cPredictor 0.4.6

Cell command line predictor

cPredictor package

This repository defines a command-line tool to predict (cPredictor) datasets according to a cell meta-atlases. At the present time only the meta-atlas for the cornea has been implemented.

Conda and pip

If you have not used Bioconda before, first set up the necessary channels (in this order!). You only have to do this once.

$ conda config --add channels defaults
$ conda config --add channels bioconda
$ conda config --add channels conda-forge

Install cPredictor into a conda environment and install with PyPI:

$ conda create -n cPredictor python=3.9 pip
$ conda activate cPredictor
$ pip install cPredictor

To see what each of the current functions do you can run these commands:

$ SVM_performance --help
$ SVM_predict --help
$ SVM_import --help
$ SVM_pseudobulk --help

Docker

Alternatively you can run the package containerized through docker:

$ docker pull artsofcoding/cpredictor:latest
$ docker tag artsofcoding/cpredictor:latest cpredictor
$ docker run -it --name cpredictor -p 8080:80 -v {path_to_H5AD_object}:/data cpredictor

In the activated docker container you can then go to the terminal:

# cd /data
# SVM_predict --query_H5AD {H5AD_object}.h5ad --OutputDir {your_output_dir} --meta_atlas

Performance with the corneal meta-atlas

Pretrained models can run on ~100.000 cells within 2 minutes on a standard laptop (4 core CPU & 8GB RAM)

To run the container locally you will need a computer with at least 28 GB of RAM and a 4-core processor.

The documentation will be extended and improved upon in later versions.

How to cite

When using this software package, please correctly cite the accompanied DOI under "Citation": https://zenodo.org/doi/10.5281/zenodo.10621121