A data-assimilation approach to predict population dynamics during epithelial-mesenchymal transition

Area of Science:

• Computational biology
• Cellular dynamics
• Systems biology

Background:

• Epithelial-mesenchymal transition (EMT) is crucial for development, regeneration, and cancer metastasis.
• Transforming growth factor-β (TGFβ) drives EMT, leading to diverse cellular states.
• Heterogeneity in cell responses to TGFβ presents challenges for computational modeling.

Purpose of the Study:

• To develop a data-assimilation approach for modeling heterogeneous cell populations undergoing EMT.
• To improve the prediction accuracy of computational models for variable cell responses.
• To identify critical model parameters distinguishing different cell sub-populations.

Main Methods:

• Applied a data-assimilation approach combining computational models with parameter estimation.
• Generated synthetic data mimicking heterogeneous epithelial cell populations.
• Performed in silico experiments with time-dependent TGFβ doses and perturbations.

Main Results:

• Identified critical model parameters for EMT-prone and EMT-resistant virtual cell sub-populations.
• Demonstrated that population-specific parameter estimation significantly enhanced prediction accuracy.
• Showcased the successful reconstruction and prediction of EMT dynamics in heterogeneous cell populations.

Conclusions:

• Data assimilation effectively models and predicts EMT dynamics in heterogeneous cell populations.
• Population-specific parameter estimation is key to improving model accuracy for biological variability.
• This approach can account for model errors and parameter uncertainty in complex biological systems.