This repository contains scripts for GWAS analysis. The preprint is online.
To replicate the analysis you will have to install the LDAK software > v5.2. Installation instructions and documentation can be found here.
First, you will have to prepare your phenotypes in PLINK Format. This pipeline expects phenotypes to be adjusted for covariates.
The analysis was performed on directly genotyped data from the UK Biobank. Data is expected to be in PLINK Format. Please also provide a list of samples to be included (homogeneous ancestry and unrelated individuals) and markers (i.e. SNP) that passed quality control (e.g. MAF > 0.01). For sex-specific analyses, provide the corresponding files of phenotypes and list of included individuals of the sex of interest.
For the LDAK-Bolt-Predict procedure, 65 functional annotations, of which 64 can be downloaded are needed. The final annotation depends on the genetic data at hand with code provided in the Snakefile.
The pipeline we followed is outlined in the Snakefile. Once all input data files and paths are specified, computation can be launched with the command snakemake. Cluster usage for snakemake is described here.