README.md

Salamander

Salamander is a non-negative matrix factorization (NMF) framework for signature analysis build on top of AnnData and MuData. It implements multiple NMF algorithms, common visualizations, and can be easily customized & expanded.

---

Installation

PyPI:

pip install salamander-learn

Usage

The following example illustrates the basic syntax:

import anndata as ad
import salamander as sal

# initialize data
adata = ad.AnnData(...)

# NMF with Poisson noise
model = sal.models.KLNMF(n_signatures=5)
model.fit(adata)

# barplot
model.plot_signatures()

# stacked barplot
model.plot_exposures()

# signature correlation
model.plot_correlation()

# sample_correlation
model.plot_correlation(data="samples")

# dimensionality reduction of the exposures
model.plot_embeddings(method="umap")

For examples of how to customize any NMF algorithm and the plots, check out the tutorial. The following algorithms are currently available:

• NMF with KL-divergence loss
• minimum-volume NMF
• a variant of correlated NMF

License

MIT

Changelog

Consult the CHANGELOG file for enhancements and fixes of each version.