Leveraging molecular quantitative trait loci to understand the genetic architecture of diseases and complex traits

Abstract

There is increasing evidence that many risk loci found using genome-wide association studies are molecular quantitative trait loci (QTLs). Here we introduce a new set of functional annotations based on causal posterior probabilities of fine-mapped molecular cis-QTLs, using data from the Genotype-Tissue Expression (GTEx) and BLUEPRINT consortia. We show that these annotations are more strongly enriched for heritability (5.84× for eQTLs; P = 1.19 × 10^-31) across 41 diseases and complex traits than annotations containing all significant molecular QTLs (1.80× for expression (e)QTLs). eQTL annotations obtained by meta-analyzing all GTEx tissues generally performed best, whereas tissue-specific eQTL annotations produced stronger enrichments for blood- and brain-related diseases and traits. eQTL annotations restricted to loss-of-function intolerant genes were even more enriched for heritability (17.06×; P = 1.20 × 10^-35). All molecular QTLs except splicing QTLs remained significantly enriched in joint analysis, indicating that each of these annotations is uniquely informative for disease and complex trait architectures.

Figure 1. S-LDSC and GCTA estimate for TopcisQTL are upward biased in simulations

Panels (a), (b), (c), and (d) illustrate the true enrichment and S-LDSC and GCTA enrichment estimates for AllcisQTL, 95%CredibleSet, TopcisQTL, and MaxCPP annotations, respectively. GCTA is not applicable to continuous annotations (MaxCPP). The Y-axis indicates the mean of enrichment and error bars represent 95% confidence intervals that are computed for 400 simulations. Numerical results are reported in Supplementary Table 2. See Supplementary Table 3 and Supplementary Table 4 for additional simulation scenarios.

Figure 2. Fine-mapped eQTL are strongly enriched for disease/trait heritability

(a) Meta-analysis results across 41 traits of enrichment for whole blood and FE-Meta-Tissue from the GTEx data set conditioning on the baselineLD model. (b) Meta-analysis results across 41 traits of $\mathrm{\tau }\ast$ for whole blood and FE-Meta-Tissue conditioning on the baselineLD model. In each panel, we report results for AllcisQTL, 95%CredibleSet, and MaxCPP. Error bars represent 95% confidence intervals. The % value under each bar indicates the proportion of SNPs in each annotation; for probabilistic annotations (MaxCPP), this is defined as the average value of the annotation. Numerical results are reported in Supplementary Table 7.

Figure 3. Relationship between eQTL sample size and the annotation effect size ( τ∗)

For each tissue, we plot the $\mathrm{\tau }\ast$ of the MaxCPP annotation, meta-analyzed across 41 traits, against the eQTL sample size. Numerical results are reported in Supplementary Table 10. For visualization purposes, we use the following abbreviations: Adipose Visceral Omentum (Adipose-Visceral), Brain Anterior cingulate cortex BA24 (Brain-ACC), Brain Caudate basal ganglia (Brain-CBG), Brain Cerebellar Hemisphere(Brain-CH), Brain Cerebellar Hemisphere (Brain-CH), Brain Frontal Cortex BA9 (Brain-FC), Brain Nucleus accumbens basal ganglia (Brain-NABG), and Brain Putamen basal ganglia (Brain-PBG), Cells EBV transformed lymphocytes (Cells-CETL), Cells Transformed fibroblasts (Cells-TF), Esophagus Gastroesophageal Junction (Esophagus-GJ), Heart Atrial Appendage (Heart-AA), Heart Left Ventricle (Heart-LV), Skin Not Sun Exposed Suprapubic (Skin-NSES), Skin Sun Exposed Lower leg (Skin-SELL), and Small Intestine Terminal Ileum (Small-Intestine).

Figure 4. Tissue-specific fine-mapped eQTL enrichments for blood and brain related traits

Meta-analysis results of (a) enrichment and (b) $\mathrm{\tau }\ast$ of FE-Meta-Tissue and tissue-specific MaxCPP annotations, conditional on each other and the baselineLD model, across six independent autoimmune diseases, five blood cell traits, and eight brain-related traits, respectively. The Y-axis is the meta-analyzed value and error bars represent 95% confidence intervals. The % value under each bar indicates the proportion of SNPs in each annotation, defined as the average value of the annotation. Numerical results are reported in Supplementary Table 13.

Figure 5. Heritability enrichment of fine-mapped eQTL is concentrated in disease-relevant gene sets

Meta-analysis results of (a) enrichment and (b) $\mathrm{\tau }\ast$ of MaxCPP(S) for various gene sets S. We report results conditional on the baselineLD model (dark blue) and results conditional on both the baselineLD model and MaxCPP(All Genes) (light blue), meta-analyzed across 41 traits. As expected, $\mathrm{\tau }\ast$ estimates are reduced by conditioning on MaxCPP(All Genes), but enrichment estimates are not affected. The Y-axis is the meta-analyzed value and error bars represent 95% confidence intervals that are computed over 41 traits. The % value under each bar indicates the proportion of SNPs in each annotation, defined as the average value of the annotation. Numerical results are reported in Supplementary Table 18.

Figure 6. Fine-mapped eQTL, hQTL, sQTL, and meQTL annotations are enriched for disease/trait heritability

Meta-analysis results of (a) enrichment and (b) $\mathrm{\tau }\ast$ of MaxCPP for various molecular QTL from GTEx and BLUEPRINT (BL). We report results conditional on the baselineLD model (dark blue) and results conditional on both the baselineLD model and MaxCPP for all six molecular QTL (orange), meta-analyzed across 41 traits. As expected, $\mathrm{\tau }\ast$ estimates are reduced by conditioning on MaxCPP for all molecular QTL, but enrichment estimates are not affected. The Y-axis is the meta-analyzed value and error bars represent 95% confidence intervals that are computed over 41 traits. The % value under each bar indicates the proportion of SNPs in each annotation, defined as the average value of the annotation. Numerical results are reported in Supplementary Table 28 and Supplementary Table 33.