SLURM profile for SnakeMake

SnakeMake is a pipeline manager for bioinformatic analyses. To make sure SnakeMake works with SLURM at Oregon Health and Science University, please follow the instructions:

$ cd ~/.config/snakemake
$ git clone https://github.com/maxsonBraunLab/slurm.git
$ cd slurm # edit cluster.yaml and save the changes
$ chmod +x *.py *.sh # add execution permission to every py and sh file

When editing the cluster.yaml file, you can set default arguments to your snakemake submission command. See the following example:

$ cat ~/.config/snakemake/slurm/config.yaml

jobscript: "slurm-jobscript.sh"
cluster: "slurm-submit.py"
cluster-status: "slurm-status.py"
max-jobs-per-second: 20
max-status-checks-per-second: 10
local-cores: 1
latency-wait: 60

use-conda: True
printshellcmds: True
conda-prefix: "/directory/to/anaconda3/envs"

With these arguments, running the command snakemake -j 8 --profile slurm is the same as snakemake -j 8 --use-conda -p --conda-prefix /directory/to/anaconda3/envs. More command line arguments can be found in the SnakeMake documentation.

Note that the cluster.yaml in this folder belongs in ~/.config/snakemake/slurm, and is not the same as the cluster.yaml in a pipeline that allocates job resources per rule. This is a one-time setup and requires minor edits over time, but the same profile will work for every pipeline you run. To add a different HPC profile, copy the whole directory structure into ~/.config/snakemake/<profile>. To read more about SnakeMake profiles, please read the comprehensive documentation.