AI-driven multiscale modeling for cell fate reconstruction using single-cell techniques

Understanding cell fate transitions is crucial for discovering the mechanisms that drive phenotypic changes in development, regeneration, and disease. In this talk, we present our approaches that combine multiscale modeling and artificial intelligence (AI) to study these transitions. Multiscale modeling explicitly describes mechanisms such as gene regulation, spatial signals, and cell-cell communications, while AI provides data-driven interface by leveraging single-cell techniques to reveal complex cellular dynamics. We will introduce a deep-learning-based model to reconstruct cellular dynamics from time-course single-cell transcriptomic data and discuss our approach on spatial transcriptomics with a focus on inferring cell-cell communications.