Merge pull request #223 from phue/bismark_genome_preparation

add bismark/genome_preparation + test

.github/filters.yml

@@ -54,6 +54,10 @@ bedtools_sort:
   - software/bedtools/sort/**
   - tests/software/bedtools/sort/**
 
+bismark_genome_preparation:
+  - software/bismark/genome_preparation/**
+  - tests/software/bismark/genome_preparation/**
+
 blast_makeblastdb:
   - software/blast/makeblastdb/**
   - tests/software/blast/makeblastdb/**

software/bismark/genome_preparation/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/genome_preparation/main.nf

@@ -0,0 +1,37 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+def options    = initOptions(params.options)
+
+process BISMARK_GENOME_PREPARATION {
+    tag "$fasta"
+    label 'process_high'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), publish_id:'') }
+
+    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:
+    path fasta, stageAs: "BismarkIndex/*"
+
+    output:
+    path "BismarkIndex" , emit: index
+    path "*.version.txt", emit: version
+
+    script:
+    def software   = getSoftwareName(task.process)
+    """
+    bismark_genome_preparation \\
+        $options.args \\
+        BismarkIndex
+
+    echo \$(bismark -v 2>&1) | sed 's/^.*Bismark Version: v//; s/Copyright.*\$//' > ${software}.version.txt
+    """
+}

software/bismark/genome_preparation/meta.yml

@@ -0,0 +1,57 @@
+name: bismark_genome_preparation
+description: |
+    Converts a specified reference genome into two different bisulfite
+    converted versions and indexes them for alignments.
+keywords:
+    - bismark
+    - 3-letter genome
+    - index
+    - methylation
+    - 5mC
+    - methylseq
+    - bisulphite
+    - fasta
+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:
+    - fasta:
+        type: file
+        description: Input genome fasta file
+output:
+    - index:
+        type: dir
+        description: Bismark genome index directory
+        pattern: "BismarkIndex"
+    - version:
+        type: file
+        description: File containing software version
+        pattern: "*.{version.txt}"
+authors:
+    - "@phue"

tests/software/bismark/genome_preparation/main.nf

@@ -0,0 +1,10 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { BISMARK_GENOME_PREPARATION } from '../../../../software/bismark/genome_preparation/main.nf' addParams( options: [:] )
+
+workflow test_bismark_genome_preparation {
+
+    BISMARK_GENOME_PREPARATION ( file("${launchDir}/tests/data/fasta/E_coli/NC_010473.fa", checkIfExists: true) )
+}

tests/software/bismark/genome_preparation/test.yml

@@ -0,0 +1,36 @@
+- name: Run bismark_genome_preparation test workflow
+  command: nextflow run ./tests/software/bismark/genome_preparation -entry test_bismark_genome_preparation -c tests/config/nextflow.config
+  tags:
+    - bismark
+    - bismark_genome_preparation
+  files:
+    - path: output/bismark/BismarkIndex/NC_010473.fa
+      md5sum: ad277e6b3da5dcf9fccfcd3a55fcec8e
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/CT_conversion/genome_mfa.CT_conversion.fa
+      md5sum: f878bc2c1d6d1404157ebf931e971f02
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/CT_conversion/BS_CT.1.bt2
+      md5sum: d865be3bfe02a3d5e277128fe7902b47
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/CT_conversion/BS_CT.2.bt2
+      md5sum: 486dfea5dde17f126cf498ead0e91421
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/CT_conversion/BS_CT.3.bt2
+      md5sum: cd201e81724f3099131aec16ef2cc53b
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/CT_conversion/BS_CT.4.bt2
+      md5sum: 4c81a7364a80bf3151a5464108593497
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/GA_conversion/genome_mfa.GA_conversion.fa
+      md5sum: 8cb16ceaf0543061151d8c1b7d1f85b4
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/GA_conversion/BS_GA.1.bt2
+      md5sum: fe77e5d17a7aa30c3a788bb90b396842
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/GA_conversion/BS_GA.2.bt2
+      md5sum: 3c6e4a48ad05f8af5f4015db7571d253
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/GA_conversion/BS_GA.3.bt2
+      md5sum: cd201e81724f3099131aec16ef2cc53b
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/GA_conversion/BS_GA.4.bt2
+      md5sum: f651bcfb9713f6ee1ce648b7c244b50e
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/CT_conversion/BS_CT.rev.1.bt2
+      md5sum: 2261afbed214da6f59f1bf648c20fb6b
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/CT_conversion/BS_CT.rev.2.bt2
+      md5sum: 622991847ea56f3c4f8583bb23ceee7a
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/GA_conversion/BS_GA.rev.1.bt2
+      md5sum: 32220350073a71a9cd69677f3aa4e577
+    - path: output/bismark/BismarkIndex/Bisulfite_Genome/GA_conversion/BS_GA.rev.2.bt2
+      md5sum: f7467830c4f33497c54c9e7e502a5def