variant-calling/tasks/ANNOVAR.wdl

task ANNOVAR {

  File vcf
  String basename = basename(vcf,".vcf")
  String hg
  File annovar_database
  Boolean only_pass
  String docker
  String cluster_config
  String disk_size


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

    if [ only_pass ]; then
      awk -F'\t' '{if(($1~"^#")||($1!~"^#" && $7=="PASS")){print $0}}' ${vcf} > INPUT.vcf
    else
      cp ${vcf} INPUT.vcf
    fi

    if [ ${hg} == "hg38" ]; then
      /installations/annovar/table_annovar.pl INPUT.vcf ${annovar_database} -buildver ${hg} -out ${basename} -remove -protocol refGene,cytoBand,genomicSuperDups,ljb26_all,dbnsfp35c,intervar_20180118,cosmic70,exac03,gnomad211_exome,clinvar_20200316 -operation g,r,r,f,f,f,f,f,f,f -nastring . -vcfinput -thread $nt
    fi

    if [ ${hg} == "hg19" ]; then
      /installations/annovar/table_annovar.pl INPUT.vcf ${annovar_database} -buildver ${hg} -out ${basename} -remove -protocol refGene,cytoBand,genomicSuperDups,ljb26_all,snp138,cosmic78,intervar_20170202,popfreq_all_20150413,clinvar_20190305 -operation g,r,r,f,f,f,f,f,f -nastring . -vcfinput -thread $nt
    fi
  >>>
  
  runtime {
    docker: docker
    cluster: cluster_config
    systemDisk: "cloud_ssd 40"
    dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"
  }

  output {
    File avinput = "${basename}.avinput"
    File multianno_txt = "${basename}.${hg}_multianno.txt"
    File multianno_vcf = "${basename}.${hg}_multianno.vcf"
  }
}