Add Miniasm module (#962)

* add racon

* add miniasm module

* edit miniasm module

* edit miniasm module

* Update tests/modules/racon/main.nf

Co-authored-by: Daniel Straub <42973691+d4straub@users.noreply.github.com>

* Update tests/modules/racon/test.yml

Co-authored-by: Daniel Straub <42973691+d4straub@users.noreply.github.com>

* Update modules/miniasm/meta.yml

Co-authored-by: Daniel Straub <42973691+d4straub@users.noreply.github.com>

* Update main.nf

Add some spaces.

* Update meta.yml

Correct DOI

* Update main.nf

* Apply suggestions from code review

* Update tests/modules/miniasm/test.yml

Co-authored-by: Daniel Straub <42973691+d4straub@users.noreply.github.com>
Co-authored-by: Sébastien Guizard <sguizard@ed.ac.uk>
Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

modules/miniasm/functions.nf

@@ -0,0 +1,78 @@
+//
+//  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()
+}
+
+//
+// Extract name of module from process name using $task.process
+//
+def getProcessName(task_process) {
+    return task_process.tokenize(':')[-1]
+}
+
+//
+// Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+//
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args            = args.args ?: ''
+    options.args2           = args.args2 ?: ''
+    options.args3           = args.args3 ?: ''
+    options.publish_by_meta = args.publish_by_meta ?: []
+    options.publish_dir     = args.publish_dir ?: ''
+    options.publish_files   = args.publish_files
+    options.suffix          = args.suffix ?: ''
+    return options
+}
+
+//
+// Tidy up and join elements of a list to return a path string
+//
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
+    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+//
+// Function to save/publish module results
+//
+def saveFiles(Map args) {
+    def ioptions  = initOptions(args.options)
+    def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+
+    // Do not publish versions.yml unless running from pytest workflow
+    if (args.filename.equals('versions.yml') && !System.getenv("NF_CORE_MODULES_TEST")) {
+        return null
+    }
+    if (ioptions.publish_by_meta) {
+        def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
+        for (key in key_list) {
+            if (args.meta && key instanceof String) {
+                def path = key
+                if (args.meta.containsKey(key)) {
+                    path = args.meta[key] instanceof Boolean ? "${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"
+    }
+}

modules/miniasm/main.nf

@@ -0,0 +1,48 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName; getProcessName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process MINIASM {
+    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), meta:meta, publish_by_meta:['id']) }
+
+    conda (params.enable_conda ? "bioconda::miniasm=0.3_r179" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/miniasm:0.3_r179--h5bf99c6_2"
+    } else {
+        container "quay.io/biocontainers/miniasm:0.3_r179--h5bf99c6_2"
+    }
+
+    input:
+    tuple val(meta), path(reads), path(paf)
+
+    output:
+    tuple val(meta), path("*.gfa.gz")  , emit: gfa
+    tuple val(meta), path("*.fasta.gz"), emit: assembly
+    path "versions.yml"                , emit: versions
+
+    script:
+    def prefix = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    """
+    miniasm \\
+        $options.args \\
+        -f $reads \\
+        $paf > \\
+        ${prefix}.gfa
+
+    awk '/^S/{print ">"\$2"\\n"\$3}' "${prefix}.gfa" | fold > ${prefix}.fasta
+
+    gzip -n ${prefix}.gfa
+    gzip -n ${prefix}.fasta
+
+    cat <<-END_VERSIONS > versions.yml
+    ${getProcessName(task.process)}:
+        ${getSoftwareName(task.process)}: \$( miniasm -V 2>&1 )
+    END_VERSIONS
+    """
+}

modules/miniasm/meta.yml

@@ -0,0 +1,51 @@
+name: miniasm
+description: A very fast OLC-based de novo assembler for noisy long reads
+keywords:
+  - assembly
+  - pacbio
+  - nanopore
+tools:
+  - miniasm:
+      description: Ultrafast de novo assembly for long noisy reads (though having no consensus step)
+      homepage: https://github.com/lh3/miniasm
+      documentation: https://github.com/lh3/miniasm
+      tool_dev_url: https://github.com/lh3/miniasm
+      doi: "10.1093/bioinformatics/btw152"
+      licence: ['MIT']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - reads:
+      type: file
+      description: List of input PacBio/ONT FastQ files.
+      pattern: "*.{fastq,fastq.gz,fq,fq.gz}"
+  - paf:
+      type: file
+      description: Alignment in PAF format
+      pattern: "*{.paf,.paf.gz}"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  - gfa:
+      type: file
+      description: Assembly graph
+      pattern: "*.gfa.gz"
+  - assembly:
+      type: file
+      description: Genome assembly
+      pattern: "*.fasta.gz"
+
+authors:
+  - "@avantonder"

tests/config/pytest_modules.yml

@@ -862,6 +862,10 @@ minia:
   - modules/minia/**
   - tests/modules/minia/**
 
+miniasm:
+  - modules/miniasm/**
+  - tests/modules/miniasm/**
+
 minimap2/align:
   - modules/minimap2/align/**
   - tests/modules/minimap2/align/**

tests/modules/miniasm/main.nf

@@ -0,0 +1,15 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { MINIASM } from '../../../modules/miniasm/main.nf' addParams( options: [suffix:'.assembly'] )
+
+workflow test_miniasm {
+    
+    input = [ [ id:'test', single_end:true ], // meta map
+             file(params.test_data['bacteroides_fragilis']['nanopore']['test_fastq_gz'], checkIfExists: true),
+             file(params.test_data['bacteroides_fragilis']['nanopore']['overlap_paf'], checkIfExists: true)
+            ]
+    
+    MINIASM ( input )
+}

tests/modules/miniasm/test.yml

@@ -0,0 +1,9 @@
+- name: miniasm test_miniasm
+  command: nextflow run tests/modules/miniasm -entry test_miniasm -c tests/config/nextflow.config
+  tags:
+    - miniasm
+  files:
+    - path: output/miniasm/test.assembly.gfa.gz
+      md5sum: c68e4c2b64338d1c0f5b79b32934da14
+    - path: output/miniasm/test.assembly.fasta.gz
+      md5sum: d2f78ae618c02744e7a57bf4706ab8b4