GRAVI: Gene Regulatory Analysis using Variable Inputs

This is a snakemake workflow for:

1. Performing sample QC
2. Calling ChIP peaks
3. Performing Differential Binding Analysis
4. Comparing results across ChIP targets

The minimum required input is one ChIP target with two conditions.

Full documentation can be found here

Snakemake Implementation

The basic workflow is written snakemake, requiring at least v7.7, and can be called using the following steps.

Firstly, setup the required conda environments

snakemake \
	--use-conda \
	--conda-prefix '/home/steveped/mambaforge/envs/' \
	--conda-create-envs-only \
	--cores 1

Secondly, create and inspect the rulegraph

snakemake --rulegraph > workflow/rules/rulegraph.dot
dot -Tpdf workflow/rules/rulegraph.dot > workflow/rules/rulegraph.pdf

Finally, the workflow itself can be run using:

snakemake \
	-p \
	--use-conda \
	--conda-prefix '/home/steveped/mambaforge/envs/' \
	--notemp \
	--rerun-triggers mtime \
	--keep-going \
	--cores 16

Note that this creates common environments able to be called by other workflows and is dependent on the user. For me, my global conda environments are stored in /home/steveped/mambaforge/envs/. For other users, this path will need to be modified.

If wishing to tidy the directory after a successful run, you can check which non-essential files can be deleted using snakemake -n --delete-temp-output --cores 1. If the files earmarked for deletion are considered to be non-essential, they can be deleted by removing the -n flag from the above code: snakemake --delete-temp-output --cores 1. As the bedgraph files produced by macs2 callpeak are typically very large, hence their conversion to bigwig files during the workflow, this step can free a considerable amount of disk space.