Hifiadapterfilt

modules/nf-core/hifiadapterfilt/environment.yml

@@ -0,0 +1,5 @@
+channels:
+  - conda-forge
+  - bioconda
+dependencies:
+  - "bioconda::hifiadapterfilt=3.0.0"

modules/nf-core/hifiadapterfilt/main.nf

@@ -0,0 +1,57 @@
+process HIFIADAPTERFILT {
+    tag "$meta.id"
+    label 'process_medium'
+
+    conda "${moduleDir}/environment.yml"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/hifiadapterfilt:3.0.0--hdfd78af_0':
+        'biocontainers/hifiadapterfilt:3.0.0--hdfd78af_0' }"
+
+    input:
+    tuple val(meta), path(fastq)
+
+    output:
+    tuple val(meta), path("*.filt.fastq.gz")        , emit: filt
+    tuple val(meta), path("*.contaminant.blastout") , emit: blast_search
+    tuple val(meta), path("*.stats")                , emit: stats
+    tuple val(meta), path("*.blocklist")            , emit: headers
+    path "versions.yml"                             , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+script:
+    def args = task.ext.args ?: ''
+    // The tool AUTOMATICALLY detects fastq files from the input folder, hence an explicit call of "fastq" is not needed.
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    """
+    hifiadapterfilt.sh \\
+        ${args}
+
+    mv *.filt.fastq.gz ${prefix}.filt.fastq.gz
+    mv *.contaminant.blastout ${prefix}.contaminant.blastout
+    mv *.stats ${prefix}.stats
+    mv *.blocklist ${prefix}.blocklist
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        hifiadapterfilt: \$(hifiadapterfilt.sh -v)
+    END_VERSIONS
+    """
+
+    stub:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    echo "" | gzip > ${prefix}.filt.fastq.gz
+    touch ${prefix}.contaminant.blastout
+    touch ${prefix}.stats
+    touch ${prefix}.blocklist
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        hifiadapterfilt: \$(hifiadapterfilt.sh -v)
+    END_VERSIONS
+    """
+}

modules/nf-core/hifiadapterfilt/meta.yml

@@ -0,0 +1,86 @@
+name: "hifiadapterfilt"
+description: "Convert .bam to .fastq and remove reads with remnant PacBio adapter
+  sequences"
+keywords:
+  - pacbio
+  - hifi
+  - filter
+  - metagenomics
+  - adapters
+tools:
+  - hifiadapterfilt:
+      description: "Convert .bam to .fastq and remove CCS reads with remnant PacBio
+        adapter sequences"
+      homepage: "https://bio.tools/hifiadapterfilt"
+      documentation: "https://github.com/sheinasim/HiFiAdapterFilt"
+      tool_dev_url: "https://github.com/sheinasim/HiFiAdapterFilt"
+      doi: "10.1186/s12864-022-08375-1"
+      licence: ["GPL-3.0"]
+      identifier: biotools:hifiadapterfilt
+
+input:
+  - - meta:
+        type: map
+        description: |
+          Groovy Map containing sample information
+          e.g. `[ id:'sample1', single_end:false ]`
+    - fastq:
+        type: file
+        description: Fastq pacbio hifi raw reads file. The tool AUTOMATICALLY detects fastq files from the input folder, hence an explicit call of "fastq" is not needed.
+        pattern: "*.fastq"
+
+output:
+  - filt:
+      - meta:
+          type: map
+          description: |
+            Groovy Map containing sample information
+            e.g. `[ id:'sample1', single_end:false ]`
+      - "*.filt.fastq.gz":
+          type: file
+          description: |
+            Fastq reads free of pacbio adapter sequence ready for assembly
+          pattern: "*.filt.fastq.gz"
+  - blast_search:
+      - meta:
+          type: map
+          description: |
+            Groovy Map containing sample information
+            e.g. `[ id:'sample1', single_end:false ]`
+      - "*.contaminant.blastout":
+          type: file
+          description: |
+            Output of BLAST search
+          pattern: "*.contaminant.blastout"
+  - stats:
+      - meta:
+          type: map
+          description: |
+            Groovy Map containing sample information
+            e.g. `[ id:'sample1', single_end:false ]`
+      - "*.stats":
+          type: file
+          description: |
+            File with simple math on number of reads removed, etc
+          pattern: "*.stats"
+  - headers:
+      - meta:
+          type: map
+          description: |
+            Groovy Map containing sample information
+            e.g. `[ id:'sample1', single_end:false ]`
+      - "*.blocklist":
+          type: file
+          description: |
+            Headers of pacbio adapter contaminated reads to be removed
+          pattern: "*.blocklist"
+  - versions:
+      - versions.yml:
+          type: file
+          description: File containing software versions
+          pattern: "versions.yml"
+
+authors:
+  - "@Ge94"
+maintainers:
+  - "@Ge94"

modules/nf-core/hifiadapterfilt/tests/main.nf.test

@@ -0,0 +1,54 @@
+nextflow_process {
+
+    name "Test Process HIFIADAPTERFILT"
+    script "../main.nf"
+    process "HIFIADAPTERFILT"
+
+    tag "modules"
+    tag "modules_nfcore"
+    tag "hifiadapterfilt"
+
+    test("hifiadapterfilt - fastq.gz") {
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id:'test' ],
+                    file(params.test_data['homo_sapiens']['pacbio']['hifi'], checkIfExists: true)
+                ]
+                """
+            }
+        }
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(process.out[0][1].findAll {
+                    file(it).name == "test.filt.fastq.gz" &&
+                    file(it).name == "test.contaminant.blastout" &&
+                    file(it).name == "test.blocklist" }).match()},
+                { assert path(process.out.stats[0].get(1)).text.contains('Number of adapter contaminated ccs reads: 0 (0% of total)') },
+                { assert snapshot(process.out.versions).match("versions") }
+            )
+        }
+    }
+
+    test("hifiadapterfilt - fastq.gz - stub") {
+        options "-stub"
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id:'test' ],
+                    file(params.test_data['homo_sapiens']['pacbio']['hifi'], checkIfExists: true)
+                ]
+                """
+            }
+        }
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(process.out).match() }
+            )
+        }
+    }
+}

modules/nf-core/hifiadapterfilt/tests/main.nf.test.snap

@@ -0,0 +1,107 @@
+{
+    "hifiadapterfilt - fastq.gz": {
+        "content": [
+            [
+
+            ]
+        ],
+        "meta": {
+            "nf-test": "0.9.2",
+            "nextflow": "24.10.4"
+        },
+        "timestamp": "2025-01-30T15:23:02.060065966"
+    },
+    "hifiadapterfilt - fastq.gz - stub": {
+        "content": [
+            {
+                "0": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test.filt.fastq.gz:md5,68b329da9893e34099c7d8ad5cb9c940"
+                    ]
+                ],
+                "1": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test.contaminant.blastout:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "2": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test.stats:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "3": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test.blocklist:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "4": [
+                    "versions.yml:md5,2fbf3754b3f590ea1ef7d80105bb1333"
+                ],
+                "blast_search": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test.contaminant.blastout:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "filt": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test.filt.fastq.gz:md5,68b329da9893e34099c7d8ad5cb9c940"
+                    ]
+                ],
+                "headers": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test.blocklist:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "stats": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test.stats:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "versions": [
+                    "versions.yml:md5,2fbf3754b3f590ea1ef7d80105bb1333"
+                ]
+            }
+        ],
+        "meta": {
+            "nf-test": "0.9.2",
+            "nextflow": "24.10.4"
+        },
+        "timestamp": "2025-01-30T15:33:57.890894465"
+    },
+    "versions": {
+        "content": [
+            [
+                "versions.yml:md5,2fbf3754b3f590ea1ef7d80105bb1333"
+            ]
+        ],
+        "meta": {
+            "nf-test": "0.9.2",
+            "nextflow": "24.10.4"
+        },
+        "timestamp": "2025-01-30T15:23:02.135872615"
+    }
+}