nextflow.config

params.qc = "fastp" // Available Options: fastp, TrimGalore, FastQC (requires own trimming)
params.mapping = "star" // Available Options: hisat2, star
params.assembly = ["Trinity","rnaspades"] // Available Options (choose multiple): STAR, Trinity, RNAspades,
params.quantification = "featureCounts" // Available Options: featureCounts (none other available at the moment)
params.trinity_type="guided" // Available options: guided, denovo

// Whether or not to publish directories to allow them to be viewed (mainly for debugging files)
PUBLISH_DIRECTORIES = 0

                                                      // Set some global variables which require linux-based commands. Can also set manually if these are known prior
params {
    USER_ID        = "id -u".execute().text.trim()    // current user ID needed for mapping directories
    GROUP_ID       = "id -g".execute().text.trim()    // current group ID needed for mapping directories

    input_dir      = '';                              // (Top-level) folder in which reads are kept
    output_dir     = "./workflow_output";             // Where to publish files i.e VCFs, CSV's
    reference_dir  = "";                              // Where to publish files i.e VCFs, CSV's
    reference_file = '${reference_dir}/*.{fa,fasta}'; // The reference genome
    gff_file       = '${reference_dir}/*.{}';         // The annotation for the reference genome
    repeats        = '';
    mirna_mature   = '';


    num_threads    = 4;
    max_memory     = '50.GB';

    star_index   = '';
    hisat_index   = '';
    known_sites    = '';

    star_two_pass  = false;                           // Useful for splice/isoform analysis, otherwise no

    // Either give samplesheet or define r1/r2 pattern
    paired         = false;                           // Whether paired data is given
    samplesheet    = null;
    strandedness   = "reverse"
    r1_pattern     = "_R1"                            // Pattern for Read1 files (for paired-end)
    r2_pattern     = "_R2"                            // Pattern for Read2 files (for paired-end)
}

// https://github.com/nextflow-io/nextflow/issues/2786
profiles {
    singularity {
        dag.overwrite = true
        singularity.enabled      = true
        singularity.autoMounts   = true
        singularity.runOptions   = '-B $SINGULARITY_TMPDIR:/tmp --no-mount /nfs/wsi/it/posth' // -B $SINGULARITYENV_TMPDIR:/tmp
        singularity.envWhitelist = ['SINGULARITY_TMPDIR', 'APPTAINERENV_TMPDIR', 'SINGULARITYENV_TMPDIR']
        executor {
            name                 = 'slurm'
            queueSize            = 20 // This multiplied by `cpus` could control max #CPUs at one time
            pollInterval         = '30 sec'
            submitRateLimit      = '50/1min'
        }

        process {
            // Labels to take into consideration: storeDir

            container            = 'file://clint.sif'
            cpus                 = 5         // General amount for CPUs
            cache                = 'lenient' // Important for shared file systems
            queue                = 'short'   // default queue - usually one of the shorter ones

            withLabel: preprocess {
                // maxForks      = 10 // If not SLURM executor
                queue            = 'short'
                memory           = 10.GB
            }
            withLabel: mapping {
                // maxForks      = 5 // If not SLURM executor
                queue            = 'short'
                memory           = 20.GB
            }
            withLabel: assembly {
                // maxForks      = 6 // If not SLURM executor
                queue            = 'long'
                memory           = 50.GB
            }
            withLabel: variant_calling {
                errorStrategy    = 'retry' // For gatk-related processes
                maxRetries       = 3          // For gatk-related processes
                queue            = 'long'
                memory           = 40.GB
            }

            // These are processes such as Mutect2 which due to
            // the inverval-scattering approach create  lot of
            // forks - limit these
            withLabel: forking_heavy {
                maxForks         = 15
            }
            withName: 'Mutect2' {
                maxForks         = 15
            }

            // This should be done quickly and once (only) hopefully
            withName: STAR_BUILD {
                cpus             = 10
                memory           = 50.GB
            }

            withName: STAR_CIRCRNA_CIRCTOOLS {
                queue = "long"
            }

            // These dont need that much regarding cpu usage
            withName: 'SAMTOOLS_INDEX|SAMTOOLS_SORT' {
                cpus             = 2
                memory           = 50.GB
            }

            withName: DCC {
                cpus = 8
                memory = 20.GB
            }

            withName: ANNOTATE {
                container = ""
            }

        }
    }

    docker {
        // Causes errors when doing 'publishDir' related stuff due to permissions
        docker.enabled           = true
        docker.temp              = 'auto'

        process {
            container            = 'file://clint.sif'
            cpus                 = 4 // Default #cpus to use
            withName: HISAT2 {
                cpus             = 6
            }
        }
    }
}