add mergebamalignment (#259)

* add mergebamalignment

* fix test.yml

* update to latest gatk4 version

* Update software/gatk4/mergebamalignment/main.nf

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Update tests/data/README.md

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Update software/gatk4/mergebamalignment/main.nf

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Update software/gatk4/mergebamalignment/meta.yml

* fixed unmapped/unaligned

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

software/gatk4/mergebamalignment/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/gatk4/mergebamalignment/main.nf

@@ -0,0 +1,44 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+def options    = initOptions(params.options)
+
+process GATK4_MERGEBAMALIGNMENT {
+    tag "$meta.id"
+    label 'process_low'
+    publishDir "${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::gatk4=4.2.0.0" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/gatk4:4.2.0.0--0"
+    } else {
+        container "quay.io/biocontainers/gatk4:4.2.0.0--0"
+    }
+
+    input:
+    tuple val(meta), path(aligned)
+    path unmapped
+    path fasta
+    path dict
+
+    output:
+    tuple val(meta), path('*.merged.bam'), emit: bam
+    path  '*.version.txt'                , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix   = options.suffix ? "${meta.id}.${options.suffix}" : "${meta.id}"
+    """
+    gatk MergeBamAlignment \\
+        ALIGNED=$aligned \\
+        UNMAPPED=$unmapped \\
+        R=$fasta \\
+        O=${prefix}.merged.bam \\
+        $options.args
+
+    gatk --version | grep Picard | sed "s/Picard Version: //g" > ${software}.version.txt
+    """
+}

software/gatk4/mergebamalignment/meta.yml

@@ -0,0 +1,66 @@
+name: gatk4_mergebamalignment
+description: Merge unmapped with mapped BAM files
+keywords:
+  - bam
+  - merge
+tools:
+  - gatk4:
+      description: |
+        Developed in the Data Sciences Platform at the Broad Institute, the toolkit offers a wide variety of tools
+        with a primary focus on variant discovery and genotyping. Its powerful processing engine
+        and high-performance computing features make it capable of taking on projects of any size.
+      homepage: https://gatk.broadinstitute.org/hc/en-us
+      documentation: https://gatk.broadinstitute.org/hc/en-us/categories/360002369672s
+      doi: 10.1158/1538-7445.AM2017-3590
+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']
+  - aligned:
+      type: file
+      description: The aligned bam file
+      pattern: "*.{bam}"
+  - unaligned:
+      type: file
+      description: The unmaped bam file
+      pattern: "*.{bam}"
+  - fasta:
+      type: file
+      description: The reference fasta file
+  - dict:
+      type: file
+      description: GATK sequence dictionary
+output:
+  - bam:
+      type: file
+      description: The merged bam file
+      pattern: "*.bam"
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.version.txt"
+authors:
+  - "@kevinmenden"

tests/config/pytest_software.yml

@@ -321,6 +321,10 @@ untar:
   - software/untar/**
   - tests/software/untar/**
 
+gatk4_mergebamalignment:
+  - software/gatk4/mergebamalignment/**
+  - tests/software/gatk4/mergebamalignment/**
+
 gatk4_variantfiltration:
   - software/gatk4/variantfiltration/**
   - tests/software/gatk4/variantfiltration/**

tests/data/README.md

@@ -14,6 +14,7 @@ When adding a new module, please check carefully whether the data necessary for
         * 'test_{,methylated}_paired_end.sorted.bam': sorted version of the above bam file
         * 'test_{,methylated}_paired_end.bam.sorted.bam.bai': bam index for the sorted bam file
         * 'test_single_end.bam': alignment (unsorted) of the 'test_1.fastq.gz' reads against test_genomic.fasta using minimap2
+        * 'test_unaligned.bam': unmapped BAM file created from 'test_1.fastq.gz' using GATK4 SamToFastq
     * bed
         * 'test.bed': exemplary bed file for the MT192765.1 genome (fasta/test_genomic.fasta)
         * 'test.2.bed': slightly modified copy of the above file
@@ -51,4 +52,3 @@ When adding a new module, please check carefully whether the data necessary for
 * 'test.txt.gar.gz' exemplary tar file for the untar module
 
 
-

tests/software/gatk4/mergebamalignment/main.nf

@@ -0,0 +1,16 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { GATK4_MERGEBAMALIGNMENT } from '../../../../software/gatk4/mergebamalignment/main.nf' addParams( options: [:] )
+
+workflow test_gatk4_mergebamalignment {
+    def input = []
+    input = [ [ id:'test' ], // meta map
+              file("${launchDir}/tests/data/genomics/sarscov2/bam/test_single_end.bam", checkIfExists: true) ]
+    unmapped = file("${launchDir}/tests/data/genomics/sarscov2/bam/test_unaligned.bam", checkIfExists: true)
+    fasta = file("${launchDir}/tests/data/genomics/sarscov2/fasta/test_genome.fasta", checkIfExists: true)
+    dict = file("${launchDir}/tests/data/genomics/sarscov2/fasta/test_genome.dict", checkIfExists: true)
+
+    GATK4_MERGEBAMALIGNMENT ( input, unmapped, fasta, dict )
+}

tests/software/gatk4/mergebamalignment/test.yml

@@ -0,0 +1,8 @@
+- name: gatk4 mergebamalignment
+  command: nextflow run ./tests/software/gatk4/mergebamalignment -entry test_gatk4_mergebamalignment -c tests/config/nextflow.config
+  tags:
+    - gatk4
+    - gatk4_mergebamalignment
+  files:
+    - path: output/gatk4/test.merged.bam
+      md5sum: 825816a6b288d644318c46cc9968e129