Add Seqfu/derep (#5958)

* add-usearch-unoise3 * add seqfu/derep * Delete modules/nf-core/usearch/unoise3/environment.yml * Delete modules/nf-core/usearch/unoise3/main.nf * remove files * update metadata * Update modules/nf-core/seqfu/derep/main.nf Co-authored-by: Simon Pearce <[email protected]> * Update modules/nf-core/seqfu/derep/main.nf Co-authored-by: Simon Pearce <[email protected]> * Update modules/nf-core/seqfu/derep/main.nf Co-authored-by: Simon Pearce <[email protected]> * Update modules/nf-core/seqfu/derep/main.nf Co-authored-by: Simon Pearce <[email protected]> * Update modules/nf-core/seqfu/derep/main.nf Co-authored-by: Simon Pearce <[email protected]> * Update modules/nf-core/seqfu/derep/tests/main.nf.test Co-authored-by: Simon Pearce <[email protected]> * Fix dataset in -stub * Clarify description * Update modules/nf-core/seqfu/derep/meta.yml Co-authored-by: Simon Pearce <[email protected]> * check for pre-existing output file && update test * Update main.nf --------- Co-authored-by: Simon Pearce <[email protected]>
nf-core · Jul 18, 2024 · e2e6613 · e2e6613
1 parent af50683
commit e2e6613
Show file tree

Hide file tree

Showing 6 changed files with 285 additions and 0 deletions.
diff --git a/modules/nf-core/seqfu/derep/environment.yml b/modules/nf-core/seqfu/derep/environment.yml
@@ -0,0 +1,9 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/environment-schema.json
+name: "seqfu_derep"
+channels:
+  - conda-forge
+  - bioconda
+  - defaults
+dependencies:
+  - "bioconda::seqfu=1.20.3"
diff --git a/modules/nf-core/seqfu/derep/main.nf b/modules/nf-core/seqfu/derep/main.nf
@@ -0,0 +1,50 @@
+process SEQFU_DEREP {
+    tag "$meta.id"
+    label 'process_low'
+    conda "${moduleDir}/environment.yml"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/seqfu:1.20.3--h1eb128b_2':
+        'biocontainers/seqfu:1.20.3--h1eb128b_2' }"
+
+    input:
+    tuple val(meta), path(fastas)
+
+    output:
+    tuple val(meta), path("*_derep.fasta.gz"), emit: fasta
+    path "versions.yml"                 , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}_derep"
+    def fasta_files = fastas.collect { it.getName() }
+    if (fasta_files.any { it == "${prefix}.fasta.gz" }) {
+        error "Input file name coincides with the output file name: ${prefix}.fasta.gz. Please set a unique prefix."
+    }
+
+    """
+    seqfu \\
+        derep \\
+        $args \\
+        $fastas | gzip -c > "${prefix}.fasta.gz"
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        seqfu: \$(seqfu version)
+    END_VERSIONS
+    """
+
+    stub:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}_derep"
+    """
+    echo "" | gzip -c > "${prefix}.fasta.gz"
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        seqfu: \$(seqfu version)
+    END_VERSIONS
+    """
+}
diff --git a/modules/nf-core/seqfu/derep/meta.yml b/modules/nf-core/seqfu/derep/meta.yml
@@ -0,0 +1,44 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/meta-schema.json
+name: "seqfu_derep"
+description: Dereplicate FASTX sequences, removing duplicate sequences and printing the number of identical sequences in the sequence header. Can dereplicate already dereplicated FASTA files, summing the numbers found in the headers.
+keywords:
+  - dereplicate
+  - fasta
+  - uniques
+tools:
+  - "seqfu":
+      description: "DNA sequence utilities for FASTX files"
+      homepage: "https://telatin.github.io/seqfu2/"
+      documentation: "https://telatin.github.io/seqfu2/"
+      tool_dev_url: "https://telatin.github.io/seqfu2/tools/derep.html"
+      doi: "10.3390/bioengineering8050059"
+      licence: ["GPL v3"]
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'sample1', single_end:false ]`
+  - fastas:
+      type: file
+      description: Input files (mainly FASTA, FASTQ supported)
+      pattern: "*.{fa,fna,faa,fasta,fq,fastq}[.gz]"
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'sample1', single_end:false ]`
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  - fasta:
+      type: file
+      description: dereplicated file (FASTA format)
+      pattern: "*.{fasta.gz}"
+authors:
+  - "@telatin"
+maintainers:
+  - "@telatin"
diff --git a/modules/nf-core/seqfu/derep/tests/main.nf.test b/modules/nf-core/seqfu/derep/tests/main.nf.test
@@ -0,0 +1,87 @@
+nextflow_process {
+
+    name "Test Process SEQFU_DEREP"
+    script "../main.nf"
+    process "SEQFU_DEREP"
+
+    tag "modules"
+    tag "modules_nfcore"
+    tag "seqfu"
+    tag "seqfu/derep"
+
+    test("derep - 3 fastas") {
+
+        when {
+            process {
+                """
+                input[0] = [
+                            [ id:'test' ],
+                            [
+                                file(params.test_data['sarscov2']['genome']['proteome_fasta'], checkIfExists: true),
+                                file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true),
+                                file(params.test_data['sarscov2']['genome']['genome_fasta_gz'], checkIfExists: true)
+                            ]
+                        ]
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(process.out).match() },
+                { assert process.out.fasta.size() == 1 }
+            )
+        }
+
+    }
+    test("derep - conflict") {
+
+        when {
+            process {
+                """
+                input[0] = Channel.of(">T1;size=300",
+                    "TTGATCACATA",
+                    ">T2;size=10",
+                    "TTGATCTCATA",
+                    "T3;size=4",
+                    "TTGATGACATA")
+                .collectFile(name: "test_derep.fasta.gz", newLine:true, sort:false)
+                .map { file -> [ [ id:'test', single_end:true ], file ] }
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert snapshot(process.out).match() }
+            )
+        }
+
+    }
+
+    test("derep - stub") {
+
+        options "-stub"
+
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id:'test', single_end:false ], // meta map
+                    file(params.test_data['sarscov2']['genome']['proteome_fasta'], checkIfExists: true)
+                    ]
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(process.out).match() }
+            )
+        }
+
+    }
+
+}
diff --git a/modules/nf-core/seqfu/derep/tests/main.nf.test.snap b/modules/nf-core/seqfu/derep/tests/main.nf.test.snap
@@ -0,0 +1,93 @@
+{
+    "derep - stub": {
+        "content": [
+            {
+                "0": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": false
+                        },
+                        "test_derep.fasta.gz:md5,68b329da9893e34099c7d8ad5cb9c940"
+                    ]
+                ],
+                "1": [
+                    "versions.yml:md5,f71fc1ed4ec36bf3a389c4aa28e1f08d"
+                ],
+                "fasta": [
+                    [
+                        {
+                            "id": "test",
+                            "single_end": false
+                        },
+                        "test_derep.fasta.gz:md5,68b329da9893e34099c7d8ad5cb9c940"
+                    ]
+                ],
+                "versions": [
+                    "versions.yml:md5,f71fc1ed4ec36bf3a389c4aa28e1f08d"
+                ]
+            }
+        ],
+        "meta": {
+            "nf-test": "0.8.4",
+            "nextflow": "23.04.0"
+        },
+        "timestamp": "2024-07-15T14:53:08.527366698"
+    },
+    "derep - 3 fastas": {
+        "content": [
+            {
+                "0": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test_derep.fasta.gz:md5,db73b7edf590972f275915ffb7810933"
+                    ]
+                ],
+                "1": [
+                    "versions.yml:md5,f71fc1ed4ec36bf3a389c4aa28e1f08d"
+                ],
+                "fasta": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test_derep.fasta.gz:md5,db73b7edf590972f275915ffb7810933"
+                    ]
+                ],
+                "versions": [
+                    "versions.yml:md5,f71fc1ed4ec36bf3a389c4aa28e1f08d"
+                ]
+            }
+        ],
+        "meta": {
+            "nf-test": "0.8.4",
+            "nextflow": "23.04.0"
+        },
+        "timestamp": "2024-07-15T14:53:02.934761389"
+    },
+    "derep - conflict": {
+        "content": [
+            {
+                "0": [
+
+                ],
+                "1": [
+
+                ],
+                "fasta": [
+
+                ],
+                "versions": [
+
+                ]
+            }
+        ],
+        "meta": {
+            "nf-test": "0.8.4",
+            "nextflow": "23.04.0"
+        },
+        "timestamp": "2024-07-15T15:17:01.598941968"
+    }
+}
diff --git a/modules/nf-core/seqfu/derep/tests/tags.yml b/modules/nf-core/seqfu/derep/tests/tags.yml
@@ -0,0 +1,2 @@
+seqfu/derep:
+  - "modules/nf-core/seqfu/derep/**"