PyUAT: An open-source Python framework for uncertainty-aware, efficient, and scalable model-driven cell tracking

Area of Science:

• Microbiology
• Cell Biology
• Bioimaging

Background:

• Live-cell imaging is crucial for understanding cell behavior and responses.
• Microbial cell tracking faces challenges due to cell movement, division, and limited imaging frame rates.

Purpose of the Study:

• To investigate the effectiveness of Uncertainty-Aware Tracking (UAT) for improving microbial cell tracking.
• To analyze the impact of cell development models on tracking quality under varying imaging intervals.

Main Methods:

• Utilized PyUAT, an open-source implementation of the UAT paradigm.
• Systematically analyzed tracking quality using statistical models of cell behavior on a 2D+t dataset.
• Evaluated performance under increasing imaging intervals.

Main Results:

• Model-driven cell tracking (UAT) demonstrated higher accuracy at low frame rates.
• UAT outperformed comparable methods in runtime efficiency.
• The choice of cell development models significantly influences tracking quality.

Conclusions:

• Probabilistic UAT robustifies microbial cell tracking quality under challenging live-imaging conditions.
• PyUAT offers an effective and efficient solution for microbial cell tracking, even with limited frame rates.
• The study provides a valuable tool for researchers studying cellular heterogeneity and responses.