GigaSOM.jl

GigaSOM.jl implements scalable clustering and dimensionality reduction for large-scale flow and mass cytometry datasets using distributed computing.

Key Features:

• Scalability: Handles datasets with billions of data points and scales horizontally across distributed computing infrastructures.
• Implementation (Julia): Implemented in Julia to leverage language performance for fast processing of large datasets.
• Performance efficiency: Enables rapid processing times for massive datasets suitable for time-sensitive analyses.
• High-dimensional data handling: Processes datasets with over 40 parameters across thousands of samples without relying on downsampling.
• Result quality: Produces clustering and dimensionality reduction results comparable in quality to state-of-the-art software tools.

Scientific Applications:

• Large-scale phenotyping (mouse): Applied to massive mouse phenotyping studies that generate extensive flow and mass cytometry datasets.
• Single-cell cytometry studies: Supports analysis of complex cellular interactions and phenotypic variation in single-cell flow and mass cytometry data.

Methodology:

Implements clustering and dimensionality reduction techniques tailored for flow and mass cytometry data and distributes computation across nodes in distributed computing environments.