disease trait. The fill of the bars indicates whether eQTL genes in local \(r_g\) s were correlated with one or more disease traits. c, d, f Heatmaps of the standardised coefficient for \(r_g\) (rho) for each significant gene expression-disease trait correlation (FDR < 0.05) within LD block (c) 1719, (d) 1273 and (f) 2351. Genes are ordered left to right on the x-axis by the genomic coordinate of their gene start. Panels are labelled by the eQTL dataset from which eQTL genes were derived (either PsychENCODE’s analysis of adult brain tissue from 1387 individuals or the eQTLGen meta-analysis of 31,684 blood samples from 37 cohorts). e Edge diagrams for representative genic regions show the rho for each significant bivariate local \(r_g\) (FDR < 0.05). GWAS and eQTL nodes are indicated by grey and white fill, respectively. Panels are labelled by the gene tested and the eQTL dataset from which eQTL genes were derived. In (c–f) significant negative and positive \(r_g\) s are indicated by blue and red colour, respectively. Coordinates for LD blocks (in the format chromosome:start-end, GRCh37): 681, chr4:90,236,972-91,309,863; 1273, chr8:27,406,512-28,344,176; 1719, chr11:112,755,447-113,889,019; 2281, chr18:52,512,524-53,762,996; 2351, chr19:45,040,933-45,893,307.