Shared Genetic Architecture Between Mental Disorders and Heart Failure Subtypes: A Comprehensive Multitrait...

Mental disorders and heart failure frequently co-occur, but their shared genetic architecture remains incompletely understood.

Prospective Cox regression analysis in the UK Biobank (UKB) quantified the risk of incident heart failure associated with mental disorders. Linkage disequilibrium score regression estimated genetic correlations, while multitrait genome-wide association and cross-phenotype association analyses identified candidate pleiotropic loci. Statistical fine mapping and gene-level association analyses prioritized candidate variants and genes. Bayesian colocalization assessed evidence for shared candidate variants, and exploratory Gene Ontology biological-process overrepresentation analysis provided descriptive biological context. Single-cell RNA sequencing and genetically informed spatial mapping (gsMap) provided cellular and spatial context for prioritized genes. Geneformer-based in silico perturbation analysis was performed using a cardiac cell atlas of dilated cardiomyopathy to prioritize candidate pleiotropic genes and identify cardiac cell states showing predicted embedding shifts.

Prospective UKB analyses demonstrated that mental disorders, such as posttraumatic stress disorder, depression, and bipolar disorder, were associated with a higher risk of incident heart failure. Multitrait analysis identified 607 pleiotropic variants, with fine-mapping and colocalization converging on prioritized candidate loci, including TMEM18, NKX2-2, and XRN2. Single-cell expression profiling and spatial mapping provided context at the cellular and tissue levels for the prioritized genes. Geneformer-based in silico perturbation prioritized FOXP1 and TMEM106B as leading pleiotropic candidate genes on the basis of their predicted changes in downstream gene embeddings in cardiac cellular contexts.

This study identifies shared genetic architecture between mental disorders and heart failure and prioritizes candidate pleiotropic genes associated with both. Transcriptomic and in silico perturbation analyses provide exploratory cellular, spatial, and cardiac functional context. Functional validation is required before mechanistic or therapeutic interpretation.