Automated identification of subpopulations in flow cytometric list mode data using cluster analysis

Area of Science:

• Biotechnology
• Computational Biology
• Immunology

Background:

• Flow cytometry is a crucial technique for analyzing cellular characteristics.
• Automated analysis of complex flow cytometry data, particularly for identifying distinct cell subpopulations, remains a challenge.
• Existing methods may not efficiently handle multi-parameter datasets or provide rapid evaluation.

Purpose of the Study:

• To describe the application of K-means (ISODATA) cluster analysis for flow cytometric data.
• To present results from analyzing fluorescent microspheres and peripheral blood mononuclear cells.
• To introduce a method for visualizing multi-parameter cluster analysis results for rapid subpopulation identification.

Main Methods:

• K-means (ISODATA) cluster analysis was applied to flow cytometry datasets.
• A novel multi-parameter data display method was utilized to visualize clustering results.
• Analysis was performed on mixtures of fluorescent microspheres and human peripheral blood mononuclear cells.

Main Results:

• Successful identification of subpopulations in both microsphere and blood cell samples.
• The multi-parameter display method facilitated rapid and effective evaluation of clustering accuracy.
• Factors influencing automated subpopulation identification were examined and optimized.

Conclusions:

• K-means cluster analysis is a viable tool for automated subpopulation identification in flow cytometry.
• The described visualization technique significantly aids in assessing the success of automated cell identification.
• Methods for optimizing automated identification parameters were established, improving data analysis efficiency.