GWAS

This repository provides a description of workflow for analysing GWAS data. All the required scripts are also included. A GWAS analysis can be subdivided into 5 major steps:

QC of unimputed data
Imputation
QC of imputed data
Association testing & filtering
Regional clustering and plotting

Note: In some cases, scripts, parameters and thresholds need to be adapted to consider certain data proporties.

Reference files

GRCh37 reference (fasta.gz)
1000 Genomes allele frequency annotations (vcf.gz)
1000GP_Phase3_combined.legend.gz

1. QC of unimputed data

Script: qc_unimputed.sh
Input data: Seperate Plink files for cases and controls with sex info and both family id and individual id set
Required programs: plink2.0, plink1.9, bcftools, python3.7, Rscript with libraries qqman and data.table
Annotation files: reference genome (fasta), allele frequency file

2. Imputation

For imputation I recommend the Sanger imputation server (free of charge) at https://imputation.sanger.ac.uk/. Either select "pre-phase with EAGLE2 and impute" or "pre-phase with SHAPEIT2 and impute" and upload the final VCF files from step 1. After successfull imputation download the VCF files.

3. QC of imputed data

Script: qc_imputed.sh
Input data: Input files (*.vcf.gz) from Sanger imputation server.
Required programs: bcftools, vcfparse, IC, Perl module Benchmark
Annotation files: allele frequency file, 1000GP_Phase3_combined.legend

4. Association testing & filtering

Script: assoc_tests.sh
Input data: Genotypes and sample information in .gen/.sample format (final output of step 3)
Required programs: snptest