flowFP

flowFP generates empirical fingerprints of multivariate flow cytometry probability distributions to enable statistical modeling and automated classification of cytometry data.

Key Features:

• Empirical fingerprinting: Produces a fingerprint representation of the multivariate probability distribution derived from raw flow cytometry data.
• Distribution-agnostic: Describes data empirically without assuming parametric forms such as Gaussian mixtures.
• Data transformation for modeling: Converts raw flow cytometry datasets into formats that can be consumed by conventional statistical and empirical modeling software.
• Bioconductor integration: Integrates with other Bioconductor tools for downstream analysis.
• Scalability and efficiency: Handles extremely large datasets with arbitrary dimensionality for high-throughput studies.
• Supports classification workflows: Produces representations usable in automated classification and machine learning pipelines.

Scientific Applications:

• Data quality control: Facilitates quality assessment of multivariate flow cytometry datasets.
• Automated classification: Enables automated classification tasks, including applications such as Acute Myeloid Leukemia (AML) classification.
• Statistical and machine learning analyses: Provides inputs for conventional statistical modeling and empirical machine learning on cytometry data.
• High-throughput cytometry studies: Applies to large-scale immunology and oncology flow cytometry experiments requiring multivariate analysis.

Methodology:

Transforms raw flow cytometry datasets into empirical "fingerprint" representations of the multivariate probability distribution without assuming parametric distributional forms, producing modeling-ready outputs scalable to arbitrary dimensionality.