Skip to content

Latest commit

 

History

History
71 lines (40 loc) · 1.49 KB

README.md

File metadata and controls

71 lines (40 loc) · 1.49 KB
Raw

COLOC scripts

Inputs:

  • QTL nominal summary statistics (per-chromosome parquet files)

Created by TensorQTL

  • GWAS summary statistics

Use EBI GWAS Catalogue as specification

Dependencies:

Utilities:

tabix (ml bcftools/1.9)

R packages

  • COLOC

Functions:

runCOLOC.R - run COLOC on with the two P-value distributions

Usage: Rscript scripts/run_COLOC.R [options]


Options:
    -o OUTFOLDER, --outFolder=OUTFOLDER
        the path to the output file

    -g GWAS, --gwas=GWAS
        the dataset ID for a GWAS in the GWAS/QTL database

    -q QTL, --qtl=QTL
        the dataset ID for a QTL dataset in the GWAS/QTL database

    --debug
        load all files and the nsave RData without running COLOC

    -h, --help
        Show this help message and exit

Putting steps together

  1. Index GWAS

  2. Extract top hits from GWAS

These are now taken care of in the GWAS scripts

  1. For a given GWAS and a top hit variant:
    • extract all nominal QTL associations within 1MB from the hit variant (ExtractNominalQTLs)
    • extract all GWAS stats within the same region for each gene: run COLOC to test for colocalisation with each gene's QTL P-value distribution and the GWAS distribution

LDlinkR

to calculate LD between the lead GWAS and lead QTL SNPs, we use the LDLinkR R package. You need to generate an API access token and add it to your ~/.Renviron LDLINK_TOKEN=" " see here