VEP/tasks/VEP.wdl

task VEP {

  File vcf
  String sample_id
  String basename = basename(vcf,".vcf")
  String tumor_id
  String normal_id
  File ref_dir
  String fasta
  String vep_path
  File cache
  String ncbi_build
  String species
  String vcf2maf_path
  String docker
  String cluster_config
  String disk_size


  command <<<
    set -o pipefail
    set -e
    nt=$(nproc)

    source /etc/profile

    awk -F'\t' '{if(($1~"^#")||($1!~"^#" && $7=="PASS")){print $0}}' ${vcf} > ${sample_id}.vcf
    
    # Judge the SAMPLE info of vcf file
    ncol=`awk -F'\t' '{if($1!~"^#"){print NF}}' ${sample_id}.vcf | uniq`
    if [ $ncol -lt 11 ]; then
      SAMPLE="--tumor-id ${tumor_id} --normal-id ${normal_id}"
    else
      SAMPLE="--tumor-id ${sample_id}"
    fi

    # Set the buffer_size based on the data size
    nrow=`awk -F'\t' '{if($1~"^chr"){print $0}}' ${sample_id}.vcf | wc -l`
    if [ $nrow -lt 5000 ]; then
      buffer_size="--buffer_size 5000"
    else
      buffer_size="--buffer_size 1000"
    fi

    # Extract the BND variants from VCF
    # awk -F'\t' '{if(($1~"^#")||($8!~".*SVTYPE=BND.*")){print $0}}' ${sample_id}.PASS.vcf > ${sample_id}.PASS.vcf2maf.vcf
    # awk -F'\t' '{if(($1~"^#")||($8~".*SVTYPE=BND.*")){print $0}}' ${sample_id}.PASS.vcf > ${sample_id}.INPUT.VEP.vcf

    # vcf2maf
    # perl ${vcf2maf_path}/vcf2maf.pl \
    # --input-vcf ${sample_id}.PASS.vcf2maf.vcf --output-maf ${basename}.maf \
    # --tumor-id ${tumor_id} --normal-id ${normal_id} \
    # --ref-fasta ${ref_dir}/${fasta} \
    # --vep-path ${vep_path} \
    # --vep-data ${cache} \
    # --ncbi-build ${ncbi_build} \
    # --species ${species} \
    # --vep-fork $nt

    # vep
    # perl ${vep_path}/vep \
    # --input_file ${sample_id}.vcf --output_file ${basename}.PASS.vep.vcf \
    # --fasta ${ref_dir}/${fasta} \
    # --dir ${cache} \
    # --assembly ${ncbi_build} \
    # --species ${species} \
    # --fork $nt \
    # --format vcf --vcf \
    # --no_progress \
    # --no_stats \
    # $buffer_size \
    # --sift b \
    # --ccds --uniprot --hgvs --symbol --numbers --domains --gene_phenotype --canonical --protein --biotype --uniprot --tsl --variant_class --shift_hgvs 1 --check_existing --total_length --allele_number --no_escape --xref_refseq --failed 1 --flag_pick_allele --pick_order canonical,tsl,biotype,rank,ccds,length --force_overwrite --offline --pubmed --regulatory
    
    # vcf2vcf: transfer into a standardized format
    perl ${vcf2maf_path}/vcf2vcf.pl \
    --input-vcf ${sample_id}.vcf --output-vcf ${basename}.norm.vcf \
    $SAMPLE \
    --ref-fasta ${ref_dir}/${fasta}

    # VEP annotation
    perl ${vep_path}/vep --format vcf --vcf \
    --assembly ${ncbi_build} \
    --species ${species} \
    --everything --af_exac \
    --offline \
    --cache --dir_cache ${cache} \
    --fasta ${ref_dir}/${fasta} \
    $buffer_size \
    --input_file ${basename}.norm.vcf --output_file ${basename}.vep.vcf

    # vcf2maf
    perl ${vcf2maf_path}/vcf2maf.pl \
    --inhibit-vep \
    --input-vcf ${basename}.vep.vcf --output-maf ${basename}.maf \
    $SAMPLE \
    --ref-fasta ${ref_dir}/${fasta} \
    --ncbi-build ${ncbi_build} \
    --species ${species} \
    --vep-fork $nt
  >>>
  
  runtime {
    docker: docker
    cluster: cluster_config
    systemDisk: "cloud_ssd 40"
    dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"
  }

  output {
    File vep_vcf = "${basename}.vep.vcf"
    File maf = "${basename}.maf"
  }
}