Myocardial infarction (MI) results in the loss of millions of cardiomyocytes, and the formation of a non-contractile scar, ultimately leading to heart failure. Despite more than a decade of research on cardiac regeneration, clinical study results using standard stem cell approaches have been disappointing and currently the only cure is heart transplantation. Our group has previously demonstrated that the epicardium can be activated during MI, through a process called Epithelial-to-Mesenchymal Transition (EMT). More specifically, epicardium-derived cells undergo morphological changes, migrate, and can replenish lost cardiac tissue and ultimately preserve heart function. This provided proof-of-principle for a novel resident cell-based therapeutic approach for heart attack treatment. The aim of this project is to identify novel small molecules that can activate the resident epicardial cells in the human heart, by promoting EMT, and regenerate the infarcted heart. To achieve this, we are conducting phenotypic small molecule screens, assessing the EMT characteristics of human stem cell-derived epicardial cells, miniaturised in a 384 well-plate format. To evaluate the results of the drug screens we developed an image analysis protocol utilising machine learning, based on software training with CellProfiler, which allows us to automatically calculate the percentage of EMT for each experimental image. We are currently performing a Pilot screen, as the first stage of a drug discovery programme.