Machine Learning Assisted Classification of Cell Lines and Cell States on Quantitative Phase Images

Area of Science:

• Biotechnology
• Cell Biology
• Machine Learning

Background:

• Accurate cell classification is crucial for biological research and drug development.
• Distinguishing between cell types and states (live, necrosis, apoptosis) is challenging using traditional methods.

Purpose of the Study:

• To develop and validate machine learning (ML) classifiers for distinguishing between different cell types and states.
• To apply the developed ML algorithm for monitoring cellular responses to photodynamic treatment.

Main Methods:

• Implementation and validation of ML classifiers using optical parameters from cell phase images.
• Analysis of HeLa, A549, and 3T3 cell lines in live, necrotic, and apoptotic states.
• Field testing of the ML algorithm to evaluate temporal dynamics after photodynamic therapy.

Main Results:

• The ML classifier achieved approximately 93% accuracy in distinguishing between three cell types.
• The classifier demonstrated about 89% accuracy in differentiating between live, necrotic, and apoptotic states within the same cell line.
• Successful evaluation of temporal dynamics of cell populations after photodynamic treatment at various doses.

Conclusions:

• Machine learning provides a robust and accurate method for cell type and state classification.
• The developed algorithm is effective for monitoring cellular responses to therapeutic interventions like photodynamic treatment.
• This approach has significant potential for applications in cell biology, drug screening, and personalized medicine.