Add metamaps/classify

modules/nf-core/metamaps/classify/environment.yml

@@ -0,0 +1,7 @@
+name: "metamaps_classify"
+channels:
+  - conda-forge
+  - bioconda
+  - defaults
+dependencies:
+  - "bioconda::metamaps=0.1.633d2e0"

modules/nf-core/metamaps/classify/main.nf

@@ -0,0 +1,68 @@
+
+process METAMAPS_CLASSIFY {
+    tag "$meta.id"
+    label 'process_single'
+
+    conda "${moduleDir}/environment.yml"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/metamaps:0.1.633d2e0--h21ec9f0_0':
+        'biocontainers/metamaps:0.1.633d2e0--h21ec9f0_0' }"
+
+    input:
+    tuple val(meta), path(classification_res)
+    tuple val(meta), path(meta_file)
+    tuple val(meta), path(meta_unmappedreadsLengths)
+    tuple val(meta), path(para_file)
+    path database_folder
+
+
+    output:
+    tuple val(meta), path("*classification_res.EM.WIMP")                              , emit: wimp
+    tuple val(meta), path("*classification_res.EM.evidenceUnknownSpecies")            , emit: evidence_unknown_species
+    tuple val(meta), path("*classification_res.EM.reads2Taxon")                       , emit: reads2taxon
+    tuple val(meta), path("*classification_res.EM")                                   , emit: em
+    tuple val(meta), path("*classification_res.EM.contigCoverage")                    , emit: contig_coverage
+    tuple val(meta), path("*classification_res.EM.lengthAndIdentitiesPerMappingUnit") , emit: length_and_id
+    tuple val(meta), path("*classification_res.EM.reads2Taxon.krona")                 , emit: krona
+    path "versions.yml"                                                               , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    metamaps \\
+        classify \\
+        $args \\
+        --mappings $classification_res \\
+        --threads $task.cpus \\
+        --DB $database_folder
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        metamaps: \$(metamaps | sed -n 2p | sed 's/^.*MetaMaps v //')
+    END_VERSIONS
+    """
+    stub:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    """
+    touch ${prefix}_classification_res.EM.WIMP
+    touch ${prefix}_classification_res.EM.evidenceUnknownSpecies
+    touch ${prefix}_classification_res.EM.reads2Taxon
+    touch ${prefix}_classification_res.EM
+    touch ${prefix}_classification_res.EM.contigCoverage
+    touch ${prefix}_classification_res.EM.lengthAndIdentitiesPerMappingUnit
+    touch ${prefix}_classification_res.EM.reads2Taxon.krona
+
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        metamaps: \$(metamaps | sed -n 2p | sed 's/^.*MetaMaps v //')
+    END_VERSIONS
+    """
+
+}

modules/nf-core/metamaps/classify/meta.yml

@@ -0,0 +1,73 @@
+name: metamaps_classify
+description: Strain-level metagenomic assignment
+keywords:
+  - metamaps
+  - long reads
+  - metagenomics
+  - taxonomy
+tools:
+  - metamaps:
+      description: MetaMaps is a tool for long-read metagenomic analysis
+      homepage: https://github.com/DiltheyLab/MetaMaps
+      documentation: https://github.com/DiltheyLab/MetaMaps/blob/master/README.md
+      tool_dev_url: https://github.com/DiltheyLab/MetaMaps
+      doi: "10.1038/s41467-019-10934-2"
+      licence: ["Public Domain"]
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - classification_res:
+      type: file
+      description: Coordinates where reads map
+      pattern: "*.{classification_res}"
+  - database_folder:
+      type: directory
+      description: Path to MetaMaps database
+      pattern: "*"
+
+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"
+  - wimp:
+      type: file
+      description: Sample composition at different taxonomic levels
+      pattern: "*.{classification_res.EM.WIMP}"
+  - evidence_unknown_species:
+      type: file
+      description: Statistics on read identities and zero-coverage regions
+      pattern: "*.{classification_res.EM.evidenceUnknownSpecies}"
+  - reads2taxon:
+      type: file
+      description: Taxon ID assignement of reads
+      pattern: "*.{classification_res.EM.reads2Taxon}"
+  - em:
+      type: file
+      description: The final and complete set of approximate read mappings
+      pattern: "*.{classification_res.EM}"
+  - contig_coverage:
+      type: file
+      description: Read coverage for contigs
+      pattern: "*.{classification_res.EM.contigCoverage}"
+  - length_and_id:
+      type: file
+      description: Read length and estimated identity for all reads
+      pattern: "*.{classification_res.EM.lengthAndIdentitiesPerMappingUnit}"
+  - krona:
+      type: file
+      description: Taxon ID assignement of reads in Krona format
+      pattern: "*.{classification_res.EM.reads2Taxon.krona}"
+
+authors:
+  - "@henningonsbring"
+  - "@sofstam"

modules/nf-core/metamaps/classify/tests/main.nf.test

@@ -0,0 +1,69 @@
+nextflow_process {
+
+    name "Test Process METAMAPS_CLASSIFY"
+    script "../main.nf"
+    process "METAMAPS_CLASSIFY"
+
+    tag "modules"
+    tag "modules_nfcore"
+    tag "metamaps"
+    tag "metamaps/mapdirectly"
+    tag "metamaps/classify"
+    tag "untar"
+
+
+    test("sarscov2_nanopore_classify") {
+        setup {
+            run("UNTAR") {
+                config "./nextflow.config"
+                script "../../../untar/main.nf"
+                process {
+                    """
+                    input[0] = [
+                        [],file(params.modules_testdata_base_path + 'genomics/sarscov2/genome/db/metamaps_db.tar.gz', checkIfExists: true)
+                    ]
+                    """
+                }
+            }
+            run("METAMAPS_MAPDIRECTLY") {
+                script "../../../metamaps/mapdirectly/main.nf"
+                process {
+                    """
+                    input[0] = [
+                    [ id:'test', single_end:true ], // meta map
+                    file(params.test_data['sarscov2']['nanopore']['test2_fastq_gz'], checkIfExists: true)
+                    ]
+                    input[1] = UNTAR.out.untar.map { id, it ->
+                    filename = it.toString() + "/DB.fa"
+                    return [ file(filename) ]
+                    }
+                    """
+                }
+            }
+        }
+
+        when {
+            process {
+                """
+                input[0] = METAMAPS_MAPDIRECTLY.out.classification_res
+                input[1] = METAMAPS_MAPDIRECTLY.out.meta_file
+                input[2] = METAMAPS_MAPDIRECTLY.out.meta_unmappedreadsLengths
+                input[3] = METAMAPS_MAPDIRECTLY.out.para_file
+                input[4] = UNTAR.out.untar.map{ it[1] }
+                """
+            }
+        }
+
+
+
+        then {
+        assertAll(
+                { assert process.success },
+                { assert snapshot ( process.out ).match() }
+            )
+        }
+
+    }
+}
+
+

modules/nf-core/metamaps/classify/tests/main.nf.test.snap

@@ -0,0 +1,145 @@
+{
+    "sarscov2_nanopore_classify": {
+        "content": [
+            {
+                "0": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM.WIMP:md5,b39a980519a382422ac6da49bea13de8"
+                    ]
+                ],
+                "1": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM.evidenceUnknownSpecies:md5,a8577e31b3d9b3a598b43fc531075e81"
+                    ]
+                ],
+                "2": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM.reads2Taxon:md5,f08d1d1865a3949a9c542b5b930fb37c"
+                    ]
+                ],
+                "3": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM:md5,c6b1f5475e3fb9c6633aa0623456eebc"
+                    ]
+                ],
+                "4": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM.contigCoverage:md5,131507f62806647f11b1dc10d546c73f"
+                    ]
+                ],
+                "5": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM.lengthAndIdentitiesPerMappingUnit:md5,56be33596aa37bc360bbec47cebdc5e2"
+                    ]
+                ],
+                "6": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM.reads2Taxon.krona:md5,9f644c743e2f4ed5617fe3d421758771"
+                    ]
+                ],
+                "7": [
+                    "versions.yml:md5,60993f8b87f7a2e25367c7dc96c9b859"
+                ],
+                "contig_coverage": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM.contigCoverage:md5,131507f62806647f11b1dc10d546c73f"
+                    ]
+                ],
+                "em": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM:md5,c6b1f5475e3fb9c6633aa0623456eebc"
+                    ]
+                ],
+                "evidence_unknown_species": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM.evidenceUnknownSpecies:md5,a8577e31b3d9b3a598b43fc531075e81"
+                    ]
+                ],
+                "krona": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM.reads2Taxon.krona:md5,9f644c743e2f4ed5617fe3d421758771"
+                    ]
+                ],
+                "length_and_id": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM.lengthAndIdentitiesPerMappingUnit:md5,56be33596aa37bc360bbec47cebdc5e2"
+                    ]
+                ],
+                "reads2taxon": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM.reads2Taxon:md5,f08d1d1865a3949a9c542b5b930fb37c"
+                    ]
+                ],
+                "versions": [
+                    "versions.yml:md5,60993f8b87f7a2e25367c7dc96c9b859"
+                ],
+                "wimp": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": true
+                        },
+                        "test.classification_res.EM.WIMP:md5,b39a980519a382422ac6da49bea13de8"
+                    ]
+                ]
+            }
+        ],
+        "meta": {
+            "nf-test": "0.8.4",
+            "nextflow": "23.04.1"
+        },
+        "timestamp": "2024-05-16T10:38:20.785516"
+    }
+}

modules/nf-core/metamaps/classify/tests/nextflow.config

@@ -0,0 +1,5 @@
+process {
+    withName: UNTAR {
+        ext.args2 = "--no-same-owner"
+    }
+}

modules/nf-core/metamaps/classify/tests/tags.yml

@@ -0,0 +1,2 @@
+metamaps/classify:
+  - "modules/nf-core/metamaps/classify/**"

modules/nf-core/metamaps/mapdirectly/main.nf

@@ -16,7 +16,7 @@ process METAMAPS_MAPDIRECTLY {
     tuple val(meta), path("*classification_res.meta")                     , emit: meta_file
     tuple val(meta), path("*classification_res.meta.unmappedReadsLengths"), emit: meta_unmappedreadsLengths
     tuple val(meta), path("*classification_res.parameters")               , emit: para_file
-    path "versions.yml"                                                   , emit: versions
+    path "versions.yml"                                                                                                                                                                                , emit: versions
 
     when:
     task.ext.when == null || task.ext.when