Add gatk4/estimatelibrarycomplexity (#965)

* Add gatk4/estimatelibrarycomplxity

* forgot to add yml

* Fix placeholder

modules/gatk4/estimatelibrarycomplexity/functions.nf

@@ -0,0 +1,78 @@
+//
+//  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()
+}
+
+//
+// Extract name of module from process name using $task.process
+//
+def getProcessName(task_process) {
+    return task_process.tokenize(':')[-1]
+}
+
+//
+// Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+//
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args            = args.args ?: ''
+    options.args2           = args.args2 ?: ''
+    options.args3           = args.args3 ?: ''
+    options.publish_by_meta = args.publish_by_meta ?: []
+    options.publish_dir     = args.publish_dir ?: ''
+    options.publish_files   = args.publish_files
+    options.suffix          = args.suffix ?: ''
+    return options
+}
+
+//
+// Tidy up and join elements of a list to return a path string
+//
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
+    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+//
+// Function to save/publish module results
+//
+def saveFiles(Map args) {
+    def ioptions  = initOptions(args.options)
+    def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+
+    // Do not publish versions.yml unless running from pytest workflow
+    if (args.filename.equals('versions.yml') && !System.getenv("NF_CORE_MODULES_TEST")) {
+        return null
+    }
+    if (ioptions.publish_by_meta) {
+        def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
+        for (key in key_list) {
+            if (args.meta && key instanceof String) {
+                def path = key
+                if (args.meta.containsKey(key)) {
+                    path = args.meta[key] instanceof Boolean ? "${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/gatk4/estimatelibrarycomplexity/main.nf

@@ -0,0 +1,54 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName; getProcessName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process GATK4_ESTIMATELIBRARYCOMPLEXITY {
+    tag "$meta.id"
+    label 'process_medium'
+    publishDir "${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::gatk4=4.2.2.0" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/gatk4:4.2.2.0--hdfd78af_1"
+    } else {
+        container "quay.io/biocontainers/gatk4:4.2.2.0--hdfd78af_1"
+    }
+
+    input:
+    tuple val(meta), path(cram)
+    path(fasta)
+    path(fai)
+    path(dict)
+
+    output:
+    tuple val(meta), path('*.metrics'), emit: metrics
+    path "versions.yml"                  , emit: versions
+
+    script:
+    def prefix = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    def crams = cram.collect(){ x -> "-I ".concat(x.toString()) }.join(" ")
+
+    def avail_mem = 3
+    if (!task.memory) {
+        log.info '[GATK EstimateLibraryComplexity] Available memory not known - defaulting to 3GB. Specify process memory requirements to change this.'
+    } else {
+        avail_mem = task.memory.giga
+    }
+    """
+    gatk EstimateLibraryComplexity \
+        ${crams} \
+        -O ${prefix}.metrics \
+        --REFERENCE_SEQUENCE ${fasta} \
+        --VALIDATION_STRINGENCY SILENT \
+        --TMP_DIR . $options.args
+
+    cat <<-END_VERSIONS > versions.yml
+    ${getProcessName(task.process)}:
+        ${getSoftwareName(task.process)}: \$(echo \$(gatk --version 2>&1) | sed 's/^.*(GATK) v//; s/ .*\$//')
+    END_VERSIONS
+    """
+}

modules/gatk4/estimatelibrarycomplexity/meta.yml

@@ -0,0 +1,56 @@
+name: gatk4_estimatelibrarycomplexity
+description: Estimates the numbers of unique molecules in a sequencing library.
+keywords:
+  - gatk4
+  - gatk4_estimatelibrarycomplexity
+  - duplication_metrics
+  - reporting
+tools:
+  - gatk4:
+      description: Genome Analysis Toolkit (GATK4)
+      homepage: https://gatk.broadinstitute.org/hc/en-us
+      documentation: https://gatk.broadinstitute.org/hc/en-us
+      tool_dev_url: https://github.com/broadinstitute/gatk
+      doi: "10.1158/1538-7445.AM2017-3590"
+      licence: ['Apache-2.0']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - cram:
+      type: file
+      description: BAM/CRAM/SAM file
+      pattern: "*.{bam,cram,sam}"
+  - fasta:
+      type: file
+      description: The reference fasta file
+      pattern: "*.fasta"
+  - fai:
+      type: file
+      description: Index of reference fasta file
+      pattern: "fasta.fai"
+  - dict:
+      type: file
+      description: GATK sequence dictionary
+      pattern: "*.dict"
+
+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"
+  - metrics:
+      type: file
+      description: File containing metrics on the input files
+      pattern: "*.{metrics}"
+
+authors:
+  - "@FriederikeHanssen"

tests/config/pytest_modules.yml

@@ -430,6 +430,10 @@ gatk4/createsomaticpanelofnormals:
   - modules/gatk4/createsomaticpanelofnormals/**
   - tests/modules/gatk4/createsomaticpanelofnormals/**
 
+gatk4/estimatelibrarycomplexity:
+  - modules/gatk4/estimatelibrarycomplexity/**
+  - tests/modules/gatk4/estimatelibrarycomplexity/**
+
 gatk4/fastqtosam:
   - modules/gatk4/fastqtosam/**
   - tests/modules/gatk4/fastqtosam/**

tests/modules/gatk4/estimatelibrarycomplexity/main.nf

@@ -0,0 +1,18 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { GATK4_ESTIMATELIBRARYCOMPLEXITY } from '../../../../modules/gatk4/estimatelibrarycomplexity/main.nf' addParams( options: [:] )
+
+workflow test_gatk4_estimatelibrarycomplexity {
+
+    input = [   [ id:'test', single_end:false ], // meta map
+                file(params.test_data['homo_sapiens']['illumina']['test_paired_end_recalibrated_sorted_bam'], checkIfExists: true)
+            ]
+
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+    fai   = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+    dict  = file(params.test_data['homo_sapiens']['genome']['genome_dict'], checkIfExists: true)
+
+    GATK4_ESTIMATELIBRARYCOMPLEXITY ( input, fasta, fai, dict )
+}

tests/modules/gatk4/estimatelibrarycomplexity/test.yml

@@ -0,0 +1,7 @@
+- name: gatk4 estimatelibrarycomplexity test_gatk4_estimatelibrarycomplexity
+  command: nextflow run tests/modules/gatk4/estimatelibrarycomplexity -entry test_gatk4_estimatelibrarycomplexity -c tests/config/nextflow.config
+  tags:
+    - gatk4/estimatelibrarycomplexity
+    - gatk4
+  files:
+    - path: output/gatk4/test.metrics