scSSGC

This is the official codebase for scSSGC: Self-supervised graph representation learning for single-cell classification..

Python environment setup with Conda

1. Install Python and Pytorch

conda create --name scSSGC python=3.8
conda install pytorch==1.8.0 torchvision==0.9.0 torchaudio==0.8.0 cudatoolkit=11.1 -c pytorch -c conda-forge
conda activate scSSGC
conda clean --all

2. Install torch_geometric

2.1 Download the required dependencies

Go to the pyg-team/pytorch_geometric page and download the following files:

Download list：
torch_scatter-2.0.8-cp39-cp39-linux_x86_64.whl
torch_sparse-0.6.12-cp39-cp39-linux_x86_64.whl
torch_cluster-1.5.9-cp39-cp39-linux_x86_64.whl 
torch_spline_conv-1.2.1-cp39-cp39-linux_x86_64.whl

2.1 Install the required dependencies and torch_geometric

pip install torch_scatter-2.0.8-cp39-cp39-linux_x86_64.whl
pip install torch_sparse-0.6.12-cp39-cp39-linux_x86_64.whl
pip install torch_cluster-1.5.9-cp39-cp39-linux_x86_64.whl 
pip install torch_spline_conv-1.2.1-cp39-cp39-linux_x86_64.whl 
pip install torch_geometric

3. Install others

pip install pandas fbpca faiss-gpu annoy matplotlib numpy==1.26.4 

Running scSSGC (Remember to update the dataset addresses)

conda activate scSSGC
# Running scSSGC 
python main.py 

Preprint and Citation