3DeeCellTracker, a deep learning-based pipeline for segmenting and tracking cells in 3D time lapse images.

Area of Science:

• Cell Biology
• Bioimaging
• Computational Biology

Background:

• Accurate cell segmentation and tracking in 3D+T microscopy is crucial for understanding cellular dynamics.
• Existing methods face limitations in handling complex datasets and diverse imaging conditions.

Purpose of the Study:

• To develop and validate a deep learning-based software pipeline, 3DeeCellTracker, for robust cell segmentation and tracking in 3D time-lapse images.
• To demonstrate the versatility and efficiency of 3DeeCellTracker across different biological samples and imaging setups.

Main Methods:

• Development of a deep learning software pipeline (3DeeCellTracker) integrating novel and existing techniques.
• Application of 3DeeCellTracker to segment and track cells in semi-immobilized worm brains, zebrafish hearts, and 3D tumor spheroids.
• Utilizing minimal training data (one volume) and initial correction for efficient model adaptation.

Main Results:

• 3DeeCellTracker successfully segmented and tracked cells in diverse biological samples, including worm brains (~100 cells), zebrafish hearts, and tumor spheroids (~1000 cells).
• Achieved high tracking accuracy (90-100% of cells) across datasets acquired with divergent optical systems.
• Demonstrated comparable or superior performance to existing cell tracking methods.

Conclusions:

• 3DeeCellTracker provides a powerful and versatile solution for analyzing challenging 3D+T cell imaging data.
• The software facilitates the study of dynamic cellular activities previously difficult to investigate.
• This advancement has the potential to accelerate discoveries in cell biology and related fields.