Genetic analyses across cardiovascular traits: leveraging genetic correlations to empower locus discovery and prediction in common cardiovascular diseases

Figures

• 

  Fig. 1. Flowchart of the study design. AF, atrial fibrillation and atrial flutter; AUC, area under the curve; BMI, body mass index; CAD, coronary artery disease; DCM, dilated cardiomyopathy; GWAS, genome-wide association studies; HF, heart failure; LDL, low-density lipoprotein; maxFDR, maximum false discovery rate, which calculates the maximum type I error in the analysis; MHIBB, Montreal Heart Institute Biobank; Neff, effective sample size; PGS, polygenic scores; SBP, systolic blood pressure; UKB, United Kingdom Biobank. aNielsen et al.6, bNelson et al.7, cShah et al.4, dAragam et al.21, eLevin et al.22. *Another large AF GWAS with partly non-overlapping samples (Roselli et al.2).

• 

  Fig. 2. Genetic correlations. Genetic correlations of the included 58 cardiovascular traits and diseases using Linkage Disequilibrium Score Regression8 all values in Supplementary Data 2. Non-significant correlations (P > 0.05 after correction for false discovery rate) are set to 0 and shaded in gray. Abbreviations specified in Supplementary Data 1.

• 

  Fig. 3. Linkage Disequilibrium Score Regression genetic correlations and hierarchical clustering for atrial fibrillation, coronary artery disease and heart failure. On the left, barplots depicting genetically correlated traits with atrial fibrillation and flutter [AF], coronary artery disease [CAD], and heart failure [HF] at a significance level PFDR ≤ 0.05. rg and 95% confidence intervals are represented; all rg values appear in Supplementary Data 4–6. On the right, hierarchical clustering of the traits with rg > 0.20 and PFDR ≤ 0.01 with AF, CAD and HF. The Euclidean distance and the complete method were used to obtain three clusters grouping the traits significantly correlated with AF, CAD, and HF. For each of the 3 clusters, the trait with the highest genetic correlation with the trait of interest was selected for multi-trait analysis of GWAS (marked with an asterisk). PFDR, P-value corrected by false discovery rate; rg, genetic correlation. Other abbreviations as per Supplementary Data 1.

• 

  Fig. 4. Miami plots of single-trait and multi-trait summary statistics. Miami plots from the summary statistics of the original studies and from the MTAG results. Panels (A, B, and C) depict Miami plots for atrial fibrillation and flutter [AF], coronary artery disease [CAD], and heart failure [HF], respectively. The plots of Nielsen et al. for AF, Nelson et al. for CAD and Shah et al. for HF, are depicted at the top, and from the disease-specific MTAG results at the bottom. In AF, MTAG increased the number of identified significant loci from 119 to 138, in CAD from 50 to 135, and in HF from 11 to 61. This led to the identification of 19 novel loci not reported in the original summary statistics for AF, 89 for CAD and 52 for HF. Red dashed line shows the genome-wide significance threshold (P = 5 × 10−8).

• 

  Fig. 5. Per SNP Z-score of the novel loci for each of the traits included in the MTAG analysis. For each novel MTAG loci the lead SNP Z-score of the disease risk allele (beta divided standard error) from all the traits included in each MTAG analysis is represented. SNP identifiers along the x-axis represent the locus number (Supplementary Data 9–11) followed by the genomic position (in GRCh37) and the risk allele. Abbreviations: CV, cardiovascular, other abbreviations and references as per Fig. 1 and Supplementary Data 1.

• 

  Fig. 6. Single-trait versus multi-trait polygenic scores. Area under the curve (AUC) from logistic regression model with the disease as an outcome (atrial fibrillation and flutter [AF], coronary artery disease [CAD] and heart failure [HF]) and the disease-specific candidate polygenic score as a predictor. The models including the single-trait polygenic scores (PGSsingleT) are represented in blue and the multi-trait ones (PGSmultiT) in red. On the left, the PGS univariate models are depicted and, on the right, the PGS models adjusted for age at last follow-up, sex and the four first principal components of genetic ancestry are shown. PGS developed by LDpred2, using the United Kingdom Biobank linkage disequilibrium reference panel (N = 362,320 individuals). The validation (N = 2500) and testing datasets (N = 12,677) consisted of unrelated individuals from the Montreal Heart Institute Biobank cohort. AUC shown are those from the testing dataset. For the female cohort, 997 individuals were included in the validation dataset and 5152 in the testing dataset. For the male cohort, 1503 individuals in the validation dataset and 7525 in the testing dataset. Representation of the AUCs corresponding…