wgs_somatic_pipeline_v1/tasks/BQSR.wdl

task Sentieon_BQSR{
    File ref_dir
    File dbsnp_dir
    File dbmills_dir

    String sample_id
    String ref_fasta
    String dbsnp
    String db_mills

    File deduped_bam
    File deduped_bam_index

    File? regions
    Int? interval_padding

    # excute env
    String docker
    String cluster_config
    String disk_size
    
    String SENTIEON_LICENSE

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


        sentieon driver -t $nt \
        -r ${ref_dir}/${ref_fasta} -i ${deduped_bam} \
        $INTERVAL \
        --algo QualCal \
        -k ${dbsnp_dir}/${dbsnp} -k ${dbmills_dir}/${db_mills} \
        ${sample_id}_recal_data.table

        sentieon driver -t $nt \
        -r ${ref_dir}/${ref_fasta} -i ${deduped_bam} \
        -q ${sample_id}_recal_data.table \
        --algo QualCal \
        -k ${dbsnp_dir}/${dbsnp} -k ${dbmills_dir}/${db_mills} \
        ${sample_id}_recal_data.table.post \
        --algo ReadWriter ${sample_id}.sorted.deduped.recaled.bam

        sentieon driver -t $nt --algo QualCal \
        --plot --before ${sample_id}_recal_data.table --after ${sample_id}_recal_data.table.post ${sample_id}_recal_data.csv

        sentieon plot bqsr -o ${sample_id}_bqsrreport.pdf ${sample_id}_recal_data.csv
    >>>

    runtime{
        docker:docker
        cluster:cluster_config
        systemDisk:"cloud_ssd 250"
        dataDisk:"cloud_ssd " + disk_size + " /cromwell_root/"


    }

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

    }
}