Learning collective cell migratory dynamics from a static snapshot with graph neural networks

Area of Science:

• Computational Biology
• Cellular Dynamics
• Bioinformatics

Background:

• Multicellular self-assembly is crucial for development and disease, involving complex cell migration.
• Understanding collective cell migration dynamics from static configurations is challenging.
• Existing methods for inferring cell motion rely heavily on physical intuition.

Purpose of the Study:

• To develop a computational method for inferring multicellular migratory dynamics from static cell configurations.
• To overcome the limitations of traditional, intuition-based approaches in identifying motion-indicative structural features.
• To enable better prediction and understanding of dynamic multicellular processes.

Main Methods:

• Utilized a graph neural network (GNN) model.
• Trained the GNN on both experimental and synthetic datasets of cell positions.
• Focused on inferring collective cell motion from static snapshots.

Main Results:

• Successfully inferred multicellular collective motion from static cell position data.
• Demonstrated the GNN's effectiveness on diverse datasets.
• Provided a novel computational approach to analyze cell migration.

Conclusions:

• Graph neural networks offer a powerful tool for analyzing collective cell migration dynamics.
• This method can predict cellular movement from static snapshots, aiding research in development and disease.
• The GNN approach provides a data-driven alternative to traditional methods for studying multicellular self-assembly.