Integrative Imaging and Machine Learning for Temporal Tracking of Cell Cycle-Resolved Phenotypic Transitions

Cells are complex high-dimensional dynamical systems with highly coupled components and simultaneous running of multiple cellular programs with a broad range of time scales. An important question is how cells coordinate different programs. Specifically, we use live-cell imaging and multiplex staining approaches to study how cell cycle progression couples to cell phenotypic transitions. We apply machine learning techniques to integrate image data from multiple modalities, enabling non-invasive determination of cell cycle progression. Our method combines image processing, feature extraction, and statistical modeling to track and analyze temporal evolution of cell cycle-resolved cell states.