parkerlab / islet_cage

Analyses of human pancreatic islet CAGE data

This repository contains code to run analyses presented in the manuscript:

Integrating Enhancer RNA signatures with diverse omics data identifies characteristics of transcription initiation in pancreatic islets

Directories in the analysis folder contain individual workflows, run using snakemake version 5.5.0. Config for a SLURM cluster execution are provided. The analysis directories follow this general pattern:

├── run.sh : for workflow execution
├── configs
│   ├── cluster.yaml : Cluster job specifications (SLURM)
│   └── config.yaml : Workflow configuration (created from mkconfig.py while using environment variables)
└── scripts : Scripts for analyses
    ├── mkconfig.py
    ├── script1.py
    ├── script2.R
    └── Snakefile : Snakemake files(s) 

Given the correct environment variables are set and the input data is downloaded from Zenodo (see below), executing the run.sh file sets base directory and paths, generates a workflow config and can be used to execute analysis or print a dry run). Most software required can be set up by a conda environment. Conda can be obtained through the Anaconda/Miniconda Python3 distribution. These instructions are for the Linux platform:

1. Set up conda and software:

Install Anaconda3 if you don't already have conda on your system.

Please manually install GREGOR, required to compute enrichment of GWAS in regulatory annotations. Edit the path to the GREGR.pl script file in env/env_vars.sh.

For atac-seq, bwa and bwa hg19 index are to be specified in env/env_vars.sh.

For MPRA analysis, install R package MPRAnalyze version 1.3.1 which is currently available from github

install.packages("remotes")
remotes::install_github("YosefLab/MPRAnalyze")

2. Prepare analysis directory

Clone this repository and change into it. Download the data archive from the Zenodo deposition and untar the archive

tar -xvzf islet-cage-zenodo.tar.gz

3. Set up conda environment

1. Fill up file env/env_vars.sh specifying email, manually installed software paths etc.

conda env create --name cage --file env/cage.yaml
conda activate cage
mkdir -p $CONDA_PREFIX/etc/conda/activate.d
mkdir -p $CONDA_PREFIX/etc/conda/deactivate.d

cp env/env_vars.sh $CONDA_PREFIX/etc/conda/activate.d/
cp env/env_vars.sh $CONDA_PREFIX/etc/conda/deactivate.d/

Analyses

Dry run of <analysis_name>

analysis/<analysis_name>/run.sh -n

Run analyses by submitting jobs

analysis/<analysis_name>/run.sh