added gvcftools/extractvariants (#1924)

* added gvcftools/extractvariants

* linting

* Update modules/gvcftools/extractvariants/meta.yml

Co-authored-by: Júlia Mir Pedrol <mirp.julia@gmail.com>

* Update modules/gvcftools/extractvariants/main.nf

Co-authored-by: Júlia Mir Pedrol <mirp.julia@gmail.com>

* added uncompressed input handling

Co-authored-by: Júlia Mir Pedrol <mirp.julia@gmail.com>

modules/gvcftools/extractvariants/main.nf

@@ -0,0 +1,37 @@
+process GVCFTOOLS_EXTRACTVARIANTS {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda (params.enable_conda ? "bioconda::gvcftools=0.17.0" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/gvcftools:0.17.0--he941832_3':
+        'quay.io/biocontainers/gvcftools:0.17.0--he941832_3' }"
+
+    input:
+    tuple val(meta), path(gvcf)
+
+    output:
+    tuple val(meta), path("*.vcf.gz"), emit: vcf
+    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 open_gvcf = gvcf.extension == "gz" ? "gzip -dc $gvcf" : "cat $gvcf"
+
+    """
+    $open_gvcf |
+    extract_variants \\
+        $args \\
+        $gvcf |
+    gzip -c > ${prefix}.vcf.gz
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        gvcftools: \$(extract_variants --help 2>&1 | grep version | sed 's/version: //')
+    END_VERSIONS
+    """
+}

modules/gvcftools/extractvariants/meta.yml

@@ -0,0 +1,45 @@
+name: "gvcftools_extractvariants"
+description: Removes all non-variant blocks from a gVCF file to produce a smaller variant-only VCF file.
+keywords:
+  - gvcftools
+  - extract_variants
+  - extractvariants
+  - gvcf
+  - vcf
+tools:
+  - "gvcftools":
+      description: "gvcftools is a package of small utilities for creating and analyzing gVCF files"
+      homepage: "{https://sites.google.com/site/gvcftools/home}"
+      documentation: "{https://sites.google.com/site/gvcftools/home/configuration-and-analysis}"
+      tool_dev_url: "{https://github.com/sequencing/gvcftools}"
+      doi: ""
+      licence: "['MIT']"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - gvcf:
+      type: file
+      description: GVCF file
+      pattern: "*.{g.vcf,gvcf}.gz"
+
+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: Converted variant-only VCF file
+      pattern: "*.vcf.gz"
+
+authors:
+  - "@nvnieuwk"

tests/config/pytest_modules.yml

@@ -1031,6 +1031,10 @@ gunzip:
   - modules/gunzip/**
   - tests/modules/gunzip/**
 
+gvcftools/extractvariants:
+  - modules/gvcftools/extractvariants/**
+  - tests/modules/gvcftools/extractvariants/**
+
 hamronization/amrfinderplus:
   - modules/hamronization/amrfinderplus/**
   - tests/modules/hamronization/amrfinderplus/**

tests/modules/gvcftools/extractvariants/main.nf

@@ -0,0 +1,25 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { GVCFTOOLS_EXTRACTVARIANTS } from '../../../../modules/gvcftools/extractvariants/main.nf'
+
+workflow test_gvcftools_extractvariants {
+    
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test_genome_vcf_gz'], checkIfExists: true)
+    ]
+
+    GVCFTOOLS_EXTRACTVARIANTS ( input )
+}
+
+workflow test_gvcftools_extractvariants_uncompressed {
+    
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test_genome_vcf'], checkIfExists: true)
+    ]
+
+    GVCFTOOLS_EXTRACTVARIANTS ( input )
+}

tests/modules/gvcftools/extractvariants/nextflow.config

@@ -0,0 +1,5 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+    
+}

tests/modules/gvcftools/extractvariants/test.yml

@@ -0,0 +1,15 @@
+- name: gvcftools extractvariants test_gvcftools_extractvariants
+  command: nextflow run ./tests/modules/gvcftools/extractvariants -entry test_gvcftools_extractvariants -c ./tests/config/nextflow.config  -c ./tests/modules/gvcftools/extractvariants/nextflow.config
+  tags:
+    - gvcftools
+    - gvcftools/extractvariants
+  files:
+    - path: output/gvcftools/test.vcf.gz
+
+- name: gvcftools extractvariants test_gvcftools_extractvariants_uncompressed
+  command: nextflow run ./tests/modules/gvcftools/extractvariants -entry test_gvcftools_extractvariants_uncompressed -c ./tests/config/nextflow.config  -c ./tests/modules/gvcftools/extractvariants/nextflow.config
+  tags:
+    - gvcftools
+    - gvcftools/extractvariants
+  files:
+    - path: output/gvcftools/test.vcf.gz