damageprofiler (#545)

* Add damageprofiler module

* Fix tests

* Bump container hopefully with font fix

* Following code review

software/damageprofiler/functions.nf

@@ -0,0 +1,68 @@
+//
+//  Utility functions used in nf-core DSL2 module files
+//
+
+//
+// Extract name of software tool from process name using $task.process
+//
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+//
+// Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+//
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args            = args.args ?: ''
+    options.args2           = args.args2 ?: ''
+    options.args3           = args.args3 ?: ''
+    options.publish_by_meta = args.publish_by_meta ?: []
+    options.publish_dir     = args.publish_dir ?: ''
+    options.publish_files   = args.publish_files
+    options.suffix          = args.suffix ?: ''
+    return options
+}
+
+//
+// Tidy up and join elements of a list to return a path string
+//
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
+    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+//
+// Function to save/publish module results
+//
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions  = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_meta) {
+            def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
+            for (key in key_list) {
+                if (args.meta && key instanceof String) {
+                    def path = key
+                    if (args.meta.containsKey(key)) {
+                        path = args.meta[key] instanceof Boolean ? "${key}_${args.meta[key]}".toString() : args.meta[key]
+                    }
+                    path = path instanceof String ? path : ''
+                    path_list.add(path)
+                }
+            }
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

software/damageprofiler/main.nf

@@ -0,0 +1,44 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process DAMAGEPROFILER {
+    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), meta:meta, publish_by_meta:['id']) }
+
+    conda (params.enable_conda ? "bioconda::damageprofiler=1.1" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/damageprofiler:1.1--hdfd78af_2"
+    } else {
+        container "quay.io/biocontainers/damageprofiler:1.1--hdfd78af_2"
+    }
+
+    input:
+    tuple val(meta), path(bam)
+    path fasta
+    path fai
+
+    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}"
+
+    """
+    damageprofiler \\
+        -i $bam \\
+        -r $fasta \\
+        -o $prefix/ \\
+        $options.args
+
+
+    echo \$(damageprofiler -v) | sed 's/^DamageProfiler v//' > ${software}.version.txt
+    """
+}

software/damageprofiler/meta.yml

@@ -0,0 +1,53 @@
+name: damageprofiler
+description: A Java based tool to determine damage patterns on ancient DNA as a replacement for mapDamage
+keywords:
+  - damage
+  - deamination
+  - miscoding lesions
+  - C to T
+  - ancient DNA
+  - aDNA
+  - palaeogenomics
+  - archaeogenomics
+  - palaeogenetics
+  - archaeogenetics
+tools:
+  - damageprofiler:
+      description: A Java based tool to determine damage patterns on ancient DNA as a replacement for mapDamage
+      homepage: https://github.com/Integrative-Transcriptomics/DamageProfiler
+      documentation: https://damageprofiler.readthedocs.io/
+      tool_dev_url: https://github.com/Integrative-Transcriptomics/DamageProfiler
+      doi: "10.1093/bioinformatics/btab190"
+      licence: ['GPL v3']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - bam:
+      type: file
+      description: BAM/CRAM/SAM file
+      pattern: "*.{bam,cram,sam}"
+  - fasta:
+      type: file
+      description: FASTA reference file
+      pattern: "*.{fasta,fna,fa}"
+  - fai:
+      type: file
+      description: FASTA index file from samtools faidx
+      pattern: "*.{fai}"
+
+output:
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+  - results:
+      type: dir
+      description: DamageProfiler results directory
+      pattern: "*/*"
+
+authors:
+  - "@jfy133"

tests/config/pytest_software.yml

@@ -190,6 +190,10 @@ cutadapt:
   - software/cutadapt/**
   - tests/software/cutadapt/**
 
+damageprofiler:
+  - software/damageprofiler/**
+  - tests/software/damageprofiler/**
+
 deeptools/computematrix:
   - software/deeptools/computematrix/**
   - tests/software/deeptools/computematrix/**

tests/software/damageprofiler/main.nf

@@ -0,0 +1,15 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { DAMAGEPROFILER } from '../../../software/damageprofiler/main.nf' addParams( options: [:] )
+
+workflow test_damageprofiler {
+
+    input = [ [ id:'test', single_end:false ], // meta map
+              [ file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_bam'], checkIfExists: true) ] ]
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+    fai = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+
+    DAMAGEPROFILER ( input, fasta, fai )
+}

tests/software/damageprofiler/test.yml

@@ -0,0 +1,36 @@
+- name: damageprofiler
+  command: nextflow run ./tests/software/damageprofiler -entry test_damageprofiler -c tests/config/nextflow.config -dump-channels
+  tags:
+    - damageprofiler
+  files:
+    - path: output/damageprofiler/test/3p_freq_misincorporations.txt
+      md5sum: da4cac90c78899a7cb6d72d415392b49
+    - path: output/damageprofiler/test/3pGtoA_freq.txt
+      md5sum: 8dab75d51a4b943b501d0995169c767f
+    - path: output/damageprofiler/test/5pCtoT_freq.txt
+      md5sum: fcc48ee5f72edff930d627c8bfdd8a5b
+    - path: output/damageprofiler/test/5p_freq_misincorporations.txt
+      md5sum: 54665474f5ef17dcc268567e5eaa7d86
+    - path: output/damageprofiler/test/DamagePlot_five_prime.svg
+    - path: output/damageprofiler/test/DamagePlot.pdf
+    - path: output/damageprofiler/test/DamagePlot_three_prime.svg
+    - path: output/damageprofiler/test/DamageProfiler.log
+      contains:
+        - "FINISHED SUCCESSFULLY"
+    - path: output/damageprofiler/test/dmgprof.json
+      md5sum: 98499024c7e937896e481f2d3cfbdd3e
+    - path: output/damageprofiler/test/DNA_comp_genome.txt
+      md5sum: f91e70760d91a1193a27e360aaddf2fd
+    - path: output/damageprofiler/test/DNA_composition_sample.txt
+      md5sum: 1257eb3eb42484647bfba2151f9ef04f
+    - path: output/damageprofiler/test/edit_distance.pdf
+    - path: output/damageprofiler/test/edit_distance.svg
+    - path: output/damageprofiler/test/editDistance.txt
+      md5sum: af2d2f4a99058ec56eae88ec27779e38
+    - path: output/damageprofiler/test/Length_plot_combined_data.svg
+    - path: output/damageprofiler/test/Length_plot_forward_reverse_separated.svg
+    - path: output/damageprofiler/test/Length_plot.pdf
+    - path: output/damageprofiler/test/lgdistribution.txt
+      md5sum: c5d029bf3a92b613310ee23f47d94981
+    - path: output/damageprofiler/test/misincorporation.txt
+      md5sum: 3aa6dd749010a492d92a815a83c196a8