flowBin

flowBin integrates multiplexed flow cytometry (FCM) data collected across separate tubes or aliquots to reconstruct higher-dimensional marker expression matrices for deep phenotypic profiling and discovery.

Key Features:

• Multiplex Data Integration: Establishes bins based on shared markers to combine data from tubes that share common markers but differ in tube-specific markers.
• Parameterization-Free Methodology: Operates without predefined clustering parameters and contrasts with nearest-neighbor (NN) imputation methods that can introduce artificial subpopulations.
• High-Dimensional Data Representation: Allocates cells to bins defined by common markers and computes aggregate expression levels of tube-specific markers within each bin to construct a comprehensive expression matrix.
• Validation with Simulated Data: Demonstrates efficacy using simulated multitube data where flowType analysis of flowBin outputs replicates results from original datasets for cell types with abundance >10%.
• Disease Profiling Application: Integrates with classifiers to distinguish normal and cancerous cells and supports identification of novel cell types associated with specific mutations.

Scientific Applications:

• High-dimensional phenotyping: Reconstruction of complete marker panels from multiplexed FCM for deep phenotypic profiling and discovery of cellular subpopulations.
• Cancer vs normal classification: Integration of flowBin outputs with classifiers to differentiate normal and cancerous cells.
• NPM1-mutated acute myeloid leukemia profiling: Profiling the immunophenotypic landscape of NPM1-mutated acute myeloid leukemia and identifying novel associated cell types.

Methodology:

Binning based on shared markers, allocation of cells to bins, aggregation of tube-specific marker expressions per bin, a parameterization-free approach contrasted with nearest-neighbor (NN) imputation, and validation using simulated multitube data with flowType analysis.