LogicDynamo
Xing Lab · ongoing · Not yet published.
Interpretable ODE models of how cells change state.
Problem
Dynamical models fitted to single-cell data rarely say which regulators drive a gene, how strongly, or how regulators interact.
My contribution
- Developed the PyTorch model-fitting workflow from single-cell RNA and RNA-velocity data to dynamical-system analysis.
- Designed a generalized additive ODE model that separates single-regulator effects from pairwise interactions.
- Implemented ensemble uncertainty estimation and computational gene perturbations.
Approach
- Readable structure
- Each gene's rate is a sum of saturating dose-response terms with sparse regulation.
Results
- Validated on 2 simulated systems with known dynamics and 5+ real single-cell datasets.