Automated gating of flow cytometry data via robust model-based clustering.

Area of Science:

• * Computational Biology
• * Bioinformatics
• * Immunology

Background:

• * Flow cytometry is crucial for health research, enabling rapid, multidimensional cell analysis.
• * Current limitations exist due to a lack of advanced statistical and bioinformatics tools for high-throughput data.
• * Manual gating is subjective, time-consuming, and prone to human error.

Purpose of the Study:

• * To introduce a flexible, model-based clustering approach for flow cytometry data.
• * To address limitations of existing methods by accounting for outliers and non-elliptical clusters.
• * To develop an automated, robust, and reproducible analysis pipeline.

Main Methods:

• * Utilized t-mixture models with Box-Cox transformation for robust clustering.
• * Implemented an Expectation-Maximization (EM) algorithm for parameter estimation and transformation selection.
• * Validated the approach using two public flow cytometry datasets and simulation studies.

Main Results:

• * The proposed method accurately mimics expert manual gating results.
• * Demonstrated superior robustness to model misspecification and accurate cluster number estimation compared to Gaussian mixture models.
• * Achieved more reproducible results in automated flow cytometry data analysis.

Conclusions:

• * The developed clustering methodology offers a flexible and robust solution for flow cytometry data.
• * This approach enhances automated analysis, reducing subjectivity and manual effort.
• * It has the potential to significantly advance cell population identification in biomedical research.