new module survivor/merge (#2648)

* new module survivor/merge * fix versions * review updates * add stubs * fix tests * changed conda according to #2654 * moved the args to input vals * update meta * update to latest decision * fix test * Update modules/nf-core/survivor/merge/main.nf Co-authored-by: Jonathan Manning <[email protected]> --------- Co-authored-by: Jonathan Manning <[email protected]>
nf-core · Feb 28, 2023 · 61cdc61 · 61cdc61
1 parent 40ce6b3
commit 61cdc61
Show file tree

Hide file tree

Showing 6 changed files with 181 additions and 0 deletions.
diff --git a/modules/nf-core/survivor/merge/main.nf b/modules/nf-core/survivor/merge/main.nf
@@ -0,0 +1,71 @@
+process SURVIVOR_MERGE {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda "bioconda::survivor=1.0.7"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/survivor:1.0.7--h9a82719_1':
+        'quay.io/biocontainers/survivor:1.0.7--h9a82719_1' }"
+
+    input:
+    tuple val(meta), path(vcfs)
+    val(max_distance_breakpoints)
+    val(min_supporting_callers)
+    val(account_for_type)
+    val(account_for_sv_strands)
+    val(estimate_distanced_by_sv_size)
+    val(min_sv_size)
+
+    output:
+    tuple val(meta), path("*.vcf")   , emit: vcf
+    path "versions.yml"              , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    vcfs.each{
+        if (it.getExtension() == "gz"){
+            error "Gzipped files are not supported by Survivor, please gunzip your VCF files first."
+            // https://github.com/fritzsedlazeck/SURVIVOR/issues/158
+        }
+    }
+
+    """
+    SURVIVOR merge \\
+        <(ls *.vcf) \\
+        ${max_distance_breakpoints} \\
+        ${min_supporting_callers} \\
+        ${account_for_type} \\
+        ${account_for_sv_strands} \\
+        ${estimate_distanced_by_sv_size} \\
+        ${min_sv_size} \\
+        ${prefix}.vcf
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        survivor: \$(echo \$(SURVIVOR 2>&1 | grep "Version" | sed 's/^Version: //'))
+    END_VERSIONS
+    """
+
+    stub:
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    vcfs.each{
+        if (it.getExtension() == "gz"){
+            error "Gzipped files are not supported by Survivor, please gunzip your VCF files first."
+            // https://github.com/fritzsedlazeck/SURVIVOR/issues/158
+        }
+    }
+
+    """
+    touch ${prefix}.vcf
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        survivor: \$(echo \$(SURVIVOR 2>&1 | grep "Version" | sed 's/^Version: //'))
+    END_VERSIONS
+    """
+}
diff --git a/modules/nf-core/survivor/merge/meta.yml b/modules/nf-core/survivor/merge/meta.yml
@@ -0,0 +1,64 @@
+name: "survivor_merge"
+description: Compare or merge VCF files to generate a consensus or multi sample VCF files.
+keywords:
+  - survivor
+  - merge
+  - vcf
+  - structural variants
+tools:
+  - "survivor":
+      description: "Toolset for SV simulation, comparison and filtering"
+      homepage: "https://github.com/fritzsedlazeck/SURVIVOR/wiki"
+      documentation: "https://github.com/fritzsedlazeck/SURVIVOR/wiki"
+      tool_dev_url: "https://github.com/fritzsedlazeck/SURVIVOR"
+      doi: "10.1038/NCOMMS14061"
+      licence: "['MIT']"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - vcfs:
+      type: files
+      description: |
+        The VCF files to be merged
+        Gzipped VCF files are not supported: https://github.com/fritzsedlazeck/SURVIVOR/issues/158
+      pattern: "*.vcf"
+  - max_distance_breakpoints:
+      type: integer
+      description: Max distance between breakpoints (0-1 percent of length, 1- number of bp)
+  - min_supporting_callers:
+      type: integer
+      description: Minimum number of supporting caller
+  - account_for_type:
+      type: integer
+      description: Take the type into account (1==yes, else no)
+  - account_for_sv_strands:
+      type: integer
+      description: Take the strands of SVs into account (1==yes, else no)
+  - estimate_distanced_by_sv_size:
+      type: integer
+      description: Estimate distance based on the size of SV (1==yes, else no)
+  - min_sv_size:
+      type: integer
+      description: Minimum size of SVs to be taken into account
+
+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"
+  - vcf:
+      type: file
+      description: The merged VCF file
+      pattern: "*.vcf"
+
+authors:
+  - "@nvnieuwk"
diff --git a/tests/config/pytest_modules.yml b/tests/config/pytest_modules.yml
@@ -3135,6 +3135,10 @@ subworkflows/vcf_gather_bcftools:
   - subworkflows/nf-core/vcf_gather_bcftools/**
   - tests/subworkflows/nf-core/vcf_gather_bcftools/**
 
+survivor/merge:
+  - modules/nf-core/survivor/merge/**
+  - tests/modules/nf-core/survivor/merge/**
+
 subworkflows/vcf_impute_glimpse:
   - subworkflows/nf-core/vcf_impute_glimpse/**
   - tests/subworkflows/nf-core/vcf_impute_glimpse/**

diff --git a/tests/modules/nf-core/survivor/merge/main.nf b/tests/modules/nf-core/survivor/merge/main.nf
@@ -0,0 +1,26 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { SURVIVOR_MERGE } from '../../../../../modules/nf-core/survivor/merge/main.nf'
+
+workflow test_survivor_merge {
+
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        [
+            file(params.test_data['homo_sapiens']['illumina']['test2_genome_vcf'], checkIfExists: true),
+            file(params.test_data['homo_sapiens']['illumina']['test_genome_vcf'], checkIfExists: true)
+        ]
+    ]
+
+    SURVIVOR_MERGE (
+        input,
+        0.2,
+        1,
+        0,
+        0,
+        0,
+        20
+    )
+}
diff --git a/tests/modules/nf-core/survivor/merge/nextflow.config b/tests/modules/nf-core/survivor/merge/nextflow.config
@@ -0,0 +1,5 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+}
diff --git a/tests/modules/nf-core/survivor/merge/test.yml b/tests/modules/nf-core/survivor/merge/test.yml
@@ -0,0 +1,11 @@
+- name: survivor merge test_survivor_merge
+  command: nextflow run ./tests/modules/nf-core/survivor/merge -entry test_survivor_merge -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/survivor/merge/nextflow.config
+  tags:
+    - survivor
+    - survivor/merge
+  files:
+    - path: output/survivor/test.vcf
+      contains:
+        - "##fileformat=VCFv4.1"
+        - "##source=SURVIVOR"
+    - path: output/survivor/versions.yml