Genome-wide association analysis reveals insights into the genetic architecture of right ventricular structure and function

Abstract

Right ventricular (RV) structure and function influence the morbidity and mortality from coronary artery disease (CAD), dilated cardiomyopathy (DCM), pulmonary hypertension and heart failure. Little is known about the genetic basis of RV measurements. Here we perform genome-wide association analyses of four clinically relevant RV phenotypes (RV end-diastolic volume, RV end-systolic volume, RV stroke volume, RV ejection fraction) from cardiovascular magnetic resonance images, using a state-of-the-art deep learning algorithm in 29,506 UK Biobank participants. We identify 25 unique loci associated with at least one RV phenotype at P < 2.27 ×10^-8, 17 of which are validated in a combined meta-analysis (n = 41,830). Several candidate genes overlap with Mendelian cardiomyopathy genes and are involved in cardiac muscle contraction and cellular adhesion. The RV polygenic risk scores (PRSs) are associated with DCM and CAD. The findings substantially advance our understanding of the genetic underpinning of RV measurements.

Fig. 1 |. Flowchart of analysis strategy for RV GWAS analyses.

UKBB, UK Biobank; IVNT, rank-based inverse normal transformation; MAF, minor allele frequency; INFO, imputation quality score; LD, linkage disequilibrium; GCTA, genome-wide complex trait analysis; PheWAS, phenome-wide association study; SIFT, sorting intolerant from tolerant; PolyPhen-2, Polymorphism Phenotyping 2; CADD, combined annotation-dependent depletion; eCaviar, expression quantitative trait loci and GWAS causal variant identification in associated regions tool; MTAG, multitrait analysis of genome-wide association; MAGMA, multimarker analysis of genomic annotation; KO, knockout; Hi-C, long-range chromatic interaction.

Fig. 2 |. Manhattan plots of genomic loci associated with CMR-derived RV phenotypes.

The red line indicates the genome-wide significant threshold at P < 2.27 ×10⁻⁸. The red text indicates the loci identified by MTAG. The blue line indicates the suggestive genome-wide threshold of P=1×10⁻⁵. P values are two sided based on the chi-squared test statistics in BOLT-LMM software.

Fig. 3 |

Pleiotropic associations between the RV GWAS variants and other traits.

Fig. 4 |. Enrichment of genes associated with RV phenotypes in g:Profiler.

Gene sets and pathways results were corrected for multiple testing by the g:Profiler ‘sets counts and sizes’ method at 5% threshold. GO:BP, Gene Ontology Biological Process; GO:CC, Gene Ontology Cellular Component; MF, molecular function; KEGG, Kyoto Encyclopedia of Genes and Genomes pathway. EMG, electromyography; P-adj, adjusted P value.

Fig. 5 |. Phenome-wide association analysis of RV PRSs.

A total of 1,041 phenotypes were evaluated for associations with RV PRSs adjusted for age, sex and the first five genetic principal components by logistic regression. The red line represents the significance threshold after accounting for multiple testing (−log₁₀ of 2.2 × 10⁻⁵). The upright triangles indicate positive correlations, and the inverted triangles indicate negative correlations. NOS, not otherwise specified.