Merge pull request #219 from phue/qualimap_bamqc

add qualimap/bamqc + test

.github/filters.yml

@@ -189,6 +189,10 @@ preseq_lcextrap:
   - software/preseq/lcextrap/**
   - tests/software/preseq/lcextrap/**
 
+qualimap_bamqc:
+  - software/qualimap/bamqc/**
+  - tests/software/qualimap/bamqc/**
+
 quast:
   - software/quast/**
   - tests/software/quast/**

software/qualimap/bamqc/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/qualimap/bamqc/main.nf

@@ -0,0 +1,61 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+def options    = initOptions(params.options)
+
+process QUALIMAP_BAMQC {
+    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), publish_id:meta.id) }
+
+    conda (params.enable_conda ? "bioconda::qualimap=2.2.2d" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/qualimap:2.2.2d--1"
+    } else {
+        container "quay.io/biocontainers/qualimap:2.2.2d--1"
+    }
+
+    input:
+    tuple val(meta), path(bam)
+    path gff
+    val use_gff
+
+    output:
+    tuple val(meta), path("${prefix}"), emit: results
+    path  "*.version.txt"             , emit: version
+
+    script:
+    def software   = getSoftwareName(task.process)
+    prefix         = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+
+    def collect_pairs = meta.single_end ? '' : '--collect-overlap-pairs'
+    def memory     = task.memory.toGiga() + "G"
+    def regions = use_gff ? "--gff $gff" : ''
+
+    def strandedness = 'non-strand-specific'
+    if (meta.strandedness == 'forward') {
+        strandedness = 'strand-specific-forward'
+    } else if (meta.strandedness == 'reverse') {
+        strandedness = 'strand-specific-reverse'
+    }
+    """
+    unset DISPLAY
+    mkdir tmp
+    export _JAVA_OPTIONS=-Djava.io.tmpdir=./tmp
+    qualimap \\
+        --java-mem-size=$memory \\
+        bamqc \\
+        $options.args \\
+        -bam $bam \\
+        $regions \\
+        -p $strandedness \\
+        $collect_pairs \\
+        -outdir $prefix \\
+        -nt $task.cpus
+
+    echo \$(qualimap 2>&1) | sed 's/^.*QualiMap v.//; s/Built.*\$//' > ${software}.version.txt
+    """
+}

software/qualimap/bamqc/meta.yml

@@ -0,0 +1,70 @@
+name: qualimap_bamqc
+description: Evaluate alignment data
+keywords:
+    - quality control
+    - qc
+    - bam
+tools:
+    - qualimap:
+        description: |
+          Qualimap 2 is a platform-independent application written in
+          Java and R that provides both a Graphical User Interface and
+          a command-line interface to facilitate the quality control of
+          alignment sequencing data and its derivatives like feature counts.
+        homepage: http://qualimap.bioinfo.cipf.es/
+        documentation: http://qualimap.conesalab.org/doc_html/index.html
+        doi: 10.1093/bioinformatics/bts503
+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
+        pattern: "*.{bam}"
+    - gff:
+        type: file
+        description: Feature file with regions of interest
+        pattern: "*.{gff,gtf,bed}"
+    - use_gff:
+        type: boolean
+        description: Specifies if feature file should be used or not
+output:
+    - meta:
+        type: map
+        description: |
+            Groovy Map containing sample information
+            e.g. [ id:'test', single_end:false ]
+    - results:
+        type: dir
+        description: Qualimap results dir
+        pattern: "*/*"
+    - version:
+        type: file
+        description: File containing software version
+        pattern: "*.{version.txt}"
+authors:
+    - "@phue"

tests/software/qualimap/bamqc/main.nf

@@ -0,0 +1,16 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { QUALIMAP_BAMQC } from '../../../../software/qualimap/bamqc/main.nf' addParams( options: [:] )
+
+workflow test_qualimap_bamqc {
+
+    def input = []
+    input = [ [ id:'test', single_end:false ], // meta map
+              file("${launchDir}/tests/data/bam/test.paired_end.sorted.bam", checkIfExists: true) ]
+    def gff = file("dummy_file.txt")
+    def use_gff = false
+
+    QUALIMAP_BAMQC ( input, gff, use_gff )
+}

tests/software/qualimap/bamqc/test.yml

@@ -0,0 +1,26 @@
+- name: Run qualimap bamqc test workflow
+  command: nextflow run ./tests/software/qualimap/bamqc -entry test_qualimap_bamqc -c tests/config/nextflow.config
+  tags:
+    - qualimap
+    - qualimap_bamqc
+  files:
+    - path: output/qualimap/test/raw_data_qualimapReport/coverage_across_reference.txt
+      md5sum: 0e2ae3deae8adbe6c66db4dadb260c76
+    - path: output/qualimap/test/raw_data_qualimapReport/coverage_histogram.txt
+      md5sum: 0e9f37aabb761969526f445ed8d3173b
+    - path: output/qualimap/test/raw_data_qualimapReport/duplication_rate_histogram.txt
+      md5sum: 2299c92a3049270b9d7b1f521cafbb86
+    - path: output/qualimap/test/raw_data_qualimapReport/genome_fraction_coverage.txt
+      md5sum: 89c8c882c9423fafe03b4fc98e2bfb5a
+    - path: output/qualimap/test/raw_data_qualimapReport/insert_size_across_reference.txt
+      md5sum: 966230f2bbe796788beac30678597e79
+    - path: output/qualimap/test/raw_data_qualimapReport/insert_size_histogram.txt
+      md5sum: b3049afc6c3c57c3b278abd66a66a690
+    - path: output/qualimap/test/raw_data_qualimapReport/mapped_reads_gc-content_distribution.txt
+      md5sum: e9e0189463430b5565ea6da1d5976e40
+    - path: output/qualimap/test/raw_data_qualimapReport/mapped_reads_nucleotide_content.txt
+      md5sum: fa1566299e1bcd2d65c95b32ae2ca3d1
+    - path: output/qualimap/test/raw_data_qualimapReport/mapping_quality_across_reference.txt
+      md5sum: 40b6951739a602da545e38a00b698fde
+    - path: output/qualimap/test/raw_data_qualimapReport/mapping_quality_histogram.txt
+      md5sum: c66579f0114a9aedc70820fa609fb1ec