STGRNS, a Transformer-based model, provides a fast and accurate tool to infer gene regulatory networks from a single-cell RNA-seq profile. By leveraging the newly designed neural network structure, STGRNS especially obtains an outperformance on GRN inference.
Instructions and examples are provided in the following tutorials.
- scikit-learn (Compatible with all versions)
- Pytorch (With Cudatoolkit is recommanded)
- Numpy > 1.20
- Scanpy > 1.9.1
For the GRN reconstruction task, the processed experimental single-cell gene expression datasets are available on Zenodo at https://doi.org/10.5281/zenodo.3378975.
The gene expression and ChIP-Seq data of bone marrow-derived macrophages, dendritic cells, and mESC(1) are available at https://github.com/xiaoyeye/CNNC.
https://academic.oup.com/bioinformatics/article/39/4/btad165/7099621