Single-cell morphodynamics predict cell fate decisions during mucociliary epithelial differentiation.

Area of Science:

• Developmental Biology
• Cell Biology
• Biophysics

Background:

• Cell state transitions, crucial for development, are traditionally studied via gene expression.
• Cellular morphology and behavior shifts during differentiation are less understood, especially in packed epithelial tissues.

Purpose of the Study:

• To develop a quantitative framework for analyzing cell morphodynamics during differentiation.
• To investigate if time-resolved morphological features can predict final cell fate in Xenopus mucociliary epithelium (MCE).

Main Methods:

• Utilized live imaging to track thousands of individual cells in differentiating MCE.
• Extracted dynamic features (shape, position, movement) to create a time-resolved morphodynamic dataset.
• Applied supervised machine learning (gradient-boosted trees, logistic regression) to predict cell fate.

Main Results:

• Integrated time-resolved morphodynamic features robustly predicted final cell type.
• Normalized Z-position, membrane-nucleus offset, and time were key predictive features.
• Cellular movement contributed minimally to fate prediction accuracy.

Conclusions:

• Morphodynamic signatures contain predictive information about cell identity.
• This study provides a framework linking physical cell dynamics to molecular state during differentiation.
• The findings offer new insights into cell fate determination in complex epithelial tissues.