Abstract 10770: Trans-Ancestry Multivariate Genome-Wide Analysis Highlights the Role of Common Genetic Variation in Cardiac Structure, Function, and Heart Failure-Related Traits

Abstract

Introduction: The contribution of common genetic variation to heart failure (HF) risk has not been fully elucidated. Here, we applied multi-ancestry and multivariate methods to summary genetic data from >1 million individuals to improve power to identify common genetic variants, genes, cells, tissues, and circulating proteins/metabolites associated with HF and related cardiac imaging traits. Methods: Trans-ancestry meta-analysis of HF (56,722 cases and 1,133,054 controls) was performed using METAL. Multivariate analysis including GWAS of HF and imaging traits (MRI and echocardiogram) was performed using N-GWAMA. Downstream transcriptome-wide association studies (TWAS; S-PrediXcan), tissue/cell enrichment (LDSC-SEG using RNAseq and snRNAseq of human left ventricle samples from the MAGnet consortium), and Mendelian randomization (MR) analyses were performed. Results: The multi-ancestry HF GWAS identified 15 loci associated with all-cause HF (p < 5 x 10 -8 ). Multivariate analysis identified 48 (16 novel) loci (p < 5 x 10 -8 ), with enrichment for loci near Mendelian cardiomyopathy genes (p < 1 x 10 -4 ). Genetic associations were enriched (FDR < 0.05) for cardiac and musculoskeletal gene expression and chromatin marks. Gene expression (p = 0.007) and splicing events (p = 0.01) were enriched for established cardiomyopathy genes. Branch chain amino acid dehydrogenase ( BCKDHA ) expression was prioritized in TWAS, and MR identified causal associations between circulating branch chain amino acids and cardiac imaging traits: LVEDV MRI (leucine β = -0.137, 95% CI -0.25 to -0.022, p = 0.02; isoleucine β = -0.276, 95% CI -0.38 to -0.17, p = 3 x 10 -7 ) and LVSEV MRI (leucine β = -.131, 95% CI -0.24 to -0.026, p = 0.01; isoleucine β = -0.217, 95% CI -0.33 to -0.11, p = 1 x 10 -4 ). Unbiased proteome-wide MR of 725 circulating proteins identified 18 significant (FDR < 0.05) causal protein-trait associations, including between lipoprotein(a) and HF (OR 1.09 per 1-SD increase in lipoprotein(a), 95% CI 1.06 to 1.11, p = 1.6 x 10 -11 ). Conclusion: These analyses implicate novel common genetic variation in the pathogenesis of HF, highlight Mendelian cardiomyopathy genes in common HF, and identify circulating metabolites and proteins that may represent new treatment targets.