Using deep reinforcement learning to speed up collective cell migration.

Area of Science:

• Cell Biology
• Computational Biology
• Biophysics

Background:

• Collective cell migration is crucial for biological processes.
• Leader-follower cell coordination impacts migration rate.
• Limited research exists on leader-to-follower stimulus signaling.

Purpose of the Study:

• To investigate the impact of stimulus signals from leader cells on follower cells.
• To develop a simulation platform for collective cell migration.
• To demonstrate the importance of intercellular communication in migration.

Main Methods:

• Agent-based modeling framework for simulating cell behavior.
• Deep reinforcement learning for training cell numbers and control signals.
• Analysis of cell movement using 3D time-lapse microscopy data.

Main Results:

• The number of stimulation signals is proportional to the rate of collective cell movement.
• Deep reinforcement learning effectively trains cell numbers and control signals.
• Experiments with single and multi-follower cells validate findings.

Conclusions:

• Agent-based modeling provides a robust framework for simulating collective cell migration.
• Deep reinforcement learning offers a novel approach to studying biological problems.
• The simulation platform highlights the critical role of leader-follower cell communication.