Understanding Endothelial Cell Junctions with Deep Learning
Author Information
Author(s): Postma Rudmer J., Fischer Susan E., Bijkerk Roel, van Zonneveld Anton Jan
Primary Institution: Leiden University Medical Center, Leiden, The Netherlands
Hypothesis
Can a deep learning image analysis pipeline effectively stratify VE-Cadherin morphologies in endothelial cells?
Conclusion
The developed image analysis pipeline can robustly stratify VE-Cadherin morphologies, aiding in the assessment of endothelial cell responses to various stimuli.
Supporting Evidence
- The pipeline successfully detected and stratified various VE-Cadherin morphologies.
- The model showed generalizability across independent experiments.
- Statistical analysis confirmed significant differences in morphology distributions between treatments.
Takeaway
This study created a computer program that helps scientists see how the borders of blood vessel cells change when they are exposed to different substances.
Methodology
An image analysis pipeline using deep convolutional neural networks was developed to analyze VE-Cadherin morphologies in endothelial cells exposed to various stimuli.
Limitations
The pipeline is optimized for specific imaging conditions and may not perform well with lower quality images.
Statistical Information
P-Value
p<0.005
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website