Snakemake Cluster Tutorial

This repository is based on the Software Carpentries lesson repository for using Snakemake, a Python-based workflow manager. The lesson and assorted materials have been adapted for use on a Moab/Torque (PBS)-based system but are easily transferrable by adding a system specific cluster profile configuration.

Datasets

The analysis parses a set of books to calculate various statistics. All of the raw texts are available in the books/ directory.

Dependencies

All dependencies are listed in envs/clusterTutorial.yaml. You can install them using conda (see below for instructions).

Conda

If you don't already have conda installed, please download Miniconda. Miniconda3 is a Python 3-based package manager and is essentially a stripped-down version of Anaconda so it should be much faster to install.

Because this tutorial is designed to work on a Linux-based computer cluster, please follow the below instructions to set up the conda environment (if you already have an updated version of Miniconda installed, please skip to step 3):

1. Login to your cluster environment and navigate to your home folder.

2. Download & install Miniconda for 64-bit Linux systems. When it asks if it should initialize conda, say yes:

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
bash Miniconda3-latest-Linux-x86_64.sh

3. Open your cluster .bashrc in a text editor and edit the initialization script added by Miniconda so it resembles the following. This will allow environments to be properly inherited when passing jobs to cluster nodes for execution:

nano ~/.bashrc

# Adding conda to the path without initializing the base environment
# >>> conda initialize >>>
# !! Contents within this block are managed by 'conda init' !!
__conda_setup="$('/home/wlclose/miniconda3/bin/conda' 'shell.bash' 'hook' 2> /dev/null)"
if [ $? -eq 0 ]; then
    eval "$__conda_setup"
else
    if [ -f "/home/wlclose/miniconda3/etc/profile.d/conda.sh" ]; then
        . "/home/wlclose/miniconda3/etc/profile.d/conda.sh"
    else
        export PATH="/home/wlclose/miniconda3/bin:$PATH"
    fi
fi
unset __conda_setup
# <<< conda initialize <<<

4. Source your .bash_profile to make all of the changes take effect and finish installing Miniconda (this should also re-source your ~/.bashrc):

source ~/.bash_profile

Using Snakemake to Submit Jobs

1. Clone this repository and move into the main project folder:

git clone https://github.com/SchlossLab/snakemake_cluster_tutorial.git
cd snakemake_cluster_tutorial

2. Create an environment called clusterTutorial with the dependencies we need:

conda env create -f envs/clusterTutorial.yaml

3. Activate the environment before running any code:

conda activate clusterTutorial

4. Before running any jobs on the cluster, don't forget to change the ACCOUNT and EMAIL fields in the following files for whichever cluster you're using:

• PBS: cluster profile configuration and the cluster submission script
• Slurm: cluster profile configuration and the cluster submission script

5. Run the Snakemake workflow. Note: If you wish to rerun the workflow after having it successfully complete, use the --forcerun or the --forceall flags or just delete the results/ directory by running snakemake clean.

• To run the entire workflow locally (without the cluster):

snakemake

• To run the rules as individual jobs on a PBS cluster:

snakemake --profile config/pbs-torque/ --latency 20

Or to run a job that manages the workflow for you instead

qsub code/clusterSnakemake.pbs

• To run the rules as individual jobs on a Slurm cluster:

snakemake --profile config/slurm/ --latency 20

Or to run a job that manages the workflow for you instead

sbatch code/clusterSnakemake.sh

See the conda documentation for more information.