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.

To install the MultiVeloVAE package through PyPI:

pip install multivelovae

And import the package inside python:

import multivelovae as vv

Package Usage

See our ReadTheDocs page for details about API and notebooks.

The example notebooks of running the mouse brain and HSPC/macrophage datasets are located in paper-notebooks. Processed AnnData objects are shared directly through figshare. Expected runtimes using RTX3060-level graphics cards can be found inside each notebook.

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

TODO

bioconda