Anaconda is a free and open source Python and R distribution, which includes binaries such as, Scipy, Numpy, Pandas along with all their dependencies.

Conda is an open source package management system and environment management system. Conda is designed to manage packages and dependencies within any software stack, more like yum and apt, conda is a language-agnostic cross-platform environment manager.

In order to accomodate Nero Users to install and customize their SW environments, we have set up a local mirror of Anaconda distributions.

The following is a list of default channels included in our anaconda distribution.

Base Conda Channels Mirrors

Channel Channel Map
main https://nero-mirror.stanford.edu/conda/main
conda-forge https://nero-mirror.stanford.edu/conda/conda-forge

Other Available Conda Channel Mirrors

bioconda  https://nero-mirror.stanford.edu/conda/bioconda
free https://nero-mirror.stanford.edu/conda/free
hcc  https://nero-mirror.stanford.edu/conda/hcc
pytorch https://nero-mirror.stanford.edu/conda/pytorch
soumith  https://nero-mirror.stanford.edu/conda/soumith

To use Anaconda 3

Connect to Stanford VPN

Login to nero.compute.stanford.edu

ssh $username@nero.compute.stanford.edu

Source Anaconda 3 with module load

module load anaconda

Search for desired software packages from base channels

Check current loaded version of package

conda list -f $PACKAGENAME

Check versions available from the conda mirror with:

conda search -f $PACKAGENAME
conda search -f boto3

Check for certain version

conda search -f boto3=1.9.21

Search for desired software packages from specific channel

conda search -f $PACKAGENAME -c <channel>
conda search -f samtools -c https://nero-mirror.stanford.edu/conda/bioconda

Create and use custom environment using base channels

A conda custom environment is a directory that contains a specific collection of conda packages that you want installed.

conda create -p <path to directory> <sw> <sw>
conda create -p /home/$USER/envs/dev numpy scipy

To specify version

conda create -p <path to directory> <sw>=<version>
conda create -p /home/$USER/envs/python_2.7 python=2.7

To utilize an environment

conda activate /home/$USER/envs/python_2.7

To deactivate current environment

conda deactivate

Create custom environment using non-base channel

conda create -p /home/$USER/envs/bio-env -c  <channel to use> <sw>=<version>
conda create -p /home/$USER/envs/bio-env -c https://nero-mirror.stanford.edu/conda/bioconda bwa=0.7.15 samtools=1.3.1

Activate and deactivate as done above

conda activate /home/$USER/envs/bio-env

Add packages to custom environment

conda activate /home/$USER/envs/python_2.7
conda install <packagename>
conda install curl

If non-base channel is needed you will need to specifiy the channel

conda activate /home/$USER/envs/bio-env
conda install bcftools -c https://nero-mirror.stanford.edu/conda/bioconda

Using pip with your custom environment

If conda does not have the package you may need to use Pip:

conda activate /home/$USER/envs/env-name
pip install --user packagename

If pip fails, make sure you have a “.pip/pip.conf” file in your home directory, with the following contents:

index-url = https://nero-mirror.stanford.edu/pypi/simple

Export environment to a file

conda actvate /home/$USER/envs/bio-env
conda env export > bio-environment.yaml

Which can then be used by someone else to create same environment

conda create -p /home/$USER/envs/bio-env --file bio-environment.yaml

Managing Quota Space Used by Anaconda

Anaconda has a tendency to use disk space - especially the package cache. You can alleviate the usage of your home directory quota (10GB) by setting your “.condarc” and specifying a envs_dir location with the following command. We suggest setting it to your “/share/pi/” directory as there is more quota available there.

conda config --add envs_dirs <path to directory>