Installation Guide

Please follow this guide to install both the velocyto command line tool and the homonymous library for interactive analysis.

Requirements

To run velocyto you will need python >=3.6.0 (we have no plans to support python<=3.5). We recommend using anaconda and the conda command to install dependencies (of course you can use pip but be aware its dependency-managing might be less robust). Feel free to use pip if some libraries are not available on the conda channels you are using.

Run the following command to make sure you have all the dependencies correctly installed:

conda install numpy scipy cython numba matplotlib scikit-learn h5py click

Note: pysam is also a requirement but currently it is preferable to install it through PyPI wheel (i.e. by pip install pysam) as we have experienced problems with the version packaged by conda-forge.

Install using PyPI

Every new stable version of velocyto gets immediately released on PyPI, so running the following command will install on your system the cutting-edge version:

pip install velocyto

pysam and loompy will be installed by pip as a dependencies.

Tip

velocyto 0.17 is an alpha release and it is updated often. If you installed with pip make sure you run pip install -U --no-deps velocyto now and then.

You can test whether the installation was successful by running velocyto --help.

To get started with velocyto you can follow our guide.

Install from source

If you plan to explore and make changes to the source code, or you have requested some bug-fix that is temporarily available only on the github dev branch, then you need to install velocyto directly from source.

First of all, make sure all the dependencies are installed, and that git is installed on your system. Then, run the following commands to complete the installation:

git clone https://github.com/velocyto-team/velocyto.py.git
cd velocyto.py
pip install -e .  # note the trailing dot

You can test whether the installation was successful by running velocyto --help.

Tip

velocyto 0.17 is an alpha release, we recommend pulling in the latest bufixes and feature improvements often. Note that adding the -e flag to the pip command installs the software in development mode, when a package is installed this way each change to the source immediatelly reflects to changes in the installed library. This means that you can simply use git pull to update your installation.

To get started with velocyto you can follow our guide.

Install using conda

Note

This installation method is not currently available. Our plan is make it available upon the 1.0 release.

Extra requirements

If you want to use velocyto in combination with DropEst you will have to install some extra requirements: R and rpy2. We suggest to this with conda:

conda install R rpy2