Machine Learning Classifies Ferroptosis and Apoptosis Cell Death Modalities with TfR1 Immunostaining

Area of Science:

• Biomedical research
• Computational biology
• Cellular pathology

Background:

• Accurate determination of cell death mechanisms in patient and animal tissues is crucial but challenging.
• Existing methods for quantifying cell death lack sufficient accuracy and unbiased analysis.

Purpose of the Study:

• To develop a machine learning approach for automated and unbiased classification of cell death.
• To establish reliable biomarkers and analytic frameworks for distinguishing ferroptosis and apoptosis.

Main Methods:

• Collected image sets of HT-1080 fibrosarcoma cells undergoing ferroptosis or apoptosis.
• Utilized anti-transferrin receptor 1 (TfR1) antibody, nuclear, and F-actin staining.
• Extracted cell features using high-content analysis and constructed a classifier with multinomial logistic lasso regression.

Main Results:

• Achieved 93% prediction accuracy for classifying control, ferroptosis, and apoptosis.
• Demonstrated that TfR1, nuclear, and F-actin staining reliably detect apoptotic and ferroptotic cells.
• Showcased the effectiveness of machine learning for unbiased cell death analysis.

Conclusions:

• TfR1 staining combined with nuclear and F-actin staining provides a reliable method for cell death detection.
• Machine learning offers an unbiased framework for analyzing modes of cell death.
• This approach enables accurate quantification of cell death mechanisms in biological samples.