QFMatch: multidimensional flow and mass cytometry samples alignment.

Area of Science:

• Immunology
• Computational Biology
• Data Science

Background:

• Flow/mass cytometry data analysis requires aligning cell subsets across samples.
• Existing cluster-matching methods are computationally intensive, suffer from the curse of dimensionality, or fail with significant population variations.

Purpose of the Study:

• Introduce QFMatch, a novel quadratic form (QF)-based algorithm for efficient and robust cell subset alignment.
• Address limitations of current methods in handling sample-to-sample variations in cell population patterns.

Main Methods:

• Developed a novel multivariate extension of the quadratic form distance for comparing flow cytometry datasets.
• Implemented the QFMatch algorithm for computationally efficient cluster matching.
• Evaluated QFMatch using sample datasets from immunology studies.

Main Results:

• QFMatch demonstrates computational efficiency.
• The algorithm effectively accommodates significant differences in population locations, including disappearing or appearing populations.
• QFMatch proves effective in aligning cell subsets across diverse sample datasets.

Conclusions:

• QFMatch offers an attractive solution for cell subset alignment in flow/mass cytometry data analysis.
• The QF distance possesses favorable computational and statistical properties for sample comparison tasks.
• QFMatch is well-suited for analyzing complex flow/mass cytometry datasets with inter-sample variability.