kn99_haploid.config

params {
    config_profile_description = 'Washington University, St Louis Center for Genome Sciences HTCF profile.'
    config_profile_contact = 'Chase Mateusiak (chasem@wustl.edu)'
    config_profile_url = "https://htcf.wustl.edu/docs/"

    snpeff_config_key  = 'ASM221672v1'
    snpeff_db          = "${projectDir}/assets/genome_files/snpeff_db"
    snpeff_db_config   = "${projectDir}/assets/genome_files/snpEff.config"
    fasta              = "https://raw.githubusercontent.com/BrentLab/callvariants/main/assets/genome_files/KN99_genome_fungidb.fasta"
    cnvpytor_conf_file = "https://github.com/BrentLab/callvariants/raw/main/assets/genome_files/kn99_cnvpytor_haploid_genome_conf.py"
    cnvpytor_gc_file   = "https://github.com/BrentLab/callvariants/raw/main/assets/genome_files/kn99_with_markers_gc.pytor"
    region_bed_mask    = "https://github.com/BrentLab/callvariants/raw/main/assets/genome_files/dustmasked.KN99.bed"

}

process{

    // set variables here that are used in multiple places in the
    // configurations below
    def CNV_BINS = "100 1000 10000 100000"
    def MIN_CNV_DEPTH_DUP = 1.7
    def MAX_CNV_DEPTH_DEL = 0.1
    def TIDDIT_COV_BIN = "10000"
    def PLOIDY = "1"
    def MIN_MAPQ = "20"
    def MIN_BASE_QUAL = "10"
    def MIN_ALTERNATE_FRACTION = "0.75"
    def MIN_COVERAGE = "10"

    withName: PICARD_MARKDUPLICATES {
        cpus   = { check_max( 1     * task.attempt, 'cpus'    ) }
        memory = { check_max( 10.GB * task.attempt, 'memory'  ) }
        time   = { check_max( 1.h   * task.attempt, 'time'    ) }
    }

    withName: CNVPYTOR_CSV_RAW {
        publishDir = [
            path: { "${params.outdir}/variants/cnvpytor" },
            mode: params.publish_dir_mode,
            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
        ]

        ext.args =[
            bins: CNV_BINS
        ]
    }

    withName: CNVPYTOR_CSV_FLTR {
        publishDir = [
            path: { "${params.outdir}/variants/cnvpytor" },
            mode: params.publish_dir_mode,
            saveAs: { filename -> filename.equals('versions.yml') ? null : filename },
            pattern: "*csv"
        ]

        ext.args =[
            bins: CNV_BINS,
            min_cnv_depth_dup: MIN_CNV_DEPTH_DUP,
            max_cnv_depth_del: MAX_CNV_DEPTH_DEL
        ]
    }

    withName: TIDDIT_COV {
        ext.args = [
            "-z " + TIDDIT_COV_BIN,
        ].join(' ').trim()

        ext.prefix = {"${meta.id}_tiddit_" + TIDDIT_COV_BIN}
    }

    withName: TIDDIT_SV {

        ext.args = [
            "-n " + PLOIDY,
            "-q 20",
            "-c 10",
            "--p_ratio 0.3",
            "--r_ratio 0.3"
        ].join(' ').trim()

    }

    // note that this could be split into FREEBAYES_INDIVIDUAL and
    // FREEBAYES_JOINT if you wish to configure them separately
    withName: '.*FREEBAYES_.*' {
        cpus   = { check_max( 1     * task.attempt, 'cpus'    ) }
        memory = { check_max( 5.GB * task.attempt, 'memory'  ) }
        time   = { check_max( 10.h   * task.attempt, 'time'    ) }

        ext.args = [
            "--genotype-qualities",
            "--ploidy " + PLOIDY,
            "--min-mapping-quality " + MIN_MAPQ,
            "--min-base-quality " + MIN_BASE_QUAL,
            "--min-alternate-fraction " + MIN_ALTERNATE_FRACTION,
            "--min-coverage " + MIN_COVERAGE
        ].join(' ').trim()

    }

    withName: '.*:SNPEFF.*' {

        ext.args = [
            "-v ${params.snpeff_config_key}"
            ].join(' ').trim()
    }

   // currently, both joint and individual vcfs go through this one filter
   // step. It may make sense to split the VCFTOOLS filter step into INDIVIDUAL
   // and JOINT at
   withName: VCFTOOLS {

        ext.args = [
            "--exclude-bed",
            "--recode",
            "--recode-INFO-all",
            "--minQ " + MIN_MAPQ,
            "--minDP " + MIN_COVERAGE
        ].join(' ').trim()

   }
}