LD Score Regression (LDSC)

This is a computational pipeline to perform the LD Score Regression (LDSC) analysis. For more information, see here: https://github.com/bulik/ldsc.

If you use this pipeline in your paper, Please cit the original publications: https://github.com/bulik/ldsc#citation

Installation

Prerequisite: install the ldsc python program first by following these instructions.

To install this pipeline, download pre-built binaries for macOS or Linux systems:

• ldsc-CentOS-x86_64: for Red Hat Enterprise Linux derivatives.

• ldsc-Ubuntu-x86_64: for Debian linux derivatives.

• ldsc-macOS-XX-XX: for macOS.

Example:

curl -L https://github.com/kaizhang/LDSC/releases/latest/download/ldsc-CentOS-x86_64 -o ldsc
chmod +x ldsc
./ldsc --help

Use the pipeline

To run this pipeline, you need to prepare the following files:

• Annotaton files containing genome regions in BED format.
• GWAS summary statistics.
• plink files
• A list of SNPs
• LD scores
• LD score weights

As an example, we will run the pipeline for GRCh38 genome and European acestry:

1. Download all files listed here: https://osf.io/86upf/files/osfstorage.
2. Go to the example directory: cd example.
3. Modify the ldsc_program_path in the configuration file config.dhall.
4. Run the pipeline: ldsc run --config config.dhall

The states are stored in a file called sciflow.db, which is used to resume the program at the last checkpoint. The results are in the output directory.

The running time depends on the number of annotations and GWAS summary statistics. You can use parallel computing to speed up the analysis.

To use parallelism in one computer: ldsc run --config config.dhall -n 10 +RTS -N10. The number 10 here is the number of parallel jobs.

To use the pipeline to in a grid computing system, adjust the relevant settings in config.dhall, and run it using: ldsc run --config config.dhall -n 10 --cloud. The default setting is configured for the TSCC environment at UCSD.