Add new module: Das Tool (#1004)

* add pydamage module

* remove TODOs

* split module by subcommands

* update version parsing

* remove forgotten TODOs

* update module names

* remove old holistic module

* Update modules/pydamage/analyze/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* add keywords

* update resource requirement

* Update modules/pydamage/filter/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/pydamage/filter/meta.yml

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* merge from upstream

* update pydamage from upstream

* add freebayes

* update pydamage test from upstream

* fix meta.yml

* update functions.nf

* update test.yml

* update version parsing

* update version parsing

* fix indentation

* Update modules/freebayes/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/freebayes/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/freebayes/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* add optional inputs

* Update modules/freebayes/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* add bed test

* add metabat2 module

* only freebayes

* remove metabat2

* update md5sum because of vcf including date of the day

* add keyword

* rescue conflicted files

* attempt to fix ECLint

* add pytest workflow for metabat

* remove -

* Update modules/metabat2/jgisummarizebamcontigdepths/meta.yml

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/metabat2/metabat2/meta.yml

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/metabat2/metabat2/meta.yml

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/metabat2/jgisummarizebamcontigdepths/meta.yml

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* add optional inputs/outpus

* remove trailing whitespace

* compressing and removing not reproducible md5sums

* follow symlinks while decompressing

* add dastool/scaffolds2bin

* add dastool

* remove non reproducible md5sum check for compressed files

* Update tests/modules/metabat2/metabat2/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update tests/modules/metabat2/metabat2/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* split tests

* export env variable

* remove metabat2 from PR

* fix linting errors

* remove traling whitespace

* Update modules/metabat2/jgisummarizebamcontigdepths/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/metabat2/jgisummarizebamcontigdepths/meta.yml

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/metabat2/metabat2/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/metabat2/metabat2/meta.yml

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* answer PR comments and switch to bgzip

* PR review updates

* update from PR review

* update test files

* add bacillus fragilis alignments

* switch tests to bacillus fragilis

* add string check

* update test string

* add pr comment answer

* last fixes for PR review

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>
Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>
Co-authored-by: Gregor Sturm <mail@gregor-sturm.de>

modules/dastool/dastool/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/dastool/dastool/main.nf

@@ -0,0 +1,73 @@
+include { initOptions; saveFiles; getSoftwareName; getProcessName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process DASTOOL_DASTOOL {
+    tag "$meta.id"
+    label 'process_medium'
+    publishDir "${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::das_tool=1.1.3" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/das_tool:1.1.3--r41hdfd78af_0"
+    } else {
+        container "quay.io/biocontainers/das_tool:1.1.3--r41hdfd78af_0"
+    }
+
+    input:
+    tuple val(meta), path(contigs), path(bins)
+    path(proteins)
+    path(db_directory)
+    val(search_engine)
+
+    output:
+    tuple val(meta), path("*.log")                                      , emit: log
+    tuple val(meta), path("*_summary.txt")                              , emit: summary
+    tuple val(meta), path("*_DASTool_scaffolds2bin.txt")                , emit: scaffolds2bin
+    tuple val(meta), path("*.eval")                     , optional: true, emit: eval
+    tuple val(meta), path("*_DASTool_bins/*.fa")        , optional: true, emit: bins
+    tuple val(meta), path("*.pdf")                      , optional: true, emit: pdfs
+    tuple val(meta), path("*.proteins.faa")             , optional: true, emit: fasta_proteins
+    tuple val(meta), path("*.archaea.scg")              , optional: true, emit: fasta_archaea_scg
+    tuple val(meta), path("*.bacteria.scg")             , optional: true, emit: fasta_bacteria_scg
+    tuple val(meta), path("*.seqlength")                , optional: true, emit: seqlength
+    path "versions.yml"                                 , emit: versions
+
+    script:
+    def prefix = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    def bin_list = bins instanceof List ? bins.join(",") : "$bins"
+    def engine = search_engine ? "--search_engine $search_engine" : "--search_engine diamond"
+    def db_dir = db_directory ? "--db_directory $db_directory" : ""
+    def clean_contigs = contigs.toString() - ".gz"
+    def decompress_contigs = contigs.toString() == clean_contigs ? "" : "gunzip -q -f $contigs"
+    def decompress_proteins = proteins ? "gunzip -f $proteins" : ""
+    def clean_proteins = proteins ? proteins.toString() - ".gz" : ""
+    def proteins_pred = proteins ? "--proteins $clean_proteins" : ""
+
+    if (! search_engine) {
+        log.info('[DAS_Tool] Default search engine (USEARCH) is proprietary software and not available in bioconda. Using DIAMOND as alternative.')
+    }
+
+    """
+    $decompress_proteins
+    $decompress_contigs
+
+    DAS_Tool \\
+        $options.args \\
+        $proteins_pred \\
+        $db_dir \\
+        $engine \\
+        -t $task.cpus \\
+        --bins $bin_list \\
+        -c $clean_contigs \\
+        -o $prefix
+
+    cat <<-END_VERSIONS > versions.yml
+    ${getProcessName(task.process)}:
+        ${getSoftwareName(task.process)}: \$( DAS_Tool --version 2>&1 | grep "DAS Tool" | sed 's/DAS Tool version //' )
+    END_VERSIONS
+    """
+}

modules/dastool/dastool/meta.yml

@@ -0,0 +1,100 @@
+name: dastool_dastool
+description: DAS Tool binning step.
+keywords:
+  - binning
+  - das tool
+  - table
+  - de novo
+  - bins
+  - contigs
+  - assembly
+  - das_tool
+tools:
+  - dastool:
+      description: |
+        DAS Tool is an automated method that integrates the results
+        of a flexible number of binning algorithms to calculate an optimized, non-redundant
+        set of bins from a single assembly.
+
+      homepage: https://github.com/cmks/DAS_Tool
+      documentation: https://github.com/cmks/DAS_Tool
+      tool_dev_url: https://github.com/cmks/DAS_Tool
+      doi: "10.1038/s41564-018-0171-1"
+      licence: ['BSD']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - contigs:
+      type: file
+      description: fasta file
+      pattern: "*.{fa.gz,fas.gz,fasta.gz}"
+  - bins:
+      type: file
+      description: "Scaffolds2bin tabular file generated with dastool/scaffolds2bin"
+      pattern: "*.scaffolds2bin.tsv"
+  - proteins:
+      type: file
+      description: Predicted proteins in prodigal fasta format (>scaffoldID_geneNo)
+      pattern: "*.{fa.gz,fas.gz,fasta.gz}"
+  - db_directory:
+      type: file
+      description: (optional) Directory of single copy gene database.
+  - search_engine:
+      type: val
+      description: Engine used for single copy gene identification. USEARCH is not supported due to it being proprietary [blast/diamond]
+
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "versions.yml"
+  - log:
+      type: file
+      description: Log file of the run
+      pattern: "*.log"
+  - summary:
+      type: file
+      description: Summary of output bins including quality and completeness estimates
+      pattern: "*summary.txt"
+  - scaffolds2bin:
+      type: file
+      description: Scaffolds to bin file of output bins
+      pattern: "*.scaffolds2bin.txt"
+  - eval:
+      type: file
+      description: Quality and completeness estimates of input bin sets
+      pattern: "*.eval"
+  - pdfs:
+      type: file
+      description: Plots showing the amount of high quality bins and score distribution of bins per method
+      pattern: "*.pdf"
+  - fasta_proteins:
+      type: file
+      description: Output from prodigal if not already supplied
+      pattern: "*.proteins.faa"
+  - fasta_archaea_scg:
+      type: file
+      description:  Results of archaeal single-copy-gene prediction
+      pattern: "*.archaea.scg"
+  - fasta_bacteria_scg:
+      type: file
+      description: Results of bacterial single-copy-gene prediction
+      pattern: "*.bacteria.scg"
+  - seqlength:
+      type: file
+      description: Summary of contig lengths
+      pattern: "*.seqlength"
+
+authors:
+  - "@maxibor"
+  - "@jfy133"

modules/dastool/scaffolds2bin/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/dastool/scaffolds2bin/main.nf

@@ -0,0 +1,46 @@
+include { initOptions; saveFiles; getSoftwareName; getProcessName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process DASTOOL_SCAFFOLDS2BIN {
+    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::das_tool=1.1.3" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/das_tool:1.1.3--r41hdfd78af_0"
+    } else {
+        container "quay.io/biocontainers/das_tool:1.1.3--r41hdfd78af_0"
+    }
+
+    input:
+    tuple val(meta), path(fasta)
+    val(extension)
+
+    output:
+    tuple val(meta), path("*.tsv"), emit: scaffolds2bin
+    path "versions.yml"                         , emit: versions
+
+    script:
+    def prefix = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    def file_extension = extension ? extension : "fasta"
+
+    """
+    gunzip -f *.${file_extension}.gz
+
+    Fasta_to_Scaffolds2Bin.sh \\
+        $options.args \\
+        -i . \\
+        -e $file_extension \\
+        > ${prefix}.tsv
+
+    cat <<-END_VERSIONS > versions.yml
+    ${getProcessName(task.process)}:
+        ${getSoftwareName(task.process)}: \$( DAS_Tool --version 2>&1 | grep "DAS Tool" | sed 's/DAS Tool version //' )
+    END_VERSIONS
+    """
+}

modules/dastool/scaffolds2bin/meta.yml

@@ -0,0 +1,58 @@
+name: dastool_scaffolds2bin
+description: Helper script to convert a set of bins in fasta format to tabular scaffolds2bin format
+keywords:
+  - binning
+  - das tool
+  - table
+  - de novo
+  - bins
+  - contigs
+  - assembly
+  - das_tool
+tools:
+  - dastool:
+      description: |
+        DAS Tool is an automated method that integrates the results
+        of a flexible number of binning algorithms to calculate an optimized, non-redundant
+        set of bins from a single assembly.
+
+      homepage: https://github.com/cmks/DAS_Tool
+      documentation: https://github.com/cmks/DAS_Tool
+      tool_dev_url: https://github.com/cmks/DAS_Tool
+      doi: "10.1038/s41564-018-0171-1"
+      licence: ['BSD']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - fasta:
+      type: file
+      description: Fasta of list of fasta files recommended to be gathered via with .collect() of bins
+      pattern: "*.{fa,fas,fasta}"
+  - binner:
+      type: val
+      description: Name of the binning software (optional)
+  - extension:
+      type: val
+      description: Fasta file extension (fa | fas | fasta | ...)
+
+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"
+  - scaffolds2bin:
+      type: file
+      description: tabular scaffolds2bin file for DAS tool input
+      pattern: "*.scaffolds2bin.tsv"
+
+authors:
+  - "@maxibor"

tests/config/pytest_modules.yml

@@ -338,6 +338,14 @@ damageprofiler:
   - modules/damageprofiler/**
   - tests/modules/damageprofiler/**
 
+dastool/dastool:
+  - modules/dastool/dastool/**
+  - tests/modules/dastool/dastool/**
+
+dastool/scaffolds2bin:
+  - modules/dastool/scaffolds2bin/**
+  - tests/modules/dastool/scaffolds2bin/**
+  
 dedup:
   - modules/dedup/**
   - tests/modules/dedup/**

tests/modules/dastool/dastool/main.nf

@@ -0,0 +1,33 @@
+#!/usr/bin/env nextflow
+nextflow.enable.dsl = 2
+
+include { METABAT2_METABAT2 } from '../../../../modules/metabat2/metabat2/main.nf' addParams( options: [args: '--minContig 1500 --minCV 0.1 --minCVSum 0.1 --minClsSize 10 --minS 2'] )
+include { METABAT2_JGISUMMARIZEBAMCONTIGDEPTHS } from '../../../../modules/metabat2/jgisummarizebamcontigdepths/main.nf' addParams( options: [:] )
+include { DASTOOL_SCAFFOLDS2BIN } from '../../../../modules/dastool/scaffolds2bin/main.nf' addParams( options: [:] )
+include { DASTOOL_DASTOOL } from '../../../../modules/dastool/dastool/main.nf' addParams( options: [args: '--score_threshold 0 --debug'] )
+
+workflow test_dastool_dastool {
+
+    input_depth = [ [ id:'test', single_end:false ], // meta map
+              file(params.test_data['bacteroides_fragilis']['illumina']['test1_paired_end_sorted_bam'], checkIfExists: true),
+              file(params.test_data['bacteroides_fragilis']['illumina']['test1_paired_end_sorted_bam_bai'], checkIfExists: true) ]
+
+    METABAT2_JGISUMMARIZEBAMCONTIGDEPTHS ( input_depth )
+
+    Channel.fromPath(params.test_data['bacteroides_fragilis']['genome']['genome_fna_gz'], checkIfExists: true)
+        .map { it -> [[ id:'test', single_end:false ], it] }
+        .join(METABAT2_JGISUMMARIZEBAMCONTIGDEPTHS.out.depth)
+        .set { input_metabat2 }
+
+    METABAT2_METABAT2 ( input_metabat2 )
+
+    DASTOOL_SCAFFOLDS2BIN ( METABAT2_METABAT2.out.fasta.collect(), "fa")
+
+    Channel.of([ [ id:'test', single_end:false ], // meta map
+                      file(params.test_data['bacteroides_fragilis']['genome']['genome_fna_gz'], checkIfExists: true)])
+        .join(DASTOOL_SCAFFOLDS2BIN.out.scaffolds2bin)
+        .set {input_dastool}
+    
+
+    DASTOOL_DASTOOL ( input_dastool, [], [], [] )
+}

tests/modules/dastool/dastool/test.yml

@@ -0,0 +1,29 @@
+- name: dastool dastool test_dastool_dastool
+  command: nextflow run tests/modules/dastool/dastool -entry test_dastool_dastool -c tests/config/nextflow.config
+  tags:
+    - dastool
+    - dastool/dastool
+  files:
+    - path: output/dastool/test.seqlength
+      md5sum: b815a5811008c36808a59b1d0dcfab24
+    - path: output/dastool/test.tsv
+      md5sum: 6e46c0be14dded7cb13af38f54feea47
+    - path: output/dastool/test_DASTool.log
+      contains: 
+        - 'DAS Tool run on'
+    - path: output/dastool/test_DASTool_scaffolds2bin.txt
+      md5sum: 6e46c0be14dded7cb13af38f54feea47
+    - path: output/dastool/test_DASTool_summary.txt
+      md5sum: a3efa8717b30dfada78dc5ae9a3dc396
+    - path: output/dastool/test_proteins.faa.archaea.scg
+      md5sum: e79d82eecee25821d1658ea4f082601d
+    - path: output/dastool/test_proteins.faa.bacteria.scg
+      md5sum: 8132cfb17cf398d41c036ead55c96ffe
+    - path: output/dastool/test_test.tsv.eval
+      md5sum: a3efa8717b30dfada78dc5ae9a3dc396
+    - path: output/metabat2/bins/test.1.fa.gz
+      md5sum: 2b297bf557cc3831b800348859331268
+    - path: output/metabat2/test.tsv.gz
+      md5sum: 619338fa5019e361d5545ce385a6961f
+    - path: output/metabat2/test.txt.gz
+      md5sum: 745a0446af6ef68b930975e9ce5a95d6

tests/modules/dastool/scaffolds2bin/main.nf

@@ -0,0 +1,25 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { METABAT2_METABAT2 } from '../../../../modules/metabat2/metabat2/main.nf' addParams( options: [args: '--minContig 1500 --minCV 0.1 --minCVSum 0.1 --minClsSize 10 --minS 2'] )
+include { METABAT2_JGISUMMARIZEBAMCONTIGDEPTHS } from '../../../../modules/metabat2/jgisummarizebamcontigdepths/main.nf' addParams( options: [:] )
+include { DASTOOL_SCAFFOLDS2BIN } from '../../../../modules/dastool/scaffolds2bin/main.nf' addParams( options: [:] )
+
+workflow test_dastool_scaffolds2bin {
+
+    input_depth = [ [ id:'test', single_end:false ], // meta map
+              file(params.test_data['bacteroides_fragilis']['illumina']['test1_paired_end_sorted_bam'], checkIfExists: true),
+              file(params.test_data['bacteroides_fragilis']['illumina']['test1_paired_end_sorted_bam_bai'], checkIfExists: true) ]
+
+    METABAT2_JGISUMMARIZEBAMCONTIGDEPTHS ( input_depth )
+
+    Channel.fromPath(params.test_data['bacteroides_fragilis']['genome']['genome_fna_gz'], checkIfExists: true)
+        .map { it -> [[ id:'test', single_end:false ], it] }
+        .join(METABAT2_JGISUMMARIZEBAMCONTIGDEPTHS.out.depth)
+        .set { input_metabat2 }
+
+    METABAT2_METABAT2 ( input_metabat2 )
+
+    DASTOOL_SCAFFOLDS2BIN ( METABAT2_METABAT2.out.fasta.collect(), "fa")
+}

tests/modules/dastool/scaffolds2bin/test.yml

@@ -0,0 +1,14 @@
+- name: dastool scaffolds2bin test_dastool_scaffolds2bin
+  command: nextflow run tests/modules/dastool/scaffolds2bin -entry test_dastool_scaffolds2bin -c tests/config/nextflow.config
+  tags:
+    - dastool
+    - dastool/scaffolds2bin
+  files:
+    - path: output/dastool/test.tsv
+      md5sum: 6e46c0be14dded7cb13af38f54feea47
+    - path: output/metabat2/bins/test.1.fa.gz
+      md5sum: 2b297bf557cc3831b800348859331268
+    - path: output/metabat2/test.tsv.gz
+      md5sum: 619338fa5019e361d5545ce385a6961f
+    - path: output/metabat2/test.txt.gz
+      md5sum: 745a0446af6ef68b930975e9ce5a95d6