bayesNMF: Bayesian Poisson NMF Software for Mutational Signatures

An open-source R software package implementing a Bayesian Poisson Non-Negative Matrix Factorization (NMF) algorithm. The package includes models and plotting capabilities for analyzing count data, with applications focused on cancer mutational signatures. It was authored by Jenna M. Landy and published on figshare in April 2026.

Use Cases

• Analyzing cancer mutational signatures based on Bayesian Poisson NMF.
• Modeling count data with automatically learned latent rank based on the BIC-based sparsity prior.
• Quantifying posterior uncertainty in matrix factorization results based on the Bayesian formulation.
• Performing fast NMF without Poisson augmentation based on the MH-within-Gibbs sampler.

Strengths

• Software package is open-source and available on GitHub.
• Algorithm removes need for computationally expensive Poisson augmentation.
• Includes plotting capabilities demonstrated in the paper.

Limitations

• Dataset row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data may reflect bias inherent to the specific applications described.

Provenance

Source

figshare

Freshness

Last updated 2026-04-13 16:29:51.