Related Experiment Video
Updated: Jan 15, 2026
Automation of the Micronucleus Assay Using Imaging Flow Cytometry and Artificial Intelligence
Published on: January 27, 2023
Advanced concept for identifying chemico-biological interactions associated with programmed cell death using a
Igor V Pantic1, Jovana Paunovic Pantic2
1University of Belgrade, Faculty of Medicine, Department of Medical Physiology, Visegradska 26/II, 11129, Belgrade, Serbia; University of Haifa, 199 Abba Hushi Blvd, Mount Carmel, Haifa IL, 3498838, Israel; Ben-Gurion University of the Negev, Faculty of Health Sciences, Department of Physiology and Cell Biology, 84105, Be'er Sheva, Israel.
This study introduces a novel Multi-Scale Attention Residual Convolutional Neural Network (MSA-RCNN) for early apoptosis detection using nuclear texture. The AI model analyzes chromatin patterns for improved accuracy in cell biology research.
Area of Science:
- Cell Biology
- Biotechnology
- Computational Biology
Background:
- Early detection of apoptosis is crucial but challenging using traditional microscopy and machine learning.
- Conventional models struggle with complex spatial relationships in nuclear architecture.
- Subtle nuclear texture changes in stained micrographs are key indicators of early apoptosis.
Purpose of the Study:
- To introduce a novel Multi-Scale Attention Residual Convolutional Neural Network (MSA-RCNN) for enhanced apoptosis detection.
- To leverage nuclear chromatin patterns and texture features for identifying early apoptotic changes.
- To explore the potential of AI in improving apoptosis detection for research and clinical applications.
Main Methods:
- Developed a novel Multi-Scale Attention Residual Convolutional Neural Network (MSA-RCNN).
- Utilized quantifiers from Gray-Level Entropy Matrix (GLEM), Run-Length Matrix (RLM), and Discrete Fourier Transform (DFT) as input features.
- Focused on specific parameters: RLM Short Run Emphasis, RLM Long Run Emphasis, GLEM Entropy, and DFT Magnitude Spectrum Mean.
Main Results:
- The MSA-RCNN learns discriminative features from nuclear chromatin patterns indicative of early apoptosis.
- The model effectively utilizes GLEM, RLM, and DFT quantifiers for apoptosis detection.
- Demonstrated a novel approach for AI-based sensing systems in cell biology.
Conclusions:
- The proposed MSA-RCNN shows promise for accurate early apoptosis detection.
- This AI approach offers advantages over traditional machine learning methods in capturing complex nuclear features.
- Future work will address model interpretability and generalization through explainability analyses and diverse dataset validation.
