Vcf2cytosure (nf-core#4312)

* adding vcf2cytosure * fixing test * updating meta
louperelo · Nov 8, 2023 · 2dba354 · 2dba354
1 parent 4ef7bec
commit 2dba354
Show file tree

Hide file tree

Showing 8 changed files with 188 additions and 0 deletions.
diff --git a/modules/nf-core/vcf2cytosure/environment.yml b/modules/nf-core/vcf2cytosure/environment.yml
@@ -0,0 +1,6 @@
+channels:
+  - conda-forge
+  - bioconda
+  - defaults
+dependencies:
+  - bioconda::vcf2cytosure=0.9.1
diff --git a/modules/nf-core/vcf2cytosure/main.nf b/modules/nf-core/vcf2cytosure/main.nf
@@ -0,0 +1,66 @@
+process VCF2CYTOSURE {
+    tag "$meta.id"
+    label 'process_single'
+
+    conda "${moduleDir}/environment.yml"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/vcf2cytosure:0.9.1--pyh7cba7a3_1':
+        'biocontainers/vcf2cytosure:0.9.1--pyh7cba7a3_1' }"
+
+    input:
+    tuple val(meta), path(sv_vcf)
+    tuple val(meta2), path(coverage_bed)
+    tuple val(meta3), path(cns)
+    tuple val(meta4), path(snv_vcf)
+    path blacklist_bed
+
+    output:
+    tuple val(meta), path("*.cgh"), emit: cgh
+    path "versions.yml"           , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def coverage = coverage_bed ? "--coverage ${coverage_bed}" : ''
+    def cnvkit = cns ? ( coverage_bed ? '' : "--cn ${cns}" ) : ''
+    def snv = snv_vcf ? ( coverage_bed ? '' : "--snv ${snv_vcf}" ) : ''
+    def blacklist = blacklist_bed ? "--blacklist ${blacklist_bed}" : ''
+    def prefix = task.ext.prefix ?: sv_vcf ? "${meta.id}" : "${meta3.id}"
+
+    if ( cns && coverage_bed || snv_vcf && coverage_bed ) error "Coverage_bed input is not compatible with cns and snv"
+
+    """
+    vcf2cytosure \\
+        --vcf $sv_vcf \\
+        --out ${prefix}.cgh \\
+        $coverage \\
+        $cnvkit \\
+        $snv \\
+        $blacklist \\
+        $args
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        vcf2cytosure: \$(echo \$(vcf2cytosure --version 2>&1) | sed 's/^.* cytosure //' )
+    END_VERSIONS
+    """
+
+    stub:
+    def args = task.ext.args ?: ''
+    def coverage = coverage_bed ? "--coverage ${coverage_bed}" : ''
+    def cnvkit = cns ? ( coverage_bed ? '' : "--cn ${cns}" ) : ''
+    def snv = snv_vcf ? ( coverage_bed ? '' : "--snv ${snv_vcf}" ) : ''
+    def blacklist = blacklist_bed ? "--blacklist ${blacklist_bed}" : ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    """
+    touch ${prefix}.xml
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        vcf2cytosure: \$(echo \$(vcf2cytosure --version 2>&1) | sed 's/^.* cytosure //' )
+    END_VERSIONS
+    """
+}
diff --git a/modules/nf-core/vcf2cytosure/meta.yml b/modules/nf-core/vcf2cytosure/meta.yml
@@ -0,0 +1,77 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/yaml-schema.json
+name: "vcf2cytosure"
+description: Convert VCF with structural variations to CytoSure format
+keywords:
+  - structural_variants
+  - array_cgh
+  - vcf
+  - cytosure
+tools:
+  - "vcf2cytosure":
+      description: "Convert VCF with structural variations to CytoSure format"
+      homepage: "https://github.com/NBISweden/vcf2cytosure"
+      documentation: "https://github.com/NBISweden/vcf2cytosure"
+      tool_dev_url: "https://github.com/NBISweden/vcf2cytosure"
+      licence: "['MIT']"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'test', single_end:false ]`
+  - meta2:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'test', single_end:false ]`
+  - meta3:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'test', single_end:false ]`
+  - meta4:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'test', single_end:false ]`
+  - sv_vcf:
+      type: file
+      description: VCF file with structural variants
+      pattern: "*.{vcf,vcf.gz}"
+  - coverage_bed:
+      type: file
+      description: Bed file with coverage data
+      pattern: "*.bed"
+  - cns:
+      type: file
+      description: CN file from CNVkit, not compatible with coverage_bed file
+  - snv_vcf:
+      type: file
+      description: |
+        VCF file with SNVs to calculate probe coverage,
+        not compatible with coverage_bed
+        pattern: "*.{vcf,vcf.gz}"
+  - blacklilst_bed:
+      type: file
+      description: Bed file with regions to exclude
+      pattern: "*.bed"
+
+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"
+  - cgh:
+      type: file
+      description: SV:s in CytoSure format
+      pattern: "*.cgh"
+
+authors:
+  - "@jemten"
diff --git a/tests/config/pytest_modules.yml b/tests/config/pytest_modules.yml
@@ -4031,6 +4031,10 @@ varlociraptor/preprocess:
   - modules/nf-core/varlociraptor/preprocess/**
   - tests/modules/nf-core/varlociraptor/preprocess/**
 
+vcf2cytosure:
+  - modules/nf-core/vcf2cytosure/**
+  - tests/modules/nf-core/vcf2cytosure/**
+
 vcf2db:
   - modules/nf-core/vcf2db/**
   - tests/modules/nf-core/vcf2db/**

diff --git a/tests/config/test_data.config b/tests/config/test_data.config
@@ -465,6 +465,7 @@ params {
 
                 cutandrun_bedgraph_test_1                               = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/bedgraph/cutandtag_h3k27me3_test_1.bedGraph"
                 cutandrun_bedgraph_test_2                               = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/bedgraph/cutandtag_igg_test_1.bedGraph"
+                na24385_chr22_coverage                                  = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/bedgraph/NA24385_coverage.bed"
 
                 empty_vcf_gz                                            = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/vcf/empty.vcf.gz"
                 empty_vcf_gz_tbi                                        = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/vcf/empty.vcf.gz.tbi"
@@ -477,6 +478,8 @@ params {
                 test_rnaseq_vcf                                         = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/vcf/test.rnaseq.vcf"
                 test_sv_vcf                                             = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/vcf/sv_query.vcf.gz"
                 test_sv_vcf_tbi                                         = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/vcf/sv_query.vcf.gz.tbi"
+                na24385_chr22_sv_vcf                                    = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/vcf/NA24385_sv.vcf.gz"
+                na24385_chr22_sv_vcf_tbi                                = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/vcf/NA24385_sv.vcf.gz.tbi"
                 genmod_vcf_gz                                           = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/vcf/genmod.vcf.gz"
                 genmod_annotate_vcf_gz                                  = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/vcf/test_annotate.vcf.gz"
                 genmod_models_vcf_gz                                    = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/vcf/test_models.vcf.gz"

diff --git a/tests/modules/nf-core/vcf2cytosure/main.nf b/tests/modules/nf-core/vcf2cytosure/main.nf
@@ -0,0 +1,19 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { VCF2CYTOSURE } from '../../../../modules/nf-core/vcf2cytosure/main.nf'
+
+workflow test_vcf2cytosure {
+
+    sv_vcf = [
+        [ id:'test' ], // meta map
+        [ file(params.test_data['homo_sapiens']['illumina']['na24385_chr22_sv_vcf'], checkIfExists: true) ]
+    ]
+    coverage = [
+        [ id:'test' ], // meta map
+        [ file(params.test_data['homo_sapiens']['illumina']['na24385_chr22_coverage'], checkIfExists: true) ]
+    ]
+    VCF2CYTOSURE ( sv_vcf, coverage, [[],[]], [[],[]], [] )
+
+}
diff --git a/tests/modules/nf-core/vcf2cytosure/nextflow.config b/tests/modules/nf-core/vcf2cytosure/nextflow.config
@@ -0,0 +1,5 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+}
diff --git a/tests/modules/nf-core/vcf2cytosure/test.yml b/tests/modules/nf-core/vcf2cytosure/test.yml
@@ -0,0 +1,8 @@
+- name: vcf2cytosure test_vcf2cytosure
+  command: nextflow run ./tests/modules/nf-core/vcf2cytosure -entry test_vcf2cytosure -c ./tests/config/nextflow.config
+  tags:
+    - vcf2cytosure
+  files:
+    - path: output/vcf2cytosure/test.cgh
+      md5sum: fc3f1ffac5b797881d992994d5d56550
+    - path: output/vcf2cytosure/versions.yml