variant-calling/tasks/BQSR.wdl

task BQSR {
  
  File ref_dir
  File dbsnp_dir
  File dbmills_dir
  
  String sample
  String SENTIEON_INSTALL_DIR
  String SENTIEON_LICENSE
  String fasta
  String dbsnp
  String db_mills
  File realigned_bam
  File realigned_bam_index
  String docker
  String cluster_config
  String disk_size

  command <<<
    set -o pipefail
    set -e
    export SENTIEON_LICENSE=${SENTIEON_LICENSE}
    nt=$(nproc)

    ${SENTIEON_INSTALL_DIR}/bin/sentieon driver -t $nt \
    -r ${ref_dir}/${fasta} -i ${realigned_bam} \
    --algo QualCal \
    -k ${dbsnp_dir}/${dbsnp} -k ${dbmills_dir}/${db_mills} \
    ${sample}_recal_data.table

    ${SENTIEON_INSTALL_DIR}/bin/sentieon driver -t $nt \
    -r ${ref_dir}/${fasta} -i ${realigned_bam} -q ${sample}_recal_data.table \
    --algo QualCal \
    -k ${dbsnp_dir}/${dbsnp} -k ${dbmills_dir}/${db_mills} \
    ${sample}_recal_data.table.post \
    --algo ReadWriter ${sample}.sorted.deduped.realigned.recaled.bam
    
    ${SENTIEON_INSTALL_DIR}/bin/sentieon driver -t $nt --algo QualCal --plot \
    --before ${sample}_recal_data.table --after ${sample}_recal_data.table.post ${sample}_recal_result.csv
    
    ${SENTIEON_INSTALL_DIR}/bin/sentieon plot QualCal -o ${sample}_bqsr.pdf ${sample}_recal_result.csv
  >>>

  runtime {
    docker: docker
    cluster: cluster_config
    systemDisk: "cloud_ssd 40"
    dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"  
  }

  output {
    File recal_table = "${sample}_recal_data.table"
    File recal_post = "${sample}_recal_data.table.post"
    File recaled_bam = "${sample}.sorted.deduped.realigned.recaled.bam"
    File recaled_bam_index = "${sample}.sorted.deduped.realigned.recaled.bam.bai"
    File recal_csv = "${sample}_recal_result.csv"
    File bqsrreport_pdf = "${sample}_bqsr.pdf"
  }
}