wrangler_by_sample

A program written to wrangle data on a sample by sample basis - each sample gets its own independent job submission on a slurm cluster (e.g. longleaf or oscar).

Alternatively, for non-cluster systems, each sample can be analyzed independently.

In both cases, If a run crashes, it doesn't have to be restarted from the beginning.

Installation:

Install mamba with: https://github.com/conda-forge/miniforge#mambaforge

This link will install a version of conda called 'mamba' that also includes conda. Mamba manages snakemake versions better than conda so you should be using mamba.

Install snakemake in an environment called snakemake with:

mamba create -c conda-forge -c bioconda -n snakemake snakemake

Make sure you have a profile that matches the computer you'd like to run this on. Example profiles are provided here for slurm clusters and for non_slurm standalone computers. You'll need to edit these profiles to be appropriate for the way that your system is currently configured (e.g. don't set memory or CPU requests that exceed your system's capacity). Copy this profile into a folder. If you put the profile in ~/.config/snakemake/your_profile_name/config.yaml, then you will be able to call it from the commandline with snakemake -s your_script.smk --profile your_profile_name. In the usage instructions below, we assume that you are using a slurm cluster, and that you put your file in ~/.config/snakemake/slurm/config.yaml, but on a non-cluster, these instructions will also work if you put your config file in ~/.config/snakemake/non_slurm/config.yaml and use --profile non_slurm

• Download the contents of this git repo with:

git clone https://github.com/bailey-lab/wrangler_by_sample.git

• cd into the folder and open the yaml file (wrangler_by_sample.yaml) and fill in the variables using instructions from the comments
• Activate snakemake with:

mamba activate snakemake

Usage:

• Run the first step with:

snakemake -s setup_run.smk --profile slurm

• Run the second step with:

snakemake -s finish_run.smk --profile slurm

• use --profile non_slurm if you are on a standalone computer (not a slurm cluster).