The MultiSP package can be run on GPU (recommend) or CPU.
This package has been tested on Windows, Linux and macOS (Ventura) operating systems, and should work in any valid python environment.
- python==3.11
- torch==2.4.0
- numpy==1.26.4
- pandas==2.22.2
- scanpy==1.10.2
- episcanpy==0.4.0
- anndata==0.10.8
- rpy2==3.5.11
- scipy==1.14.0
- scikit-learn==1.5.1
- tqdm==4.66.5
- matplotlib==3.9.2
- R==4.3.1
It's prefered to create a new environment for MultiSP
conda create -n MultiSP python==3.11
conda activate MultiSP
MultiSP is available on PyPI and can be installed via
pip install multisp
Install all the required packages
pip install -r requirements.txt
Installation of MultiSP should take less than a minute and it may take several minutes to install the dependencies.
The use of the mclust algorithm requires the rpy2 package (Python) and the mclust package (R). See https://pypi.org/project/rpy2/ and https://cran.r-project.org/web/packages/mclust/index.html for detail.
The details of all datasets used are available in the Methods section. all datasets are available at https://drive.google.com/file/d/15iN5XumcEFptHSrS2YudhllwX4iFX9H7/view?usp=drive_link. Step-by-step tutorials are included in the Tutorial folder to show how to use MultiSP.
- 1.Tutorial for spatial RNA-ADT human lymph node dataset (It takes about 20 seconds to run on GeForce RTX 3090 GPU )
- 2.Tutorial for spatial RNA-ATAC MISAR_seq dataset (It takes about 3 minutes to run on GeForce RTX 3090 GPU)
- 3.Tutorial for spatial P22 mouse brain dataset (It takes about 50 seconds to run on GeForce RTX 3090 GPU)
- 4.Tutorial for human tonsil trimodal dataset (It takes about 1 minutes to run on GeForce RTX 3090 GPU)
Tutorial for inferring spatially multimodal cell-cell communication is available at the Github repository of CellChat toolkit