SDT-PICS

SDT-PICS extracts three-dimensional cell geometries over time from membrane-labeled confocal time-lapse microscopy to enable automated cell tracking and quantitative shape analysis for developmental biology.

Key Features:

• Automated tracking and 3D segmentation: Uses a sphere clustering approach with local thresholding and logical rules for unseeded segmentation of variable cell shapes across developmental stages.
• Biomechanical refinement: Refines initial segmentations via a discrete element method simulation constrained by a biomechanical cell shape model to produce biologically plausible geometries.
• Application to Caenorhabditis elegans: Applied to 7- and 8-cell stage C. elegans embryos to quantify volume, contact area, and shape dynamics over time.
• Open-source code and datasets: Provides Python code for the algorithm and performance measurement tools with datasets on Zenodo (10.5281/zenodo.5108416, 10.5281/zenodo.4540092) and a repository at https://bitbucket.org/pgmsembryogenesis/sdt-pics.

Scientific Applications:

• Embryo development analysis: Enables quantitative analysis of temporal changes in cell geometry for studies of cellular and mechanical processes in embryogenesis.
• Biomechanical modeling: Supplies geometry data and biomechanically constrained segmentations useful for modeling cell shape and tissue mechanics.
• Reproducibility and data sharing: Facilitates reproducible analyses and cross-laboratory comparison via shared Python code and Zenodo-hosted datasets.

Methodology:

Initial segmentation and tracking use a sphere clustering approach with local thresholding and logical rules for unseeded 3D segmentation, followed by refinement using a discrete element method simulation constrained by a biomechanical cell shape model.