eQTL-Catalogue/genimpute workflow

Genotype imputation and quality control workflow used by the eQTL Catalogue.

Performs the following main steps:

Pre-imputation QC:

• Convert raw array genotypes to GRCh38 coordinates with CrossMap.py v0.4.1
• Align array genotypes to the 1000 Genomes 30x on GRCh38 reference panel with Genotype Harmonizer.
• Convert the genotypes to the VCF format with PLINK.
• Exclude variants with Hardy-Weinberg p-value < 1e-6, missingness > 0.05 and minor allele frequency < 0.01 with bcftools
• Calculate individual-level missingness using vcftools.

Imputation:

• Genotype pre-phasing with Egale 2.4.1
• Genotype imputation with Minimac4

Post-imputation QC:

• Exclude variants with imputation R2 < 0.4
• Keep variants on chromosomes 1-22 and X
• Keep variants with MAF > 0.01
• Multiply genotype dosage of male samples on the Non-PAR region of the X chromsome by two for easier QTL mapping

Input parameters

--bfile

Raw genotypes in PLINK format (bed, bim, fam). Assumed to be in GRCh37 coordinates. Genotypes in VCF format can be converted to PLINK format with:

plink --vcf <path_to_vcf_file> --make-bed --out <plink_file_prefix>

The PAR and non-PAR regions of the X chromosome should be merged together and the name of the X chromsome should be 'X'. This can be achieved with PLINK:

plink --bfile Young_2019 --merge-x --make-bed --output-chr MT --out Young_2019_mergedX

Example command

Imputing genotypes from the open access CEDAR dataset.

nextflow run main.nf \
  -profile eqtl_catalogue -resume\
  --bfile plink_genimpute/CEDAR\
  --output_name CEDAR\
  --outdir CEDAR\
  --impute_PAR true\
  --impute_non_PAR true

Contributors

• Ralf Tambets
• Kaur Alasoo
• Liina Anette Pärtel
• Mark-Erik Kodar