FlowSOM

FlowSOM applies a Self-Organizing Map (SOM)-based approach to analyze and visualize high-dimensional flow and mass cytometry data, enabling identification and summarization of cellular populations and marker behavior.

Key Features:

• Self-Organizing Map (SOM): Uses a Self-Organizing Map (SOM) algorithm to project high-dimensional cytometry measurements onto a grid of nodes.
• Two-level clustering: Performs two-level clustering that first organizes data into clusters and then refines those clusters to reveal detailed substructure.
• Star chart visualization: Generates star charts that display marker expression profiles across clusters to visualize marker behavior within and between cell groups.
• High-dimensional cytometry support: Targets datasets from flow cytometry and mass cytometry with large numbers of markers.
• Cell population and subset identification: Detects distinct cell populations and subtle cellular subsets within complex datasets.
• Marker interaction summarization: Summarizes marker interactions and cell subset distributions across the dataset.

Scientific Applications:

• Discovery of novel phenotypes: Supports discovery of novel cellular phenotypes from high-dimensional cytometry data.
• Immune-response characterization: Facilitates characterization of intricate immune responses by identifying population-level and subset-level changes.
• Large-scale cytometry analysis: Enables analysis and visualization of large-scale flow and mass cytometry datasets.

Methodology:

Implements a Self-Organizing Map (SOM) algorithm followed by two-level clustering (initial clustering and further subdivision) and visualizes marker profiles with star charts.