Automatic phenotyping using exhaustive projection pursuit

Area of Science:

• * Computational Biology
• * Immunology
• * Data Science

Background:

• * Flow cytometry is a critical technique for analyzing cell populations.
• * Identifying distinct cell phenotypes is a common but challenging objective.
• * Manual gating can be subjective and time-consuming.

Purpose of the Study:

• * To develop an automated, comprehensive method for cell phenotype identification in flow cytometry data.
• * To introduce Exhaustive Projection Pursuit (EPP) as a novel approach for automated gating.
• * To provide accessible and reusable code for the scientific community.

Main Methods:

• * Exhaustive Projection Pursuit (EPP) algorithm was developed for automated phenotype identification.
• * The method systematically evaluates all two-dimensional projections of the data.
• * Statistically significant gating regions are generated to delineate cell populations.

Main Results:

• * EPP successfully identified cell phenotypes across four well-characterized literature datasets.
• * The automated approach provides a comprehensive analysis of cell populations.
• * The method's effectiveness is demonstrated through validation on existing datasets.

Conclusions:

• * Exhaustive Projection Pursuit (EPP) offers an automated and effective solution for flow cytometry data analysis.
• * The tool facilitates objective and comprehensive identification of cell phenotypes.
• * Freely available C++ and MATLAB code promote wider adoption and integration.