CardamomOT: A mechanistic optimal transport-based framework for gene regulatory network inference, trajectory reconstruction and generative modeling.

CardamomOT improves gene regulatory network inference and trajectory reconstruction accuracy by integrating exact time labels and mechanistic optimal transport, reducing hyperparameters.

• Why it matters: Accurately inferring gene regulatory networks and cellular trajectories from single-cell data is challenging due to unmeasured protein dynamics and limited temporal information, hindering causal understanding.
• What they did: The authors developed CardamomOT, a method that jointly reconstructs gene regulatory networks and protein trajectories from scRNA-seq data using a mechanistic optimal transport framework, incorporating prior knowledge and exact timing.
• The result: CardamomOT outperforms existing methods in accuracy and robustness across simulated and experimental datasets, enabling realistic data generation and advancing causal cellular process modeling.