scLeap is a pre-trained deep learning framework designed to predict genes driving cell-state transitions with single-cell CRISPR screening data.
- Clone the repo
git clone https://github.com/DLS5-Omics/scLeap.git
- Install python dependencies
pip install -r requirements.txt
Please refer to tutorials/README.md
.
The full descriptions of the datasets and the studies of origin can be found in the manuscript. Here we provide the links to the pretrained model and the example datasets.
- Pretrained model checkpoint and gene2token file: link
- Example training and testing datasets, with adjacency matrix graph and shortest path graph provided: link