Merge pull request #166 from MaxUlysse/master_bwa-mem2

chores: add bwa-mem2 index + mem

.github/filters.yml

@@ -96,6 +96,14 @@ bwa_mem:
   - software/bwa/mem/**
   - tests/software/bwa/mem/**
 
+bwamem2_index:
+  - software/bwamem2/index/**
+  - tests/software/bwamem2/index/**
+
+bwamem2_mem:
+  - software/bwamem2/mem/**
+  - tests/software/bwamem2/mem/**
+
 cat_fastq:
   - software/cat/fastq/**
   - tests/software/cat/fastq/**

software/bwamem2/index/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/bwamem2/index/main.nf

@@ -0,0 +1,35 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+def options    = initOptions(params.options)
+
+process BWAMEM2_INDEX {
+    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::bwa-mem2=2.1=he513fc3_0" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/bwa-mem2:2.1--he513fc3_0"
+    } else {
+        container "quay.io/biocontainers/bwa-mem2:2.1--he513fc3_0"
+    }
+
+    input:
+    path fasta
+
+    output:
+    path "bwamem2"      , emit: index
+    path "*.version.txt", emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    """
+    mkdir bwamem2
+    bwa-mem2 index $options.args $fasta -p bwamem2/${fasta}
+    echo \$(bwa-mem2 version 2>&1) > ${software}.version.txt
+    """
+}

software/bwamem2/index/meta.yml

@@ -0,0 +1,50 @@
+name: bwamem2_index
+description: Create BWA-mem2 index for reference genome
+keywords:
+    - index
+    - fasta
+    - genome
+    - reference
+tools:
+    - bwa:
+        description: |
+            BWA-mem2 is a software package for mapping DNA sequences against
+            a large reference genome, such as the human genome.
+        homepage: https://github.com/bwa-mem2/bwa-mem2
+        documentation: https://github.com/bwa-mem2/bwa-mem2#usage
+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: file
+        description: BWA genome index files
+        pattern: "*.{0132,amb,ann,bwt.2bit.64,pac}"
+    - version:
+        type: file
+        description: File containing software version
+        pattern: "*.{version.txt}"
+authors:
+    - "@maxulysse"

software/bwamem2/mem/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/bwamem2/mem/main.nf

@@ -0,0 +1,46 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+def options    = initOptions(params.options)
+
+process BWAMEM2_MEM {
+    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::bwa-mem2=2.1=he513fc3_0 bioconda::samtools=1.11=h6270b1f_0" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/mulled-v2-e5d375990341c5aef3c9aff74f96f66f65375ef6:e6f0d20c9d78572ddbbf00d8767ee6ff865edd4e-0"
+    } else {
+        container "quay.io/biocontainers/mulled-v2-e5d375990341c5aef3c9aff74f96f66f65375ef6:e6f0d20c9d78572ddbbf00d8767ee6ff865edd4e-0"
+    }
+
+    input:
+    tuple val(meta), path(reads)
+    path  index
+
+    output:
+    tuple val(meta), path("*.bam"), emit: bam
+    path  "*.version.txt"         , emit: version
+
+    script:
+    def software   = getSoftwareName(task.process)
+    def prefix     = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    def read_group = meta.read_group ? "-R ${meta.read_group}" : ""
+    """
+    INDEX=`find -L ./ -name "*.amb" | sed 's/.amb//'`
+
+    bwa-mem2 mem \\
+        $options.args \\
+        $read_group \\
+        -t $task.cpus \\
+        \$INDEX \\
+        $reads \\
+        | samtools view $options.args2 -@ $task.cpus -bhS -o ${prefix}.bam -
+
+    echo \$(bwa-mem2 version 2>&1) > ${software}.version.txt
+    """
+}

software/bwamem2/mem/meta.yml

@@ -0,0 +1,65 @@
+name: bwamem2_mem
+description: Performs fastq alignment to a fasta reference using BWA
+keywords:
+    - mem
+    - bwa
+    - alignment
+    - map
+    - fastq
+    - bam
+    - sam
+tools:
+    - bwa:
+        description: |
+            BWA is a software package for mapping DNA sequences against
+            a large reference genome, such as the human genome.
+        homepage: http://bio-bwa.sourceforge.net/
+        documentation: http://www.htslib.org/doc/samtools.html
+        arxiv: arXiv:1303.3997
+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 ]
+    - reads:
+        type: file
+        description: |
+            List of input FastQ files of size 1 and 2 for single-end and paired-end data,
+            respectively.
+    - index:
+        type: file
+        description: BWA genome index files
+        pattern: "Directory containing BWA index *.{amb,ann,bwt,pac,sa}"
+output:
+    - bam:
+        type: file
+        description: Output BAM file containing read alignments
+        pattern: "*.{bam}"
+    - version:
+        type: file
+        description: File containing software version
+        pattern: "*.{version.txt}"
+authors:
+    - "@maxulysse"

tests/data/index/E_coli/bwamem2/NC_010473.fa.amb

@@ -0,0 +1,3 @@
+4686137 1 2
+20895 1 Y
+142347 1 R

tests/data/index/E_coli/bwamem2/NC_010473.fa.ann

@@ -0,0 +1,3 @@
+4686137 1 11
+0 gi|170079663|ref|NC_010473.1| Escherichia coli str. K-12 substr. DH10B, complete genome
+0 4686137 2

tests/software/bwamem2/index/main.nf

@@ -0,0 +1,9 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { BWAMEM2_INDEX } from '../../../../software/bwamem2/index/main.nf' addParams( options: [:] )
+
+workflow test_bwamem2_index {
+    BWAMEM2_INDEX ( file("${launchDir}/tests/data/fasta/E_coli/NC_010473.fa", checkIfExists: true) )
+}

tests/software/bwamem2/index/test.yml

@@ -0,0 +1,16 @@
+- name: bwamem2 index
+  command: nextflow run ./tests/software/bwamem2/index -entry test_bwamem2_index -c tests/config/nextflow.config
+  tags:
+    - bwamem2
+    - bwamem2_index
+  files:
+    - path: output/bwamem2/bwamem2/NC_010473.fa.0123
+      md5sum: e112e4d2fd893fb939d306b27ece9de5
+    - path: output/bwamem2/bwamem2/NC_010473.fa.amb
+      md5sum: 942a990ae872f1c0b8d72dda2db405d5
+    - path: output/bwamem2/bwamem2/NC_010473.fa.ann
+      md5sum: ebf1a0279cf5b8d7f1a8cb855a3a3705
+    - path: output/bwamem2/bwamem2/NC_010473.fa.bwt.2bit.64
+      md5sum: f0154573be12440aee294b726cd88c86
+    - path: output/bwamem2/bwamem2/NC_010473.fa.pac
+      md5sum: 4d5e6fc45bbc968f7f859e9ca2cc89ad

tests/software/bwamem2/mem/main.nf

@@ -0,0 +1,36 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { BWAMEM2_MEM } from '../../../../software/bwamem2/mem/main.nf' addParams( options: [:] )
+
+/*
+ * Test with single-end data
+ */
+workflow test_bwamem2_mem_single_end {
+
+    def input = []
+    input = [ [ id:'test', single_end:true ], // meta map
+              [ file("${launchDir}/tests/data/fastq/dna/Ecoli_DNA_R1.fastq.gz", checkIfExists: true) ] ]
+
+    BWAMEM2_MEM (
+        input,
+        file("${launchDir}/tests/data/index/E_coli/bwamem2/", checkIfExists: true)
+    )
+}
+
+/*
+ * Test with paired-end data
+ */
+workflow test_bwamem2_mem_paired_end {
+
+    def input = []
+    input = [ [ id:'test', single_end:false ], // meta map
+              [ file("${launchDir}/tests/data/fastq/dna/Ecoli_DNA_R1.fastq.gz", checkIfExists: true),
+                file("${launchDir}/tests/data/fastq/dna/Ecoli_DNA_R2.fastq.gz", checkIfExists: true) ] ]
+
+    BWAMEM2_MEM (
+        input,
+        file("${launchDir}/tests/data/index/E_coli/bwamem2/", checkIfExists: true)
+    )
+}

tests/software/bwamem2/mem/test.yml

@@ -0,0 +1,19 @@
+- name: bwamem2 mem single-end
+  command: nextflow run ./tests/software/bwamem2/mem -entry test_bwamem2_mem_single_end -c tests/config/nextflow.config
+  tags:
+    - bwamem2
+    - bwamem2_mem
+    - bwamem2_mem_single_end
+  files:
+    - path: output/bwamem2/test.bam
+      md5sum: 354acd3b7033a2d4ee69452df18c0a4d
+
+- name: bwamem2 mem paired-end
+  command: nextflow run ./tests/software/bwamem2/mem -entry test_bwamem2_mem_paired_end -c tests/config/nextflow.config
+  tags:
+    - bwamem2
+    - bwamem2_mem
+    - bwamem2_mem_paired_end
+  files:
+    - path: output/bwamem2/test.bam
+      md5sum: 26187528a7bde13a2a9e9dd549b9bcd0