DetecDiv, a generalist deep-learning platform for automated cell division tracking and survival analysis

Area of Science:

• Cell biology
• Microscopy
• Bioinformatics

Background:

• Characterizing dynamical biological processes requires analyzing temporal information from microscopy images.
• Limitations in single-cell trajectory analysis hinder large-scale studies of yeast replicative senescence.

Purpose of the Study:

• To develop a high-throughput platform for automated single-cell division tracking.
• To accurately reconstruct cellular replicative lifespans and quantify cellular dynamics.

Main Methods:

• Developed DetecDiv, a microfluidic-based image acquisition platform.
• Integrated deep learning-based software for automated single-cell division tracking.
• Utilized time-series classification and image semantic segmentation for temporal metrics.

Main Results:

• DetecDiv accurately reconstructs cellular replicative lifespans.
• The platform performs consistently across different imaging setups and microfluidic trap designs.
• Demonstrated application in quantifying cellular adaptation and survival under environmental stress.

Conclusions:

• DetecDiv offers an all-in-one toolbox for high-throughput phenotyping.
• Enables detailed analysis of cell cycle, stress response, and replicative lifespan assays.
• Advances the study of dynamical biological processes at the single-cell level.