Automatic Phenotyping Using Exhaustive Projection Pursuit

Area of Science:

• * Computational Biology
• * Immunology
• * Data Science

Background:

• * Flow cytometry is crucial for analyzing cell populations and identifying distinct phenotypes.
• * Manual identification of cell phenotypes can be subjective and time-consuming.
• * Automated methods are needed for comprehensive and objective cell population analysis.

Purpose of the Study:

• * To introduce Exhaustive Projection Pursuit (EPP), an automated computational method for cell phenotype identification.
• * To demonstrate EPP's capability in comprehensively identifying all statistically supported cell populations within flow cytometry data.
• * To provide accessible and integrated EPP software for the scientific community.

Main Methods:

• * Exhaustive Projection Pursuit (EPP) algorithm evaluates all two-dimensional projections of flow cytometry data.
• * Identifies statistically significant gating regions to delineate distinct cell phenotypes.
• * Validated using four well-characterized datasets from existing literature.

Main Results:

• * EPP successfully identified known cell phenotypes across diverse datasets.
• * The method provides a comprehensive and automated approach to cell population delineation.
• * EPP is implemented in C++ and integrates with common analysis environments like MATLAB and FlowJo.

Conclusions:

• * Exhaustive Projection Pursuit (EPP) offers an effective automated solution for flow cytometry data analysis.
• * The tool enables objective and comprehensive identification of cell phenotypes.
• * Freely available source code promotes wider adoption and integration in biological research.