Merge pull request #245 from phue/bismark_deduplicate

add bismark/deduplicate + tests

.github/filters.yml

@@ -54,6 +54,10 @@ bedtools_sort:
   - software/bedtools/sort/**
   - tests/software/bedtools/sort/**
 
+bismark_deduplicate:
+  - software/bismark/deduplicate/**
+  - tests/software/bismark/deduplicate/**
+
 bismark_genome_preparation:
   - software/bismark/genome_preparation/**
   - tests/software/bismark/genome_preparation/**

software/bismark/deduplicate/functions.nf

@@ -0,0 +1,59 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args          = args.args ?: ''
+    options.args2         = args.args2 ?: ''
+    options.publish_by_id = args.publish_by_id ?: false
+    options.publish_dir   = args.publish_dir ?: ''
+    options.publish_files = args.publish_files
+    options.suffix        = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }  // Remove empty entries
+    paths = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_id) {
+            path_list.add(args.publish_id)
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

software/bismark/deduplicate/main.nf

@@ -0,0 +1,41 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+def options    = initOptions(params.options)
+
+process BISMARK_DEDUPLICATE {
+    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::bismark==0.23.0" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/bismark:0.23.0--0"
+    } else {
+        container "quay.io/biocontainers/bismark:0.23.0--0"
+    }
+
+    input:
+    tuple val(meta), path(bam)
+
+    output:
+    tuple val(meta), path("*.deduplicated.bam")        , emit: bam
+    tuple val(meta), path("*.deduplication_report.txt"), emit: report
+    path  "*.version.txt"                              , emit: version
+
+    script:
+    def software   = getSoftwareName(task.process)
+    def prefix     = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    def seqtype    = meta.single_end ? '-s' : '-p'
+    """
+    deduplicate_bismark \\
+        $options.args \\
+        $seqtype \\
+        --bam $bam
+
+    echo \$(bismark -v 2>&1) | sed 's/^.*Bismark Version: v//; s/Copyright.*\$//' > ${software}.version.txt
+    """
+}

software/bismark/deduplicate/meta.yml

@@ -0,0 +1,72 @@
+name: bismark_deduplicate
+description: |
+    Removes alignments to the same position in the genome
+    from the Bismark mapping output.
+keywords:
+    - bismark
+    - 3-letter genome
+    - map
+    - methylation
+    - 5mC
+    - methylseq
+    - bisulphite
+    - bam
+tools:
+    - bismark:
+        description: |
+            Bismark is a tool to map bisulfite treated sequencing reads
+            and perform methylation calling in a quick and easy-to-use fashion.
+        homepage: https://github.com/FelixKrueger/Bismark
+        documentation: https://github.com/FelixKrueger/Bismark/tree/master/Docs
+        doi: 10.1093/bioinformatics/btr167
+params:
+    - outdir:
+        type: string
+        description: |
+            The pipeline's output directory. By default, the module will
+            output files into `$params.outdir/<SOFTWARE>`
+    - publish_dir_mode:
+        type: string
+        description: |
+            Value for the Nextflow `publishDir` mode parameter.
+            Available: symlink, rellink, link, copy, copyNoFollow, move.
+    - enable_conda:
+        type: boolean
+        description: |
+            Run the module with Conda using the software specified
+            via the `conda` directive
+    - singularity_pull_docker_container:
+        type: boolean
+        description: |
+            Instead of directly downloading Singularity images for use with Singularity,
+            force the workflow to pull and convert Docker containers instead.
+input:
+    - meta:
+        type: map
+        description: |
+            Groovy Map containing sample information
+            e.g. [ id:'test', single_end:false ]
+    - bam:
+        type: file
+        description: BAM file containing read alignments
+        pattern: "*.{bam}"
+output:
+    - meta:
+        type: map
+        description: |
+            Groovy Map containing sample information
+            e.g. [ id:'test', single_end:false ]
+    - bam:
+        type: file
+        description: Deduplicated output BAM file containing read alignments
+        pattern: "*.{deduplicated.bam}"
+    - report:
+        type: file
+        description: Bismark deduplication reports
+        pattern: "*.{deduplication_report.txt}"
+    - version:
+        type: file
+        description: File containing software version
+        pattern: "*.{version.txt}"
+authors:
+    - "@phue"

tests/software/bismark/deduplicate/main.nf

@@ -0,0 +1,14 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { BISMARK_DEDUPLICATE } from '../../../../software/bismark/deduplicate/main.nf' addParams( options: [:] )
+
+workflow test_bismark_deduplicate {
+
+    def input = []
+    input = [ [ id:'test', single_end:false ], // meta map
+              [ file("${launchDir}/tests/data/bismark/Ecoli_10K_methylated_R1_bismark_bt2_pe.bam", checkIfExists: true) ] ]
+
+    BISMARK_DEDUPLICATE ( input )
+}

tests/software/bismark/deduplicate/test.yml

@@ -0,0 +1,10 @@
+- name: Run bismark deduplicate test workflow
+  command: nextflow run ./tests/software/bismark/deduplicate -entry test_bismark_deduplicate -c tests/config/nextflow.config
+  tags:
+    - bismark
+    - bismark_deduplicate
+  files:
+    - path: output/bismark/Ecoli_10K_methylated_R1_bismark_bt2_pe.deduplicated.bam
+      md5sum: 35794f413740d9097010e8411f03b785
+    - path: output/bismark/Ecoli_10K_methylated_R1_bismark_bt2_pe.deduplication_report.txt
+      md5sum: 125b01ad4f1784a77b7802667d4bfee7