UKBB_CLUMP

A Snakemake pipeline for clumping GWAS results from the UK-Biobank using plink. First, GWAS results are downloaded from the UK-Biobank. After some formatting steps, plink --clump is used in order to find a set of independent and significantly associated variants (index variants). Finally, allele counts of the index variants are extracted for all individuals found in the plink-binary genotype data.

Usage

For use within the lab you can activate the conda environment prs:

conda activate prs

or use the environment.yaml to create a new conda environment with the necessary dependencies.

Clumping parameters, the location of the plink2 binary genotype data and the output directory are specified in the config.yaml.

The phenotype_annotation csv-file specified in the config.yaml lists all phenotypes and GWAS-results which are to be clumped: The column phenotype contains the name of the trait and the column GWAS_result_file the filename of the GWAS-results as specified in the UKBB manifest in the field filename (column BW).

Call the pipeline as follows to run the clumping for all traits listed in the phenotype_annotation csv-file:

snakemake -c16

The runtime of the pipeline strongly depends on the number of variants in the GWAS results that have a p-value below the p_threshold.

See ./notebooks/gene_var_intersect.ipynb for how to read the allele counts for some given gene ID after running the pipeline.