MultiVeloVAE - Velocity inference from multi-lineage, multi-omic, and multi-sample single-cell data

Package Installation

The package depends on several popular packages in computational biology and machine learning, including scanpy, scVelo, PyTorch, and scikit-learn. We suggest using a GPU to accelerate the training process.

We will make the package available on PyPI and Bioconda soon, but you can download and use it now by adding the path of the package to the system path:

import sys
sys.path.append(< path to the package >)
import velovae as vv

Please feel free to test our method on our previously published 10X Multiome datasets. See https://multivelo.readthedocs.io/en/latest/MultiVelo_Demo.html. The example of running the mouse brain dataset is located in paper-notebooks. Alternatively, you can apply the same training and analysis steps on our previous single-sample HSPC dataset for which we provide the AnnData objects directly in figshare. Expected run times can be found inside the notebooks.

This file lists the versions of packages used to generate manuscript figures.