Yleaf module added @3.3.0

modules/nf-core/yleaf/environment.yml

@@ -0,0 +1,16 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/environment-schema.json
+channels:
+  - bioconda
+  - conda-forge
+dependencies:
+  - python=3.9
+  - numpy=1.24.3
+  - pandas=1.0.0
+  - bcftools=1.18
+  # Edit this below only once yleaf 3.3.0 has been released to bioconda.
+  # - yleaf=3.3.0
+  # Otherwise, install from my local branch with pip:
+  - pip
+  - pip:
+    - -e /home/a/Yleaf

modules/nf-core/yleaf/main.nf

@@ -0,0 +1,95 @@
+process YLEAF {
+    tag "$meta.id"
+    label 'process_medium'
+
+    // TODO AH: Remove this once yleaf 3.3.0 is released to bioconda
+    conda "nf-core"
+
+    input:
+    tuple val(meta), path(input_file)
+    path reference_fasta, stageAs: "reference.fa"
+    path y_reference_fasta, stageAs: "y_reference.fa"
+    val reference_genome
+    val reads_threshold
+    val quality_thresh
+    val base_majority
+    val prediction_quality
+    val draw_haplogroups
+    val collapsed_draw_mode
+    val ancient_dna
+    val private_mutations
+    val minor_allele_frequency
+
+    output:
+    tuple val(meta), path("${meta.id}/hg_prediction.hg"), emit: haplogroup
+    tuple val(meta), path("${meta.id}/*.log"), emit: log
+    tuple val(meta), path("${meta.id}/hg_tree_image.pdf"), emit: tree
+    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}"
+    def ref_genome = reference_genome ?: "hg38"
+    // These defaults correspond to the defaults used in the Yleaf package
+    def reads_thresh = reads_threshold ?: 10
+    def qual_thresh = quality_thresh ?: 20
+    def base_maj = base_majority ?: 90
+    def pred_qual = prediction_quality ?: 0.95
+    def draw_hg = draw_haplogroups ?: false
+    def collapsed_mode = collapsed_draw_mode ?: false
+    def ancient = ancient_dna ?: false
+    def private_mut = private_mutations ?: false
+    def maf = minor_allele_frequency ?: 0.01
+
+    """
+    mkdir -p ${prefix}
+
+    echo "Input file: ${input_file}"
+    echo "Reference genome: ${ref_genome}"
+
+    PYTHONPATH=\$PWD:/home/a/Yleaf Yleaf \\
+        --vcffile ${input_file} \\
+        --output ${prefix} \\
+        --reference_genome ${ref_genome} \\
+        --full_genome_reference ${reference_fasta} \\
+        --y_chromosome_reference ${y_reference_fasta} \\
+        --threads ${task.cpus} \\
+        --force \\
+        --reanalyze \\
+        --reads_treshold ${reads_thresh} \\
+        --quality_thresh ${qual_thresh} \\
+        --base_majority ${base_maj} \\
+        --prediction_quality ${pred_qual} \\
+        ${draw_hg ? '--draw_haplogroups' : ''} \\
+        ${collapsed_mode ? '--collapsed_draw_mode' : ''} \\
+        ${ancient ? '--ancient_DNA' : ''} \\
+        ${private_mut ? '--private_mutations' : ''} \\
+        --minor_allele_frequency ${maf} \\
+        ${args}
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        yleaf: \$(Yleaf --help | grep "version" | sed 's/.*version //g' | sed 's/).*//g')
+    END_VERSIONS
+    """
+
+    stub:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    """
+    mkdir -p ${prefix}
+
+    touch ${prefix}/run.log
+    touch ${prefix}/hg_prediction.hg
+    touch ${prefix}/hg_tree_image.pdf
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        yleaf: "3.3.0"
+    END_VERSIONS
+    """
+}

modules/nf-core/yleaf/meta.yml

@@ -0,0 +1,65 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/meta-schema.json
+name: "yleaf"
+description: Software for human Y-chromosomal haplogroup inference from next generation sequencing data
+keywords:
+  - genomics
+  - y-chromosome
+  - haplogroup
+  - prediction
+  - next-generation-sequencing
+tools:
+  - "yleaf":
+      description: "Yleaf is software for human Y-chromosomal haplogroup inference from next generation sequencing data"
+      homepage: "https://github.com/genid/Yleaf"
+      documentation: "https://academic.oup.com/mbe/article/35/5/1291/4922696"
+      tool_dev_url: "https://github.com/genid/Yleaf"
+      doi: "10.1093/molbev/msy032"
+      licence: ['GPL-3.0']
+      identifier: "biotools:yleaf"
+
+input:
+  - - meta:
+        type: map
+        description: |
+          Groovy Map containing sample information
+          e.g. `[ id:'sample1', single_end:false ]`
+    - input_file:
+        type: file
+        description: Input sequencing file - VCF
+        pattern: "*.{vcf,vcf.gz}"
+        ontologies:
+          - edam: "http://edamontology.org/format_3016" # VCF
+  - - reference_file:
+        type: file
+        description: Reference genome file
+        pattern: "*.{fa,fasta,fna}"
+        ontologies:
+          - edam: "http://edamontology.org/format_1929" # FASTA
+
+output:
+  - haplogroup:
+    - meta:
+        type: map
+        description: |
+          Groovy Map containing sample information
+          e.g. `[ id:'sample1', single_end:false ]`
+    - "*.hg.txt":
+        type: file
+        description: Predicted Y-chromosome haplogroup output
+        pattern: "*.hg.txt"
+  - log:
+    - "*.log":
+        type: file
+        description: Log file generated by Yleaf
+        pattern: "*.log"
+  - versions:
+    - "versions.yml":
+        type: file
+        description: File containing software versions
+        pattern: "versions.yml"
+
+authors:
+  - "@trianglegrrl"
+maintainers:
+  - "@trianglegrrl"

modules/nf-core/yleaf/tests/environment.yml

@@ -0,0 +1,16 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/environment-schema.json
+channels:
+  - bioconda
+  - conda-forge
+dependencies:
+  - python=3.9
+  - numpy=1.24.3
+  - pandas=1.0.0
+  - bcftools=1.18
+  # Edit this below only once yleaf 3.3.0 has been released to bioconda.
+  # - yleaf=3.3.0
+  # Otherwise, install from my local branch with pip:
+  - pip
+  - pip:
+    - -e /home/a/Yleaf

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

@@ -0,0 +1,90 @@
+nextflow_process {
+
+    name "Test Process YLEAF"
+    script "../main.nf"
+    process "YLEAF"
+
+    tag "modules"
+    tag "modules_nfcore"
+    tag "yleaf"
+
+    def VCF_FILE = "/20tb/2025-01-project-drive/2024-12-19-yhaplo/ancient0003.chrY.vcf.gz"
+    def REF_GENOME_FULL = "/references/reference_genomes/hg38.analysisSet.fa"
+    def REF_GENOME_Y = "/references/reference_genomes/hg38.chrY.analysisSet.fa"
+
+    test("real_data - vcf - with all parameters") {
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id:'ancient0003' ],
+                    file("${VCF_FILE}", checkIfExists: true)
+                ]
+                input[1] = file("${REF_GENOME_FULL}", checkIfExists: true)
+                input[2] = file("${REF_GENOME_Y}", checkIfExists: true)
+                input[3] = "hg38"
+                input[4] = 10
+                input[5] = 20
+                input[6] = 90
+                input[7] = 0.95
+                input[8] = true
+                input[9] = false
+                input[10] = true
+                input[11] = true
+                input[12] = 0.01
+                """
+            }
+            """
+            process.ext.args = "--force --reanalyze"
+            """
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert process.out.versions != null },
+                { assert process.out.haplogroup != null },
+                { assert process.out.tree != null },
+                { assert process.out.log != null }
+            )
+        }
+    }
+
+    test("real_data - vcf - minimal parameters") {
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id:'ancient0003' ],
+                    file("${VCF_FILE}", checkIfExists: true)
+                ]
+                input[1] = file("${REF_GENOME_FULL}", checkIfExists: true)
+                input[2] = file("${REF_GENOME_Y}", checkIfExists: true)
+                input[3] = "hg38"
+                input[4] = null
+                input[5] = null
+                input[6] = null
+                input[7] = null
+                input[8] = null
+                input[9] = null
+                input[10] = null
+                input[11] = null
+                input[12] = null
+                """
+            }
+            """
+            process.ext.args = "--force --reanalyze"
+            """
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert process.out.versions != null },
+                { assert process.out.haplogroup != null },
+                { assert process.out.tree != null },
+                { assert process.out.log != null }
+            )
+        }
+    }
+}

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

@@ -0,0 +1,99 @@
+{
+    "vcf_with_references - stub": {
+        "content": [
+            {
+                "0": [
+                    [
+                        {
+                            "id": "test_sample"
+                        },
+                        "hg_prediction.hg:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "1": [
+                    [
+                        {
+                            "id": "test_sample"
+                        },
+                        "run.log:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "2": [
+                    [
+                        {
+                            "id": "test_sample"
+                        },
+                        "hg_tree_image.pdf:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "3": [
+                    [
+                        {
+                            "id": "test_sample"
+                        },
+                        "test_sample.filtered.out:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "4": [
+                    [
+                        {
+                            "id": "test_sample"
+                        },
+                        "/home/a/nf-core/modules/.nf-test/tests/b999677892a41ca0bc784311e59f5db8/work/e3/a6c0c3f9c52a0daeffc19569c5a905/test_sample/filtered_vcf_files/test_sample.filtered.vcf.gz"
+                    ]
+                ],
+                "5": [
+                    "versions.yml:md5,664e30ba361578a9b5c7f897a3e6d429"
+                ],
+                "filtered_out": [
+                    [
+                        {
+                            "id": "test_sample"
+                        },
+                        "test_sample.filtered.out:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "filtered_vcf": [
+                    [
+                        {
+                            "id": "test_sample"
+                        },
+                        "/home/a/nf-core/modules/.nf-test/tests/b999677892a41ca0bc784311e59f5db8/work/e3/a6c0c3f9c52a0daeffc19569c5a905/test_sample/filtered_vcf_files/test_sample.filtered.vcf.gz"
+                    ]
+                ],
+                "haplogroup": [
+                    [
+                        {
+                            "id": "test_sample"
+                        },
+                        "hg_prediction.hg:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "log": [
+                    [
+                        {
+                            "id": "test_sample"
+                        },
+                        "run.log:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "tree": [
+                    [
+                        {
+                            "id": "test_sample"
+                        },
+                        "hg_tree_image.pdf:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "versions": [
+                    "versions.yml:md5,664e30ba361578a9b5c7f897a3e6d429"
+                ]
+            }
+        ],
+        "meta": {
+            "nf-test": "0.9.2",
+            "nextflow": "24.10.5"
+        },
+        "timestamp": "2025-04-02T05:49:24.004827"
+    }
+}