Related Experiment Video
Updated: Oct 23, 2025
08:44
Quantitative Cell Biology of Neurodegeneration in Drosophila Through Unbiased Analysis of Fluorescently Tagged Proteins Using ImageJ
Published on: August 3, 2018
10.2K
Deep Learning Approach for Quantification of Fluorescently Labeled Blood Cells in Danio rerio (Zebrafish)
Samrat Thapa1, David L Stachura1
1Department of Biological Sciences, California State University, Chico, Chico, CA, USA.
Bioinformatics and Biology Insights
|August 20, 2021
Summary
We developed a machine learning algorithm to automatically count neutrophils in zebrafish, significantly reducing time and subjectivity compared to manual methods. This deep learning model offers a faster, more accessible approach for analyzing immune cells in this model organism.
Area of Science:
- Immunology
- Computational Biology
- Zebrafish Model Systems
Background:
- Neutrophils are crucial for innate immunity.
- Zebrafish (Danio rerio) are used as vertebrate models for neutrophil studies.
- Transgenic zebrafish allow in vivo imaging of fluorescently labeled neutrophils.
Purpose of the Study:
- To automate the identification and counting of fluorescently labeled neutrophils in zebrafish.
- To overcome the laborious and subjective nature of manual neutrophil counting.
- To introduce a deep learning approach for high-throughput analysis of immune cells.
Main Methods:
- Development of a custom-trained "you only look once" (YOLO) machine learning algorithm.
- Application of the YOLO algorithm to identify and count fluorescently labeled neutrophils in zebrafish.
- Comparison of automated counts with manual counts performed by human observers.
Main Results:
- The YOLO model achieved a correlation coefficient of r = 0.8207 with human counting.
- The model exhibited an 8.65% error rate, comparable to the 5%-12% variation among human counters.
- The algorithm successfully validated results from a previous study using manual quantification.
- Automated counting was completed in minutes, versus hours for manual methods.
Conclusions:
- A custom-trained YOLO algorithm effectively automates neutrophil identification and counting in zebrafish.
- This deep learning model provides a rapid, objective, and accessible method for neutrophil quantification.
- The approach supports the use of deep learning for high-throughput analysis in zebrafish immunology research.
