lyh970817/tngwas

A simple and extensible GWAS pipeline for SLURM

tngwas

This is a simple and extensible GWAS pipeline writtien in bash to be used with Slurm Workload Manager. Currenly the pipeline has pipeline templates to run association analyses (including manhattan and QQ plots) in regenie and plink2 as well as heritabiltiy estimation in ldsc.

The lib/ directory contains default pipeline templates which are structured as directoris containing a *.sbatch job submission script. The pipeline templates can also use scripts in utils/ (which is design to keep scripts usable by multiple pipeline templates and can be set by the UTILS variable in the config file).

Usage

Clone the repository

git clone https://github.com/lyh970817/tngwas

Generate the default project directory structure in the repository direcotry

cd tngwas
./main.sh --init

An example config file which is required to run the pipeline will be generated. If ldsc is to be used, make sure to specify the ldsc, snp_list and ld_chr variables.

Before actually running the pipeline, make sure to have the GWAS software in the selected pipeline template, parallel and R in your PATH variable (for example, using module load). The qqman and data.table R packages should also be installed.

To run the pipeline, specify the config file and a pipeline template directory.

./main.sh --config config --pipe ./lib/regenie-bt

To clear the file content in LOG and OUT before running the pipeline, add the --fresh flag.

Extending the pipeline

To extend the pipeline, it is as simple as creating a directory containing a set of scripts to be submitted as sbatch jobs and a job submission script named DIRECTORYNAME.sbatch. The scripts can utilize a number of pre-defined variables and use sharable scripts kept in the directory set by UTILS in the config file. Check the exsiting pipeline templates for examples.