Cellcano is an open source software for supervised cell type identification (celltyping) in scATAC-seq data, published in Nature Communications. The motivation to develop Cellcano are:
- Supervised methods are more accurate, robust and efficient than unsupervised clustering methods in scATAC-seq data
- With more high-quality scATAC-seq datasets being generated, methods using scATAC-seq as references can have better prediction performances and are in high demand
More details and tutorial: https://marvinquiet.github.io/Cellcano/.
Table of Contents
Cellcano package requires only a standard computer with enough RAM to support the in-memory operations. Cellcano can use GPU if the computer has the GPU resource but it is not required.
Cellcano supports macOS, Linux and Windows. It has been tested on all three systems.
(However, Cellcano has not been tested on M1 because I do not have the test environment. Thanks to @nleroy917, who has helped with the installation on M1, which can be referred in Issue #6.)
Cellcano requires the following:
- python (3.8 recommended)
- R
- tensorflow (2.7.1)
- anndata (0.7.4)
- scanpy (1.8.2)
- numpy (1.19.2)
- h5py (2.10.0)
- keras (version compatible with tensor flow)
- rpy2 (version compatible with both Python and R)
cuda toolkitandnvidia cudnnif using GPU, more information can be found here
If the input is scATAC-seq raw data (i.e. fragment file or bam file), ArchR package has to be installed.
The most convinient way is to install with pip.
pip install CellcanoTo upgrade to a newer release use the --upgrade flag.
pip install --upgrade CellcanoWe have a detailed tutorial on installation in our documentation.
This project is covered under the MIT license.
For usage of the package and associated manuscript, please cite:
@article{ma23cellcano,
title = {Cellcano: supervised cell type identification for single cell ATAC-seq data},
author = {Ma, Wenjing and Lu, Jiaying and Wu, Hao},
journal = {Nature Communications},
year = {2023},
month = {Apr.},
day = {03},
volume={14},
number={1},
pages={1864},
issn={2041-1723},
doi={10.1038/s41467-023-37439-3},
url={https://doi.org/10.1038/s41467-023-37439-3}
}