scmallet 0.1.2

Python wrapper of MALLET for LDA analysis on single-cell data

scmallet

Python wrapper of MALLET for LDA analysis on single-cell data.

MALLET is the LDA backend chosen by pycistopic. The implementation in this package has several difference than pycistopic:

1. Improved paralization.
2. Allow train LDA model with cell subset and then inference ramaining cells. This is nicely supported by MALLET itself.
3. Allow the training process to be resumable.
4. Work with anndata.

Installation

You need to have Python 3.9 or newer installed on your system. If you don't have Python installed, we recommend installing Mambaforge.

There are several alternative options to install scmallet:

1. Install the latest release of scmallet from PyPI:

pip install scmallet

2. Install the latest development version:

pip install git+https://github.com/lhqing/scmallet.git@main

Usage

See example usage here: https://github.com/lhqing/scmallet/blob/main/tests/example.ipynb

Citation

• Mallet Package: https://mimno.github.io/Mallet/

  McCallum, Andrew Kachites. "MALLET: A Machine Learning for Language Toolkit." http://mallet.cs.umass.edu. 2002.

• PyCistopic: https://github.com/aertslab/pycisTopic

  Bravo Gonzalez-Blas, C. & De Winter, S. et al. (2022). SCENIC+: single-cell multiomic inference of enhancers and gene regulatory networks