DISCO

DISCO performs automated single-cell segmentation and tracking in images from microfluidic devices by integrating microfluidic trap constraints, budding yeast morphological priors, and temporal growth dynamics to improve long-term single-cell analyses.

Key Features:

• Physical Constraints Utilization: Leverages the physical constraints imposed by microfluidic traps to improve segmentation and tracking accuracy.
• Morphological Constraints: Applies shape-based morphological constraints specific to budding yeast to ensure segmentation aligns with biological morphology.
• Temporal Information Integration: Incorporates temporal data on cell growth and motion to track cellular dynamics across time-lapse experiments.
• Performance Improvement: Demonstrates improved segmentation and tracking performance using manually curated datasets compared to existing software under non-uniform fields of view and multiple cell layers.

Scientific Applications:

• High-content image cytometry: Enables precise single-cell analysis in microfluidic devices for high-content image cytometry over extended imaging periods.
• Yeast cellular dynamics: Supports studies of cellular dynamics, growth patterns, and morphological changes in budding yeast and potentially other cell types.

Methodology:

Integrates microfluidic device features (microfluidic trap design and flow characteristics), shape-based morphological constraints derived from budding yeast, temporal dynamics of cell growth and motion, and evaluates performance using manually curated ground-truth datasets.