A machine learning based model accurately predicts cellular response to electric fields in multiple cell types.

Area of Science:

• Cell Biology
• Biophysics
• Machine Learning

Background:

• Cells naturally respond to electric fields, a phenomenon known as galvanotaxis.
• External electric fields show potential for optimizing biological processes like wound healing.
• Current predictive models for cell migration lack generalizability.

Purpose of the Study:

• To develop a machine learning model for forecasting cell migration directedness.
• To enable precise cell guidance through predictive modeling.
• To explore the application of electric fields in cellular control mechanisms.

Main Methods:

• Trained a machine learning model on time-series galvanotaxis data of mammalian cranial neural crest cells.
• Utilized transfer learning to adapt the model for different cell types (keratocytes, keratinocytes) and conditions with limited data.
• Simulated in silico cell migration under time-varying electric fields and implemented feedback control using a PID controller.

Main Results:

• The machine learning model accurately forecasts cell migration directedness based on electric field stimuli.
• Transfer learning enabled successful model adaptation to diverse cell types and experimental conditions with minimal training data.
• In silico simulations demonstrated the model's capability for controlling cell migration patterns.

Conclusions:

• A data-driven approach provides generalizable predictive models for cell migration.
• The developed model can be instrumental in designing electric field-based cellular control for precision medicine.
• This technology holds promise for applications such as enhanced wound healing.