In collaboration with the Department of Biology at the University of York and the Department of Computer Science at the University of Sheffield, programmable agent-based computational models have been developed to help interpret how populations of cells arise and repair from individual cell:cell interactions. Comparison of computational simulations against time-lapse videos of equivalent cells in culture has been useful to identify where behaviour is poorly modelled, leading to new hypothesis testing. However, due to the complexity of cell behaviours, it is currently not possible to use non-biased techniques to identify principle differences between in vitro and in virtuo performances. In fact the same also holds true for replicate cultures modulated by drugs. This project will use machine-learning approaches to identify principal features of cell cultures to test the equivalence of agent-based simulations with the aim of providing new tools for drug screening, develop more accurate in virtuo models, and identify and predict cell behaviours in real-time, targeting single cell selection for downstream analysis and characterisation.