A Stain-Free Apoptosis Detection and Classification Method Based on Machine Learning Technique
Feng Jingwen, Sa Yu*
Cell apoptosis detection and classification are very important in biological and medical studies. In this study, we established an apoptosis detection and classification method based on the polarization diffraction imaging flow cytometry system and machine learning techniques, which has higher time efficiency and applicability comparing with the previous result. K562 and HL60 cells were induced to undergo apoptosis. The cells were sorted into three subpopulations (viable, early apoptotic and late apoptotic/necrotic cells) using fluorescence-activated cell sorter in combination with double fluorescent stain after the apoptosis induction, and then measured by polarization diffraction imaging flow cytometry for diffraction image acquisition. Different classification algorithms were explored. The model performance and efficiency were analyzed to obtain a high-efficiency model. The new method can achieve a high accuracy of 90% and has higher time efficiency. A fast stain-free apoptosis detection method was developed. Cells after measurement and classification can be directly used in further experimental studies.
CN
EN