Strelka germline (#340)

* nf-core template created

* boilerplate and sarek_dsl2 code merged

* adding an option to give it args

* bai got away

* seperating vcf files and vcf index files into seperate streams

* some minor spacefixes

* adding standard information about the module

* removing typos

* some basic tests based on tiddit tests

* removed the bed parameter, should be provided via options.args instead

* removing typos

* adding indexed bam file instead

* Adding changing out fasta with reference to deal with empty fasta input

* adding the correct fasta

* removing the empty test

* adding the correct data to the testoutput and removing the md5sum since
it constantly changes

* adding  target_bed to input

* adding info on target bed

* adding target bed to test

* adding more files to the test

* adding meta for target bed test

* adding a test for target_bed

* typo

* fixing pytest with master

software/strelka/germline/functions.nf

@@ -0,0 +1,60 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args          = args.args ?: ''
+    options.args2         = args.args2 ?: ''
+    options.args3         = args.args3 ?: ''
+    options.publish_by_id = args.publish_by_id ?: false
+    options.publish_dir   = args.publish_dir ?: ''
+    options.publish_files = args.publish_files
+    options.suffix        = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }  // Remove empty entries
+    paths = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_id) {
+            path_list.add(args.publish_id)
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

software/strelka/germline/main.nf

@@ -0,0 +1,54 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process STRELKA_GERMLINE {
+    tag "$meta.id"
+    label 'process_high'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), publish_id:meta.id) }
+
+    conda (params.enable_conda ? "bioconda::strelka=2.9.10" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/strelka:2.9.10--0"
+    } else {
+        container "quay.io/biocontainers/strelka:2.9.10--0"
+    }
+
+    input:
+    tuple val(meta), path(bam), path(bai)
+    path fasta
+    path fai
+    path target_bed
+
+    output:
+    tuple val(meta), path("*_variants.vcf.gz"), emit: vcf
+    tuple val(meta), path("*_variants.vcf.gz.tbi"), emit: vcf_tbi
+    tuple val(meta), path("*_genome.vcf.gz"), emit: genome_vcf
+    tuple val(meta), path("*_genome.vcf.gz.tbi"), emit: genome_vcf_tbi
+    path "*.version.txt", emit: version 
+
+    script:
+    def software = getSoftwareName(task.process)
+    def ioptions = initOptions(options)
+    def prefix   = ioptions.suffix ? "strelka_${meta.id}${ioptions.suffix}" : "strelka_${meta.id}"
+    
+    options_strelka = params.target_bed ? "--exome --callRegions ${target_bed}" : ""
+    """
+    configureStrelkaGermlineWorkflow.py \
+        --bam ${bam} \
+        --referenceFasta ${fasta} \
+        ${options_strelka} \
+        ${options.args} \
+        --runDir strelka
+    python strelka/runWorkflow.py -m local -j ${task.cpus}
+    mv strelka/results/variants/genome.*.vcf.gz     ${prefix}_genome.vcf.gz
+    mv strelka/results/variants/genome.*.vcf.gz.tbi ${prefix}_genome.vcf.gz.tbi
+    mv strelka/results/variants/variants.vcf.gz     ${prefix}_variants.vcf.gz
+    mv strelka/results/variants/variants.vcf.gz.tbi ${prefix}_variants.vcf.gz.tbi
+    echo configureStrelkaGermlineWorkflow.py --version &> ${software}.version.txt #2>&1
+    """
+}

software/strelka/germline/meta.yml

@@ -0,0 +1,63 @@
+name: strelka_germline
+description: Strelka2 is a fast and accurate small variant caller optimized for analysis of germline variation 
+keywords:
+  - variantcalling
+  - germline
+  - wgs
+  - vcf
+  - variants
+tools:
+  - strelka:
+      description: Strelka calls somatic and germline small variants from mapped sequencing reads
+      homepage: https://github.com/Illumina/strelka
+      documentation: https://github.com/Illumina/strelka/blob/v2.9.x/docs/userGuide/README.md
+      tool_dev_url: https://github.com/Illumina/strelka
+      doi: 10.1038/s41592-018-0051-x
+      licence: ['GPL v3']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test']
+  - bam:
+      type: file
+      description: BAM file
+      pattern: "*.{bam}"
+  - bai:
+      type: file
+      description: BAM index file
+      pattern: "*.{bai}"
+  - target_bed:
+      type: file
+      description: An optional bed file
+      pattern: "*.{bed}"
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test']
+  - vcf:
+      type: file
+      description: gzipped germline variant file
+      pattern: "*.{vcf.gz}"
+  - vcf_tbi:
+     type: file
+     description: index file for the vcf file
+     pattern: "*.vcf.gz.tbi"
+  - genome_vcf:
+     type: file
+     description: variant records and compressed non-variant blocks
+     pattern: "*_genome.vcf.gz"
+  - genome_vcf_tbi:
+     type: file
+     description: index file for the genome_vcf file
+     pattern: "*_genome.vcf.gz.tbi"
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+authors:
+  - "@arontommi"

tests/config/pytest_software.yml

@@ -411,6 +411,10 @@ star_genomegenerate:
   - software/star/genomegenerate/**
   - tests/software/star/genomegenerate/**
 
+strelka_germline:
+  - software/strelka/germline/**
+  - tests/software/strelka/germline/**
+
 stringtie:
   - software/stringtie/**
   - tests/software/stringtie/**

tests/software/strelka/germline/main.nf

@@ -0,0 +1,33 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { STRELKA_GERMLINE } from '../../../../software/strelka/germline/main.nf' addParams( options: [:] )
+
+workflow test_strelka_germline {
+    
+    def input = []
+
+    def fasta = file("${launchDir}/tests/data/genomics/sarscov2/fasta/test_genome.fasta", checkIfExists: true)
+    def fai   = file("${launchDir}/tests/data/genomics/sarscov2/fasta/test_genome.fasta.fai", checkIfExists: true)
+    def target_bed = []
+    input = [ [ id:'test'], // meta map
+              file("${launchDir}/tests/data/genomics/sarscov2/bam/test_paired_end.sorted.bam", checkIfExists: true),
+              file("${launchDir}/tests/data/genomics/sarscov2/bam/test_paired_end.sorted.bam.bai", checkIfExists: true) ]
+
+    STRELKA_GERMLINE ( input, fasta, fai, target_bed )
+}
+workflow test_strelka_germline_target_bed {
+    
+    def input = []
+
+    def fasta = file("${launchDir}/tests/data/genomics/sarscov2/fasta/test_genome.fasta", checkIfExists: true)
+    def fai   = file("${launchDir}/tests/data/genomics/sarscov2/fasta/test_genome.fasta.fai", checkIfExists: true)
+    def target_bed = file("${launchDir}/tests/data/genomics/sarscov2/bed/test.bed", checkIfExists: true)
+    input = [ [ id:'test'], // meta map
+              file("${launchDir}/tests/data/genomics/sarscov2/bam/test_paired_end.sorted.bam", checkIfExists: true),
+              file("${launchDir}/tests/data/genomics/sarscov2/bam/test_paired_end.sorted.bam.bai", checkIfExists: true) ]
+
+    STRELKA_GERMLINE ( input, fasta, fai, target_bed )
+}
+

tests/software/strelka/germline/test.yml

@@ -0,0 +1,22 @@
+## TODO nf-core: Please run the following command to build this file:
+#                nf-core modules create-test-yml strelka/germline
+- name: strelka germline
+  command: nextflow run ./tests/software/strelka/germline -entry test_strelka_germline -c tests/config/nextflow.config
+  tags:
+    - strelka
+    - strelka_germline
+  files:
+    - path: output/strelka/strelka_test_variants.vcf.gz
+    - path: output/strelka/strelka_test_variants.vcf.gz.tbi
+    - path: output/strelka/strelka_test_genome.vcf.gz
+    - path: output/strelka/strelka_test_genome.vcf.gz.tbi
+- name: strelka germline target bed
+  command: nextflow run ./tests/software/strelka/germline -entry test_strelka_germline_target_bed -c tests/config/nextflow.config
+  tags:
+    - strelka
+    - strelka_germline
+  files:
+    - path: output/strelka/strelka_test_variants.vcf.gz
+    - path: output/strelka/strelka_test_variants.vcf.gz.tbi
+    - path: output/strelka/strelka_test_genome.vcf.gz
+    - path: output/strelka/strelka_test_genome.vcf.gz.tbi