mRNA-Seq Workflow

Workflow for the analysis of mRNA-seq data for Yan et. al, "An endogenous small RNA-binding protein safeguards prime editing" (in press).

The workflow is written using Snakemake and Quarto.

Dependencies are installed using Bioconda where possible.

The workflow consists of two pieces, one written in Snakemake, the other is composed of Quarto notebooks.

Running the Snakemake workflow

Here, we create two workflows to work with the two subsets of data separately: HEK3_1TtoA and PRNP_6GtoT. Run the workflow in each directory separately.

The workflows use the publicly available rna-seq-star-deseq2 workflow. Citation: https://doi.org/10.5281/zenodo.4737358

1. Clone workflow into working directory

   git clone <repository> <dir>
   cd <dir>

2. Download input data (or skip and use demo-data)

   Copy the fastq files into data directory

3. Edit the configuration as needed (not needed if using demo-data)

   # Edit location of fastq files
   nano HEK3_1TtoA/config/units.yaml
   nano PRNP_6GtoT/config/units.yaml
   
   # Generally, these can remain unchanged 
   nano HEK3_1TtoA/config/samples.yaml
   nano PRNP_6GtoT/config/samples.yaml
   nano HEK3_1TtoA/config/config.yaml
   nano PRNP_6GtoT/config/config.yaml

4. Install Snakemake and Snakedeploy

   mamba create -c conda-forge -c bioconda --name snakemake snakemake snakedeploy mamba activate snakemake

5. Setup workflow specific resources

   1. Modify workflow/profiles/default/config.yaml to ensure rules have the required resources to run on the cluster.

6. Run the workflow (using cluster options is recommended)

   snakemake --use-conda -cores 1

7. Generate a report

   snakemake --report report.zip

Quarto notebooks

The Quarto notebooks utilize R and are run separately.

1. Run the workflows as above

2. Load the Rproject ./pe-mrna-seq-diffexp.Rproj in RStudio.

3. This project uses renv to keep track of installed packages. Install renv if not installed and load dependencies with renv::restore().

4. Load the quarto notebook ./mrna-seq-venn-diag.qmd and run all of the cells or use the "Render" button in RStuido.