Add dsl2 module for deepvariant

README.md

@@ -89,6 +89,7 @@ Main authors:
 
 Helpful contributors:
 
+* [Abhinav Sharma](https://github.com/abhi18av)
 * [Adrian Lärkeryd](https://github.com/adrlar)
 * [Alexander Peltzer](https://github.com/apeltzer)
 * [Chela James](https://github.com/chelauk)

bin/scrape_software_versions.py

@@ -14,6 +14,7 @@
     "FastQC": ["v_fastqc.txt", r"FastQC v(\S+)"],
     "FreeBayes": ["v_freebayes.txt", r"version:  v(\d\.\d\.\d+)"],
     "GATK": ["v_gatk.txt", r"Version:(\S+)"],
+    "DeepVariant": ["v_deepvariant.txt", r"Version:(\S+)"],
     "htslib": ["v_samtools.txt", r"htslib (\S+)"],
     "Manta": ["v_manta.txt", r"([0-9.]+)"],
     "MultiQC": ["v_multiqc.txt", r"multiqc, version (\S+)"],

conf/modules.config

@@ -178,6 +178,13 @@ params {
             publish_dir      = 'variant_calling'
             publish_files    = ['vcf.gz':'strelka', 'vcf.gz.tbi':'strelka']
         }
+        'deepvariant' {
+            args             = '--model_type=WGS'
+            publish_by_meta  = true
+            publish_dir      = 'variant_calling'
+            publish_files    = false
+        }
+
 // TUMOR_VARIANT_CALLING
 
 // PAIR_VARIANT_CALLING

conf/test.config

@@ -58,7 +58,7 @@ profiles {
   tool_pair {
     params.input               = 'https://raw.githubusercontent.com/nf-core/test-datasets/sarek/testdata/csv/tiny-recal-pair-https.csv'
     params.step                = 'variant_calling'
-    }
+  }
   trimming {
     params.clip_r1             = 1
     params.clip_r2             = 1

modules/local/deepvariant/functions.nf

@@ -0,0 +1,70 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args            = args.args ?: ''
+    options.args2           = args.args2 ?: ''
+    options.args3           = args.args3 ?: ''
+    options.publish_by_meta = args.publish_by_meta ?: []
+    options.publish_dir     = args.publish_dir ?: ''
+    options.publish_files   = args.publish_files
+    options.suffix          = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
+    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions  = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_meta) {
+            def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
+            for (key in key_list) {
+                if (args.meta && key instanceof String) {
+                    def path = key
+                    if (args.meta.containsKey(key)) {
+                        path = args.meta[key] instanceof Boolean ? "${key}_${args.meta[key]}".toString() : args.meta[key]
+                    }
+                    path = path instanceof String ? path : ''
+                    path_list.add(path)
+                }
+            }
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

modules/local/deepvariant/main.nf

@@ -0,0 +1,61 @@
+// TODO Remove the module declaration
+nextflow.enable.dsl=2
+
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+// TODO nf-core: A module file SHOULD only define input and output files as command-line parameters.
+//               All other parameters MUST be provided as a string i.e. "options.args"
+//               where "params.options" is a Groovy Map that MUST be provided via the addParams section of the including workflow.
+//               Any parameters that need to be evaluated in the context of a particular sample
+//               e.g. single-end/paired-end data MUST also be defined and evaluated appropriately.
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process DEEPVARIANT {
+    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), meta:meta, publish_by_meta:['id']) }
+
+    conda (params.enable_conda ? "bioconda::deepvariant=1.1.0" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/deepvariant:1.1.0--py36hf3e76ba_2"
+    } else {
+        // TODO update the bioconda container to work with run_deepvariant.sh script
+        // container "quay.io/biocontainers/deepvariant:1.1.0--py36hf3e76ba_2"
+        container "google/deepvariant:1.1.0"
+    }
+
+    input:
+    // TODO nf-core: Where applicable all sample-specific information e.g. "id", "single_end", "read_group"
+    //               MUST be provided as an input via a Groovy Map called "meta".
+    //               This information may not be required in some instances e.g. indexing reference genome files:
+    //               https://github.com/nf-core/modules/blob/master/software/bwa/index/main.nf
+    tuple val(meta), path(bam), path(bai)
+    path fasta
+    path fai
+
+    output:
+    tuple val(meta), path("*.vcf.gz"),  emit: vcf
+    tuple val(meta), path("*g.vcf.gz"),  emit: gvcf
+    path "*.version.txt"          , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    """
+    /opt/deepvariant/bin/run_deepvariant \\
+        --ref=${fasta} \\
+        --reads=${bam} \\
+        --output_vcf=${prefix}.vcf.gz \\
+        --output_gvcf=${prefix}.g.vcf.gz \\
+        --num_shards=${task.cpus} \\
+        ${options.args}
+
+    echo \$(/opt/deepvariant/bin/run_deepvariant --version)  > ${software}.version.txt
+    """
+
+}

subworkflows/local/germline_variant_calling.nf

@@ -5,12 +5,14 @@
 */
 
 params.haplotypecaller_options        = [:]
+params.deepvariant_options            = [:]
 params.genotypegvcf_options           = [:]
 params.concat_gvcf_options            = [:]
 params.concat_haplotypecaller_options = [:]
 params.strelka_options                = [:]
 
 include { GATK4_HAPLOTYPECALLER as HAPLOTYPECALLER } from '../../modules/nf-core/software/gatk4/haplotypecaller/main' addParams(options: params.haplotypecaller_options)
+include { DEEPVARIANT }                              from '../../modules/local/deepvariant/main'                      addParams(options: params.deepvariant_options)
 include { GATK4_GENOTYPEGVCF as GENOTYPEGVCF }       from '../../modules/nf-core/software/gatk4/genotypegvcf/main'    addParams(options: params.genotypegvcf_options)
 include { CONCAT_VCF as CONCAT_GVCF }                from '../../modules/local/concat_vcf/main'                       addParams(options: params.concat_gvcf_options)
 include { CONCAT_VCF as CONCAT_HAPLOTYPECALLER }     from '../../modules/local/concat_vcf/main'                       addParams(options: params.concat_haplotypecaller_options)
@@ -94,6 +96,17 @@ workflow GERMLINE_VARIANT_CALLING {
         haplotypecaller_vcf = CONCAT_HAPLOTYPECALLER.out.vcf
     }
 
+    if ('deepvariant' in params.tools.toLowerCase()) {
+
+        DEEPVARIANT(
+            bam,
+            fasta,
+            fai)
+
+        haplotypecaller_vcf = DEEPVARIANT.out.vcf
+        haplotypecaller_gvcf = DEEPVARIANT.out.gvcf
+    }
+
     if ('strelka' in params.tools.toLowerCase()) {
         STRELKA(
             bam,

workflows/sarek.nf

@@ -259,7 +259,7 @@ workflow SAREK {
     msisensorpro_scan = BUILD_INDICES.out.msisensorpro_scan
     target_bed_gz_tbi = BUILD_INDICES.out.target_bed_gz_tbi
 
-    // PREPREOCESSING
+    // PREPROCESSING
 
     bam_mapped          = Channel.empty()
     bam_mapped_qc       = Channel.empty()