Protocol-dependent cardiomyocyte states determine disease modelling capacity of human iPSCs

Human induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) are widely used to model cardiovascular disease, yet numerous differentiation protocols generate cardiomyocytes with heterogeneous molecular and functional properties, complicating experimental design. Here we systematically compare sixteen commonly used cardiomyocyte differentiation protocols and characterize their resulting cell states using single-nucleus RNA sequencing, functional phenotyping and computational integration with human genetic data. Despite similar cardiomyocyte yields, protocols produced distinct transcriptional programs, subtype compositions and physiological properties. By integrating protocol-specific gene expression signatures with genome-wide association studies of cardiovascular traits, we identify cardiomyocyte states enriched for genetic architectures underlying specific diseases. These analyses accurately predict protocols most suitable for modelling particular disease contexts, including electrophysiological defects associated with Brugada syndrome and metabolic vulnerability relevant to myocardial infarction. Our results demonstrate that differentiation protocols encode biologically distinct cardiomyocyte states with differential disease relevance and establish a framework for aligning stem-cell differentiation strategies with human complex trait genetics to guide model selection. This approach enables rational design of iPSC-based disease models and highlights how population-scale genetic data can inform experimental systems in stem cell biology.