bioNMF

bioNMF applies nonnegative matrix factorization (NMF) to decompose high-dimensional biological datasets into non-negative components for discovery of latent patterns in genomics, proteomics, gene expression, protein interactions and systems biology.

Key Features:

• Nonnegative Matrix Factorization: Implements NMF to decompose datasets into non-negative components for identification of underlying patterns and structures.
• Parallel Implementation: Provides a computationally efficient parallel implementation of NMF algorithms to improve scalability on large datasets.
• Versatile Analysis Contexts: Applies to gene expression, protein interactions and other complex biological systems for exploratory analysis.
• Web Services Interface: Exposes a public web services interface for programmatic submission of NMF jobs and integration into computational workflows.

Scientific Applications:

• Genomics: Reveals latent components in genomic datasets to assist in pattern discovery across high-dimensional data.
• Proteomics: Identifies protein expression patterns and interaction modules in proteomic data.
• Systems Biology: Decomposes high-dimensional data to support identification of biomarkers and exploration of regulatory networks.

Methodology:

Performs NMF by decomposing a dataset V into two non-negative matrices W (basis) and H (coefficient) such that V ≈ WH, and implements parallel NMF algorithmic variants for computational efficiency.