cadast 0.1.0

A python implement of CadaST

Probabilistic-Graph Based Spatial Context-Aware Framework for Interpretable Spatial Omics Denoising and Augmentation

Description

This project is a python implementation of CadaST, a spatial omics data denoising and augmenting method based on similarity graph. The method is proposed in the paper Probabilistic-Graph Based Spatial Context-Aware Denoising and Augmenting in Spatial Omics. The method is implemented in the class CadaST in model.py.

Installation

Use pip to install the package in your python environment.

conda create -n cadaST python=3.9
conda activate cadaST
pip install cadaST

To enable clustering method 'mclust', you shall install R package 'mclust'

pip install rpy2
Rscript -e "install.packages('mclust')"

Usage

To get detailed tutorial, please refer to the tutorial.

Parameters

• beta $\beta$: The scaling weight for the similarity graph.
• alpha $\alpha$: The scaling weight for correlation matrix.
• theta $\theta$: The decay value for non-matching labels in the label matrix.
• n_jobs: The number of processes used for parallel computing.
• kneighbors: The k-Nearest Neighbors to construct graph.

Contributing

Contribution is welcomed! Here’s how you can help:

• Report bugs or suggest features using GitHub Issues.
• Fork the repository and create a new branch for your feature or bugfix.
• Submit a pull request and ensure that all tests pass.

License

This project is licensed under the MIT License - see the LICENSE file for details.