Computational Biology

Training foundational models to reproduce embryogenesis — predicting both cell-level gene expression dynamics and 3D morphogenesis.

A reinforcement-learning framework for adaptive gene panel selection that preserves single-cell structure while minimizing panel size.