Related Experiment Video
Updated: Dec 7, 2025
09:34
Scratch Migration Assay and Dorsal Skinfold Chamber for In Vitro and In Vivo Analysis of Wound Healing
Published on: September 26, 2019
13.9K
DeepScratch: Single-cell based topological metrics of scratch wound assays
Avelino Javer1, Jens Rittscher1,2, Heba Z Sailem1,2
1Institute of Biomedical Engineering, Department of Engineering Science, Old Road Campus Research Building, University of Oxford OX3 7DQ, UK.
Computational and Structural Biotechnology Journal
|October 2, 2020
Summary
DeepScratch, a new AI tool, accurately detects cells in scratch wound assays. This method reveals how genetic changes impact cell organization and migration, offering deeper insights than traditional wound area measurements.
Area of Science:
- Cell biology
- Bioinformatics
- Biophysics
Background:
- Tissue architecture changes are implicated in diseases like cancer and cardiovascular conditions.
- Scratch wound assays assess cell migration but struggle with analyzing complex organizational patterns.
- Analyzing imaging data from cell migration assays is a significant challenge.
Purpose of the Study:
- To develop a novel neural network, DeepScratch, for accurate cell detection in scratch assays.
- To introduce topological measures, including local cell density (LCD), for characterizing tissue remodeling.
- To demonstrate the utility of DeepScratch and LCD metrics in analyzing genetic perturbations affecting cell migration.
Main Methods:
- Development of DeepScratch, a neural network for cell detection in heterogeneous marker sets.
- Application of DeepScratch to analyze over 232,000 lymphatic endothelial cell images.
- Implementation of topological measures like cell connectivity and local cell density (LCD).
Main Results:
- DeepScratch accurately detects cells in complex scratch assay images.
- LCD-based metrics successfully classify genetic perturbations (CDH5, CDC42) affecting cell migration.
- Analysis revealed differences in cell migration mechanisms not discernible by wound area alone.
Conclusions:
- DeepScratch enables detailed single-cell analysis in scratch wound healing models.
- Topological measures, particularly LCD, provide a more nuanced understanding of collective cell migration.
- This approach enhances the investigation of genetic factors influencing tissue remodeling and cell behavior.
