Fig. 4

Human-inserted and mouse-deleted promoters are enriched for eQTL across tissues. a Odds ratios (log₂) showing eQTL depletion within functional-turnover relative to conserved promoters (x=0, green line). Expression QTL for n=44 tissues (y-axis) from GTEx consortium, rank ordered by the total number of promoter-associated eQTL identified in the tissue. Symbols (tissue-type indicated in key) show the odds-ratio point estimate (positive values are enrichment) and horizontal lines the 95% confidence interval from Fisher’s exact test. b As for (a) showing human-inserted promoter eQTL enrichment. c As for (a, b) showing mouse-deleted promoter eQTL enrichment. d Heatmap of beta-coefficients from multivariate regression considering evolutionary history and additional promoter and gene features (x-axis). Tissues (y-axis) ordered as in a–c. Red denotes eQTL enrichment relative to conserved promoters and grey depletion. Non-significant (p>0.05) associations have beta rounded to zero (displayed as white). e Spatial enrichment of eQTL across promoter regions, normalised to the 2–4 kb flank upstream from the TSS (x-axis, arrow). f Distribution of derived beta coefficients for all significant prostate eQTL overlapping the indicated promoter class. * indicates a Bonferroni-corrected p value < 0.05 for Mann-Whitney tests comparing evolutionarily volatile promoters to those with conserved expression. g Consensus across tissues, for the direction of change in expression for the derived allele. Boxplots show the distribution of median derived beta-coefficients over n=44 GTEx tissue types (Additional file 1: Fig. S14 for individual tissue analyses). ** and *** indicates Bonferroni-corrected p value < 0.01 and < 0.001, respectively, for two-tailed Mann-Whitney tests between conserved promoters and the indicated evolutionarily volatile promoter class