Fig. 5: sQTLs and GWAS.

a Multidimensional scaling-based representation of the semantic dissimilarities between GWAS traits and diseases whose associated variants are enriched among sQTLs with respect to non-sQTLs (two-sided Fisher’s exact test FDR < 0.05). Each GWAS term is represented by a dot, whose size corresponds to the enrichment odds ratio (OR), and its color to the Experimental Factor Ontology (EFO) parent category the term belongs to. GWAS terms that lie close to each other are semantically similar. Eight representative traits with available summary statistics are highlighted. To help visualization, only the labels for the non-redundant, confidently and highly enriched terms are displayed (p-value < 10⁻⁸, lower bound of the 95% confidence interval (CI) for the OR estimate > 1.5, width of the 95% CI for the OR estimate below the median). b Maximum likelihood estimates and 95% CIs (shown as dots and error bars, respectively) for the GWAS association effect size of variants affecting splicing (S), and variants affecting expression, but not splicing (E), for eight traits and diseases. c Quantile-quantile plot of p-values for association with asthma in⁴² for sQTLs (black dots), eQTLs without effects on splicing (blue line and area), and variants with effects neither on expression nor on splicing (black line and gray area). Lines and colored areas represent, respectively, medians and middle 95% observed −log₁₀ p-values across 10,000 random samplings from the corresponding variant set, with the same size as the sQTL set. The identity line is shown in red. d p-values for association with asthma in⁴² (left y-axis) of variants in the region chr17:38,010,000-38,130,000, around the GSDMB gene (highlighted). The larger dots correspond to variants identified as sQTLs for the GSDMB gene in the lung. Linkage disequilibrium patterns (color-coded) and recombination rates are also displayed. The lower panels represent the location of RBP eCLIP peaks, H3K36me3 marked-regions and other GWAS Catalog associations with asthma (shown as arrows). The highlighted variant (rs2305480) is in perfect LD with rs11078928, previously shown to have an impact on GSDMB splicing⁵⁴. Source data for a–d are provided as a Source Data file.