VeloCycle

Manifold-constrained variational inference for RNA velocity of the cell cycle. This is the repository for the VeloCycle framework. Installation instructions and tutorials can be found below.

Getting started

Please refer to the installation instructions to install VeloCycle. Once you are ready to begin exploring the package, please refer to the tutorials contained in this repo. The HTML versions of each tutorial show the expected output of running VeloCycle as well as expected runtimes obtained using a Macbook Pro 2019 edition (faster runtimes will be achieved on newer computers or when using GPUs).

Installation

You need to have Python 3.8 or newer installed. All other package versions required are indicated in the requirements.txt file in this repo. Installation of VeloCycle should take only a few minutes on a standard operating system.

We suggest installing VeloCycle in a separate conda environment, which for example can be created with the command:

conda create --name velocycle_env python==3.8

You will probably need to install git next:

conda install git

Then you can install VeloCycle using one of the following two approaches:

1. Install the latest release on PyPI:

pip install velocycle

2. Install the latest development version:

pip install git+https://github.com/lamanno-epfl/velocycle.git@main

Release notes

This is the initial release of VeloCycle corresponding to our preprint: "Statistical inference with a manifold-constrained RNA velocity model uncovers cell cycle speed modulations".

Link: https://www.biorxiv.org/content/10.1101/2024.01.18.576093v1

These software are still under continuous development.

Contact

For questions and help requests, you can reach out to Alex Lederer and Gioele La Manno. We are happy to hear your feedback and comments!

Additional materials

For notebooks and data files not used in the tutorials included in this repo, but used in the original publication, please see the following Google Drive folder and GEO. In the future, files will be transferred from the Google Drive to Zenodo.

Citation

Lederer, A. R., Leonardi, M., Talamanca, L., Herrera, A., Droin, C., Khven, I., Carvalho, H. J. F., Valente, A., Dominguez Mantes, A., Mulet Arabí, P., Pinello, L., Naef, F., & La Manno, G. (2024). Statistical inference with a manifold-constrained RNA velocity model uncovers cell cycle speed modulations. In bioRxiv. https://doi.org/10.1101/2024.01.18.576093