Genetic-exposome axis orchestrates cardiac motion patterns

Cardiac motion involves coordinated electrical, contractile, and developmen-tal processes but little is known about how genes and environment influence these patterns. Using four-dimensional models of the heart in 78,113 participants, we applied an unsupervised deep-learning framework to derive latent representations of left-ventricular motion and mapped their genetic, environmental, and clinical correlates. Genome-wide association of spatiotemporal traits identified 39 loci, the majority undetected by conventional measures. Integrative single-cell and functional enrichment analyses localised signals to active cis-regulatory elements, with latent-specific loci revealing a non-myocyte axis and transcriptional regulators of cardiac patterning. Rare cardiomyopathy-associated variants displayed spatially heteroge-neous contraction patterns, and environmental exposures, including smoking and air pollution, also shaped motion profiles. Unsupervised tree-based clustering strat-ified motion into subgroups with differential genetic burden and clinical risk. Our results advance a holistic perspective on cardiac motion revealing how a genetic-exposome axis activates functional and morphogenetic programmes in regulating heart dynamics.