Commit 3d2a6734 authored by Chris Cheshire's avatar Chris Cheshire
Browse files

Switched to nf-core bedtools genome cov scale

parent 8c9ac3e4
Loading
Loading
Loading
Loading
+3 −0
Original line number Diff line number Diff line
@@ -6,6 +6,9 @@
            "bedtools/bamtobed": {
                "git_sha": "e937c7950af70930d1f34bb961403d9d2aa81c7d"
            },
            "bedtools/genomecov": {
                "git_sha": "e8b33e6eb1be2a03abdd2c06068f9a5cebc02bb4"
            },
            "bedtools/intersect": {
                "git_sha": "49da8642876ae4d91128168cd0db4f1c858d7792"
            },
+0 −38
Original line number Diff line number Diff line
// Import generic module functions
include { initOptions; saveFiles; getSoftwareName } from './common/functions'

params.options = [:]
options        = initOptions(params.options)

process BEDTOOLS_GENOMECOV_SCALE {
    tag "$meta.id"
    label 'process_high'
    publishDir "${params.outdir}",
        mode: params.publish_dir_mode,
        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), publish_id:meta.id) }

    conda (params.enable_conda ? "bioconda::bedtools=2.30.0" : null)
    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
        container "https://depot.galaxyproject.org/singularity/bedtools:2.30.0--hc088bd4_0"
    } else {
        container "quay.io/biocontainers/bedtools:2.30.0--hc088bd4_0"
    }

    input:
    tuple val(meta), path(bam), val(scale)

    output:
    tuple val(meta), path("*.bedGraph"), emit: bedgraph
    path "*.version.txt"               , emit: version

    script:
    def software = getSoftwareName(task.process)
    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"

    """
    bedtools genomecov -ibam $bam $options.args -bg -scale $scale \\
    | bedtools sort > ${prefix}.bedGraph

    bedtools --version | sed -e "s/bedtools v//g" > ${software}.version.txt
    """
}
+78 −0
Original line number Diff line number Diff line
//
//  Utility functions used in nf-core DSL2 module files
//

//
// Extract name of software tool from process name using $task.process
//
def getSoftwareName(task_process) {
    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
}

//
// Extract name of module from process name using $task.process
//
def getProcessName(task_process) {
    return task_process.tokenize(':')[-1]
}

//
// Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
//
def initOptions(Map args) {
    def Map options = [:]
    options.args            = args.args ?: ''
    options.args2           = args.args2 ?: ''
    options.args3           = args.args3 ?: ''
    options.publish_by_meta = args.publish_by_meta ?: []
    options.publish_dir     = args.publish_dir ?: ''
    options.publish_files   = args.publish_files
    options.suffix          = args.suffix ?: ''
    return options
}

//
// Tidy up and join elements of a list to return a path string
//
def getPathFromList(path_list) {
    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
    return paths.join('/')
}

//
// Function to save/publish module results
//
def saveFiles(Map args) {
    def ioptions  = initOptions(args.options)
    def path_list = [ ioptions.publish_dir ?: args.publish_dir ]

    // Do not publish versions.yml unless running from pytest workflow
    if (args.filename.equals('versions.yml') && !System.getenv("NF_CORE_MODULES_TEST")) {
        return null
    }
    if (ioptions.publish_by_meta) {
        def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
        for (key in key_list) {
            if (args.meta && key instanceof String) {
                def path = key
                if (args.meta.containsKey(key)) {
                    path = args.meta[key] instanceof Boolean ? "${key}_${args.meta[key]}".toString() : args.meta[key]
                }
                path = path instanceof String ? path : ''
                path_list.add(path)
            }
        }
    }
    if (ioptions.publish_files instanceof Map) {
        for (ext in ioptions.publish_files) {
            if (args.filename.endsWith(ext.key)) {
                def ext_list = path_list.collect()
                ext_list.add(ext.value)
                return "${getPathFromList(ext_list)}/$args.filename"
            }
        }
    } else if (ioptions.publish_files == null) {
        return "${getPathFromList(path_list)}/$args.filename"
    }
}
+68 −0
Original line number Diff line number Diff line
// Import generic module functions
include { initOptions; saveFiles; getSoftwareName; getProcessName } from './functions'

params.options = [:]
options        = initOptions(params.options)

process BEDTOOLS_GENOMECOV {
    tag "$meta.id"
    label 'process_medium'
    publishDir "${params.outdir}",
        mode: params.publish_dir_mode,
        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), meta:meta, publish_by_meta:['id']) }

    conda (params.enable_conda ? "bioconda::bedtools=2.30.0" : null)
    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
        container "https://depot.galaxyproject.org/singularity/bedtools:2.30.0--hc088bd4_0"
    } else {
        container "quay.io/biocontainers/bedtools:2.30.0--hc088bd4_0"
    }

    input:
    tuple val(meta), path(intervals), val(scale)
    path  sizes
    val   extension

    output:
    tuple val(meta), path("*.${extension}"), emit: genomecov
    path  "versions.yml"                   , emit: versions

    script:
    def prefix     = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
    def args_token = options.args.tokenize()
    def args       = options.args
    args += (scale > 0 && scale != 1) ? " -scale $scale" : ""

    if (!args_token.contains('-bg') && (scale > 0 && scale != 1)) {
        args += " -bg"
    }

    if (intervals.name =~ /\.bam/) {
        """
        bedtools \\
            genomecov \\
            -ibam $intervals \\
            $args \\
            > ${prefix}.${extension}

        cat <<-END_VERSIONS > versions.yml
        ${getProcessName(task.process)}:
            ${getSoftwareName(task.process)}: \$(bedtools --version | sed -e "s/bedtools v//g")
        END_VERSIONS
        """
    } else {
        """
        bedtools \\
            genomecov \\
            -i $intervals \\
            -g $sizes \\
            $args \\
            > ${prefix}.${extension}

        cat <<-END_VERSIONS > versions.yml
        ${getProcessName(task.process)}:
            ${getSoftwareName(task.process)}: \$(bedtools --version | sed -e "s/bedtools v//g")
        END_VERSIONS
        """
    }
}
+51 −0
Original line number Diff line number Diff line
name: bedtools_genomecov
description: Computes histograms (default), per-base reports (-d) and BEDGRAPH (-bg) summaries of feature coverage (e.g., aligned sequences) for a given genome.
keywords:
    - bed
    - bam
    - genomecov
tools:
    - bedtools:
        description: |
            A set of tools for genomic analysis tasks, specifically enabling genome arithmetic (merge, count, complement) on various file types.
        documentation: https://bedtools.readthedocs.io/en/latest/content/tools/genomecov.html
        licence: ['MIT']
input:
    - meta:
        type: map
        description: |
            Groovy Map containing sample information
            e.g. [ id:'test', single_end:false ]
    - intervals:
        type: file
        description: BAM/BED/GFF/VCF
        pattern: "*.{bam|bed|gff|vcf}"
    - scale:
        type: value
        description: Number containing the scale factor for the output. Set to 1 to disable. Setting to a value other than 1 will also get the -bg bedgraph output format as this is required for this command switch
    - sizes:
        type: file
        description: Tab-delimited table of chromosome names in the first column and chromosome sizes in the second column
    - extension:
        type: string
        description: Extension of the output file (e. g., ".bg", ".bedgraph", ".txt", ".tab", etc.) It is set arbitrarily by the user and corresponds to the file format which depends on arguments.
output:
    - meta:
        type: map
        description: |
            Groovy Map containing sample information
            e.g. [ id:'test', single_end:false ]
    - genomecov:
        type: file
        description: Computed genome coverage file
        pattern: "*.${extension}"
    - versions:
        type: file
        description: File containing software versions
        pattern: "versions.yml"
authors:
    - "@Emiller88"
    - "@sruthipsuresh"
    - "@drpatelh"
    - "@sidorov-si"
    - "@chris-cheshire"