Mathematical Modelling of NHU Cell Dynamics and Proliferation Using Interpretable Machine Learning

This paper introduces a novel approach to predicting the number of cells in the Normal Human Urothelial (NHU) environment in untreated monolayer cultures over 48 hours. We employ symbolic regression to develop a mathematical model that delineates the relationship between cell growth over time, migration speed, and angular velocity. This data-driven technique scrutinises the cell growth curve, a concept well-known among biologists. Our methodology utilises advanced computer vision techniques to monitor time-lapse videos, extracting detailed features such as cell growth, mean migration speed, and angular velocity. A significant innovation in our analysis is the application of Genetic Programming for symbolic regression, which enhances our ability to accurately predict cellular behaviours under control conditions. This approach allows the evolved network to precisely identify features crucial for understanding the responses of cells in untreated scenarios, providing unbiased, efficient insights into cell dynamics. This method supports the development of mathematical models for computational simulations, enhancing our understanding of cellular behaviours under normal conditions.