GenOMICC_StrepA

Install the project dependencies

• Install Julia

• Install Julia dependencies:

julia --project=. --startup-file=no -e'using Pkg; Pkg.instantiate()'

Imputing EGAD00001004558

We integrate data from this study. In principle, this has already been done and the imputed genotypes can be downloaded from ODAP (ask admin member) or accessed from datastore at /exports/cmvm/datastore/eb/groups/baillielab2/strepA_EGAD00001004558 ().

Here are the steps that were performed to obtain these imputed genotypes. We assume the source data is located in the following directory:

EGAD_DATA_DIR=/gpfs/igmmfs01/eddie/ISARIC4C/olivier/data/EGAD00001004558
KGP_DATA_DIR=/home/olabayle/isaric/olivier/data/kgp-merged-unrelated-or3
mkdir imputation_work
mkdir imputation_work/mappings
# Get sexes
awk -F',' 'NR>1 {print $6 "\t" $6 "\t" $5}' $EGAD_DATA_DIR/GSA2024_1140_DATA/samples.csv > imputation_work/sexes.tsv
# Make BED file for: STREPGENE_HG19_GSA
plink2 \
--vcf $EGAD_DATA_DIR/EGAF00002318977/STREPGENE_HG19_GSA.vcf \
--id-delim _ \
--update-sex imputation_work/sexes.tsv \
--make-bed \
--out imputation_work/STREPGENE_HG19_GSA
plink2 \
--bfile imputation_work/STREPGENE_HG19_GSA \
--freq \
--out imputation_work/STREPGENE_HG19_GSA
# Make file to map to GRCh38
awk '{
  chr=$1
  pos=$4
  start=pos-1
  end=pos
  print "chr"chr"\t"start"\t"end"\t"$2
}' imputation_work/STREPGENE_HG19_GSA.bim > imputation_work/STREPGENE_HG19_GSA.variants.GRCh37.bed

# Make BED file for: STREPGENE_HG19_GSA
plink2 \
--vcf $EGAD_DATA_DIR/EGAF00002318978/STREPGENE_HG19_HC24.vcf \
--id-delim _ \
--update-sex imputation_work/sexes.tsv \
--make-bed \
--out imputation_work/STREPGENE_HG19_HC24
plink2 \
--bfile imputation_work/STREPGENE_HG19_HC24 \
--freq \
--out imputation_work/STREPGENE_HG19_HC24
# Make file to map to GRCh38
awk '{
  chr=$1
  pos=$4
  start=pos-1
  end=pos
  print "chr"chr"\t"start"\t"end"\t"$2
}' imputation_work/STREPGENE_HG19_HC24.bim > imputation_work/STREPGENE_HG19_HC24.variants.GRCh37.bed

Now go on ucsc and lift over the variants.GRCh37.bed files to GRCh38. We can update the coordinates in the bim file using the bin/update_coordinates.jl:

julia --project=. --startup-file=no bin/update_coordinates.jl imputation_work/STREPGENE_HG19_GSA imputation_work/mappings/STREPGENE_HG19_GSA.variants.GRCh38.bed $KGP_DATA_DIR/kgp.merged.unrelated.afreq
plink2 --bfile imputation_work/STREPGENE_HG19_GSA.qced --make-bed --set-all-var-ids @:#:\$r:\$a --chr 1-22,X --freq --out imputation_work/STREPGENE_HG19_GSA.qced.ready
# This is because TOPMed requires at least 20 samples to proceed
cp imputation_work/STREPGENE_HG19_GSA.qced.ready.bed imputation_work/STREPGENE_HG19_GSA.qced.ready.bis.bed
cp imputation_work/STREPGENE_HG19_GSA.qced.ready.bim imputation_work/STREPGENE_HG19_GSA.qced.ready.bis.bim
cp imputation_work/STREPGENE_HG19_GSA.qced.ready.fam imputation_work/STREPGENE_HG19_GSA.qced.ready.bis.fam
awk 'BEGIN{OFS="\t"} {$1 = $1 "_copy"; $2 = $2 "_copy"; print}' imputation_work/STREPGENE_HG19_GSA.qced.ready.bis.fam > imputation_work/STREPGENE_HG19_GSA.qced.ready.bis.temp.fam
mv imputation_work/STREPGENE_HG19_GSA.qced.ready.bis.temp.fam imputation_work/STREPGENE_HG19_GSA.qced.ready.bis.fam
plink -bfile imputation_work/STREPGENE_HG19_GSA.qced.ready -bmerge imputation_work/STREPGENE_HG19_GSA.qced.ready.bis --make-bed --out imputation_work/STREPGENE_HG19_GSA.qced.ready.duplicated

julia --project=. --startup-file=no bin/update_coordinates.jl imputation_work/STREPGENE_HG19_HC24 imputation_work/mappings/STREPGENE_HG19_HC24.variants.GRCh38.bed $KGP_DATA_DIR/kgp.merged.unrelated.afreq
plink2 --bfile imputation_work/STREPGENE_HG19_HC24.qced --make-bed --set-all-var-ids @:#:\$r:\$a --chr 1-22,X --freq --out imputation_work/STREPGENE_HG19_HC24.qced.ready

Then open a persistent session and run:

nextflow run nf-topmed-imputation/main.nf -profile eddie -resume -with-report -with-trace -c config/imputation_STREPGENE_HG19_GSA.config

and

nextflow run nf-topmed-imputation/main.nf -profile eddie -resume -with-report -with-trace -c config/imputation_STREPGENE_HG19_HC24.config

to impute the genotypes.

Finally, the resulting genotypes can be merged via:

julia --project=. --startup-file=no bin/merge_gsa_hc24.jl results_GSA/ results_HC24/

Merging EGAD00001004558 with GenOMICC

Reproducing The GWAS

To run the GWAS, we only need the WDL-GWAS pipeline, make sure you have followed the installation isntructions and run the following:

input_config=config/genomicc-inputs.strepA.meta.plink2.json
compiled_input_config=config/genomicc-inputs.strepA.meta.plink2.dx.json
output_dir=/or3_strepa_plink2
commit=98c732469053100dff96cf3b2a4ebb9c70fafa43

# Clone WDL-GWAS pipeline
git clone https://github.com/olivierlabayle/WDL-GWAS
cd WDL-GWAS && git checkout $commit && cd ..

# Compile the workflow
java -jar $DX_COMPILER_PATH compile WDL-GWAS/workflows/gwas.wdl \
    -f -project $RAP_PROJECT_ID \
    -extras WDL-GWAS/config/extras.json \
    -reorg \
    -folder /workflows/gwas \
    -inputs ${input_config}

# Run the workflow
dx run -y \
    -f ${compiled_input_config} \
    --priority high \
    --preserve-job-outputs \
    --destination ${output_dir} \
    /workflows/gwas/gwas

Debugging a Run With A DNA Nexus Instance

To connect to a DNA Nexus instance, run:

instance_type=mem2_ssd1_v2_x8
dx run \
    --instance-type $instance_type \
    -imax_session_length="48h" \
    -y \
    --ssh app-cloud_workstation

You can download files via:

dx download file_id_1 file_id_2 ... 

And run the container with:

docker_tag=optional_fp
docker run -it --rm -v $PWD:/mnt/data olivierlabayle/wdl-gwas:$docker_tag /bin/bash

Then you can run code as normal, for instance to enter the Julia REPL:

julia --project=/opt/PopGen --startup-file=no --sysimage=/opt/PopGen/sysimage.so --threads=auto