scSLAT: single cell spatial alignment tools

Installation • Tutorial • Publication

scSLAT package implements the SLAT (Spatial Linked Alignment Tool) model to align single cell spatial omics data. For more details, please check out our publication.

Directory structure

.
├── scSLAT/                  # Main Python package
├── env/                     # Extra environment
├── data/                    # Data files
├── evaluation/              # SLAT evaluation pipeline
├── benchmark/               # Benchmark pipeline
├── case/                    # Case studies in paper
├── docs/                    # Documentation files
├── resource/                # Other useful resource 
├── pyproject.toml           # Python package metadata
└── README.md

Tutorial

Tutorial of scSLAT is here, if you have any question please open an issue on github

Installation

Docker

Dockerfile of scSLAT is available at env/Dockerfile. You can also pull the docker image directly from here by:

docker pull huhansan666666/slat:0.2.1

PyPI

Note

Installing scSLAT within a new conda environment is recommended.

First, we create a clean environment and install scSLAT from PyPI. Then we also need install dependencies for pyg manually. We default install with CUDA 11.7. Please refer here for CPU version or different CUDA versions.

Warning

old NVIDIA driver may raise error, please update NVIDIA driver to the latest version.

conda create -n scSLAT python=3.8 -y && conda activate scSLAT
pip install scSLAT
python -c "import torch; print(torch.__version__)"
pip install pyg_lib torch_scatter torch_sparse torch_cluster torch_spline_conv -f https://data.pyg.org/whl/torch-2.0.0+cu117.html  # replace torch and CUDA version to yours

Development version

For development purpose, clone this repo and install:

git clone git@github.com:gao-lab/SLAT.git
cd SLAT
pip install -e ".[dev,docs]"
pip install pyg_lib torch_scatter torch_sparse torch_cluster torch_spline_conv -f https://data.pyg.org/whl/torch-2.0.0+cu117.html

Reproduce manuscript results

1. Please follow the env/README.md to install all dependencies. Please checkout the repository to v0.2.1 before install scSLAT.
2. Download and pre-process data follow the data/README.md.
3. Whole benchmark and evaluation procedure can be found in /benchmark and /evaluation, respectively.
4. Every case study is recorded in the /case directory in the form of jupyter notebook.