From bd68797ffb0997f3c62d5964d78c3bd5cd7e4736 Mon Sep 17 00:00:00 2001
From: GCJMackenzie <43276267+GCJMackenzie@users.noreply.github.com>
Date: Mon, 13 Sep 2021 16:16:23 +0100
Subject: [PATCH] Mutect2 (#682)

* mutect2 files added, first draft of module code entered

* removing comment from main.nf

* removing comment from main.nf

* test added, commit made before editing yaml

* tests added, still needs final check and info/comments added

* gatk4 version changed to gatk4=4.2.0.0

* multiple sample support added, information added to module yaml file

* Update meta.yml

fixed linting error

* add keywords to meta.yml

* Corrections made to meta.yml

* removed whitespace from meta.yml

Co-authored-by: Gavin.Mackenzie <gavin.mackenzie@nibsc.org>
Co-authored-by: Maxime U. Garcia <max.u.garcia@gmail.com>
---
 modules/gatk4/mutect2/functions.nf   | 68 +++++++++++++++++++++
 modules/gatk4/mutect2/main.nf        | 76 ++++++++++++++++++++++++
 modules/gatk4/mutect2/meta.yml       | 89 ++++++++++++++++++++++++++++
 tests/config/pytest_modules.yml      |  4 ++
 tests/modules/gatk4/mutect2/main.nf  | 62 +++++++++++++++++++
 tests/modules/gatk4/mutect2/test.yml | 27 +++++++++
 6 files changed, 326 insertions(+)
 create mode 100644 modules/gatk4/mutect2/functions.nf
 create mode 100644 modules/gatk4/mutect2/main.nf
 create mode 100644 modules/gatk4/mutect2/meta.yml
 create mode 100644 tests/modules/gatk4/mutect2/main.nf
 create mode 100644 tests/modules/gatk4/mutect2/test.yml

diff --git a/modules/gatk4/mutect2/functions.nf b/modules/gatk4/mutect2/functions.nf
new file mode 100644
index 00000000..da9da093
--- /dev/null
+++ b/modules/gatk4/mutect2/functions.nf
@@ -0,0 +1,68 @@
+//
+//  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)}/$args.filename"
+        }
+    }
+}
diff --git a/modules/gatk4/mutect2/main.nf b/modules/gatk4/mutect2/main.nf
new file mode 100644
index 00000000..2fcdbbd7
--- /dev/null
+++ b/modules/gatk4/mutect2/main.nf
@@ -0,0 +1,76 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process GATK4_MUTECT2 {
+    tag "$meta.id"
+    label 'process_medium'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), meta:meta, publish_by_meta:['id']) }
+
+    conda (params.enable_conda ? "bioconda::gatk4=4.2.0.0" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/gatk4:4.2.0.0--0"
+    } else {
+        container "quay.io/biocontainers/gatk4:4.2.0.0--0"
+    }
+
+    input:
+    tuple val(meta) , path(bam) , path(bai) , val(which_norm)
+    val run_single
+    val run_pon
+    path fasta
+    path fastaidx
+    path dict
+    path germline_resource
+    path germline_resource_idx
+    path panel_of_normals
+    path panel_of_normals_idx
+
+    output:
+    tuple val(meta), path("*.vcf.gz"), emit: vcf
+    tuple val(meta), path("*.tbi")   , emit: tbi
+    tuple val(meta), path("*.f1r2.tar.gz"), optional:true, emit: f1r2
+    path "*.version.txt"          , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    def inputsList = []
+    def normalsList = []
+    def inputsCommand = ''
+    def panelsCommand = ''
+    def normalsCommand = ''
+
+    bam.each() {a -> inputsList.add(" -I " + a ) }
+    inputsCommand = inputsList.join( ' ')
+
+    if(run_pon) {
+        panelsCommand = ''
+        normalsCommand = ''
+
+    } else if(run_single) {
+        panelsCommand = " --germline-resource $germline_resource --panel-of-normals $panel_of_normals"
+        normalsCommand = ''
+
+    } else {
+        panelsCommand = " --germline-resource $germline_resource --panel-of-normals $panel_of_normals --f1r2-tar-gz ${prefix}.f1r2.tar.gz"
+        which_norm.each() {a -> normalsList.add(" -normal " + a ) }
+        normalsCommand = normalsList.join( ' ')
+    }
+
+    """
+    gatk Mutect2 \\
+        -R ${fasta} \\
+        ${inputsCommand} \\
+        ${normalsCommand} \\
+        ${panelsCommand} \\
+        -O ${prefix}.vcf.gz \\
+        $options.args
+
+    echo \$(gatk --version 2>&1) | sed 's/^.*(GATK) v//; s/ .*\$//' > ${software}.version.txt
+    """
+}
diff --git a/modules/gatk4/mutect2/meta.yml b/modules/gatk4/mutect2/meta.yml
new file mode 100644
index 00000000..8ba803df
--- /dev/null
+++ b/modules/gatk4/mutect2/meta.yml
@@ -0,0 +1,89 @@
+name: gatk4_mutect2
+description:  Call somatic SNVs and indels via local assembly of haplotypes.
+keywords:
+  - gatk4
+  - mutect2
+  - haplotype
+  - somatic
+tools:
+  - gatk4:
+      description: |
+        Developed in the Data Sciences Platform at the Broad Institute, the toolkit offers a wide variety of tools
+        with a primary focus on variant discovery and genotyping. Its powerful processing engine
+        and high-performance computing features make it capable of taking on projects of any size.
+      homepage: https://gatk.broadinstitute.org/hc/en-us
+      documentation: https://gatk.broadinstitute.org/hc/en-us/categories/360002369672s
+      doi: 10.1158/1538-7445.AM2017-3590
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test']
+  - bam:
+      type: list
+      description: list of BAM files
+      pattern: "*.bam"
+  - bai:
+      type: list
+      description: list of BAM file indexes
+      pattern: "*.bam.bai"
+  - which_norm:
+      type: list
+      description: optional list of sample headers contained in the normal sample bam files (these are required for tumor_normal_pair mode)
+  - run_single:
+      type: boolean
+      description: Specify whether or not to run in tumor_single mode instead of tumor_normal_pair mode (will be ignored if run_pon is also true)
+  - run_pon:
+      type: boolean
+      description: Specify whether or not to run in panel_of_normal mode instead of tumor_normal_pair mode
+  - fasta:
+      type: file
+      description: The reference fasta file
+      pattern: "*.fasta"
+  - fastaidx:
+      type: file
+      description: Index of reference fasta file
+      pattern: "fasta.fai"
+  - dict:
+      type: file
+      description: GATK sequence dictionary
+      pattern: "*.dict"
+  - germline_resource:
+      type: file
+      description: Population vcf of germline sequencing, containing allele fractions.
+      pattern: "*.vcf.gz"
+  - germline_resource_idx:
+      type: file
+      description: Index file for the germline resource.
+      pattern: "*.vcf.gz_tbi"
+  - panel_of_normals:
+      type: file
+      description: vcf file to be used as a panel of normals.
+      pattern: "*.vcf.gz"
+  - panel_of_normals_idx:
+      type: file
+      description: Index for the panel of normals.
+      pattern: "*.vcf.gz_tbi"
+
+output:
+  - vcf:
+      type: file
+      description: compressed vcf file
+      pattern: "*.vcf.gz"
+  - tbi:
+      type: file
+      description: Index of vcf file
+      pattern: "*vcf.gz.tbi"
+  - f1r2:
+      type: file
+      description: file containing information to be passed to LearnReadOrientationModel (only outputted when tumor_normal_pair mode is run)
+      pattern: "*.f1r2.tar.gz"
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.version.txt"
+
+authors:
+  - "@GCJMackenzie"
diff --git a/tests/config/pytest_modules.yml b/tests/config/pytest_modules.yml
index 6c7993a3..a10bccf1 100644
--- a/tests/config/pytest_modules.yml
+++ b/tests/config/pytest_modules.yml
@@ -358,6 +358,10 @@ gatk4/mergevcfs:
   - modules/gatk4/mergevcfs/**
   - tests/modules/gatk4/mergevcfs/**
 
+gatk4/mutect2:
+  - modules/gatk4/mutect2/**
+  - tests/modules/gatk4/mutect2/**
+
 gatk4/revertsam:
   - modules/gatk4/revertsam/**
   - tests/modules/gatk4/revertsam/**
diff --git a/tests/modules/gatk4/mutect2/main.nf b/tests/modules/gatk4/mutect2/main.nf
new file mode 100644
index 00000000..072b3125
--- /dev/null
+++ b/tests/modules/gatk4/mutect2/main.nf
@@ -0,0 +1,62 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { GATK4_MUTECT2 } from '../../../../modules/gatk4/mutect2/main.nf' addParams( options: [:] )
+
+workflow test_gatk4_mutect2_tumor_normal_pair {
+    input = [ [ id:'test'], // meta map
+              [ file(params.test_data['homo_sapiens']['illumina']['test_paired_end_recalibrated_sorted_bam'], checkIfExists: true) , file(params.test_data['homo_sapiens']['illumina']['test2_paired_end_recalibrated_sorted_bam'], checkIfExists: true)],
+              [ file(params.test_data['homo_sapiens']['illumina']['test_paired_end_recalibrated_sorted_bam_bai'], checkIfExists: true) , file(params.test_data['homo_sapiens']['illumina']['test2_paired_end_recalibrated_sorted_bam_bai'], checkIfExists: true)],
+              ["testN"]
+            ]
+    run_single = false
+    run_pon = false
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+    fastaidx = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+    dict = file(params.test_data['homo_sapiens']['genome']['genome_dict'], checkIfExists: true)
+    germline_resource = file(params.test_data['homo_sapiens']['genome']['gnomad_r2_1_1_vcf_gz'], checkIfExists: true)
+    germline_resource_idx = file(params.test_data['homo_sapiens']['genome']['gnomad_r2_1_1_vcf_gz_tbi'], checkIfExists: true)
+    panel_of_normals = file(params.test_data['homo_sapiens']['genome']['mills_and_1000g_indels_vcf_gz'], checkIfExists: true)
+    panel_of_normals_idx = file(params.test_data['homo_sapiens']['genome']['mills_and_1000g_indels_vcf_gz_tbi'], checkIfExists: true)
+
+    GATK4_MUTECT2 ( input , run_single , run_pon , fasta , fastaidx , dict , germline_resource, germline_resource_idx , panel_of_normals , panel_of_normals_idx )
+}
+
+workflow test_gatk4_mutect2_tumor_single {
+    input = [ [ id:'test'], // meta map
+              [ file(params.test_data['homo_sapiens']['illumina']['test2_paired_end_recalibrated_sorted_bam'], checkIfExists: true)],
+              [ file(params.test_data['homo_sapiens']['illumina']['test2_paired_end_recalibrated_sorted_bam_bai'], checkIfExists: true)],
+              []
+            ]
+    run_single = true
+    run_pon = false
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+    fastaidx = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+    dict = file(params.test_data['homo_sapiens']['genome']['genome_dict'], checkIfExists: true)
+    germline_resource = file(params.test_data['homo_sapiens']['genome']['gnomad_r2_1_1_vcf_gz'], checkIfExists: true)
+    germline_resource_idx = file(params.test_data['homo_sapiens']['genome']['gnomad_r2_1_1_vcf_gz_tbi'], checkIfExists: true)
+    panel_of_normals = file(params.test_data['homo_sapiens']['genome']['mills_and_1000g_indels_vcf_gz'], checkIfExists: true)
+    panel_of_normals_idx = file(params.test_data['homo_sapiens']['genome']['mills_and_1000g_indels_vcf_gz_tbi'], checkIfExists: true)
+
+    GATK4_MUTECT2 ( input , run_single , run_pon , fasta , fastaidx , dict , germline_resource, germline_resource_idx , panel_of_normals , panel_of_normals_idx )
+}
+
+workflow test_gatk4_mutect2_generate_pon {
+    input = [ [ id:'test'], // meta map
+              [ file(params.test_data['homo_sapiens']['illumina']['test_paired_end_recalibrated_sorted_bam'], checkIfExists: true)],
+              [ file(params.test_data['homo_sapiens']['illumina']['test_paired_end_recalibrated_sorted_bam_bai'], checkIfExists: true)],
+              []
+            ]
+    run_single = false
+    run_pon = true
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+    fastaidx = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+    dict = file(params.test_data['homo_sapiens']['genome']['genome_dict'], checkIfExists: true)
+    germline_resource = []
+    germline_resource_idx = []
+    panel_of_normals = []
+    panel_of_normals_idx = []
+
+    GATK4_MUTECT2 ( input , run_single , run_pon , fasta , fastaidx , dict , germline_resource, germline_resource_idx , panel_of_normals , panel_of_normals_idx )
+}
diff --git a/tests/modules/gatk4/mutect2/test.yml b/tests/modules/gatk4/mutect2/test.yml
new file mode 100644
index 00000000..03205266
--- /dev/null
+++ b/tests/modules/gatk4/mutect2/test.yml
@@ -0,0 +1,27 @@
+- name: gatk4 mutect2 test_gatk4_mutect2_tumor_normal_pair
+  command: nextflow run tests/modules/gatk4/mutect2 -entry test_gatk4_mutect2_tumor_normal_pair -c tests/config/nextflow.config
+  tags:
+    - gatk4
+    - gatk4/mutect2
+  files:
+    - path: output/gatk4/test.f1r2.tar.gz
+    - path: output/gatk4/test.vcf.gz
+    - path: output/gatk4/test.vcf.gz.tbi
+
+- name: gatk4 mutect2 test_gatk4_mutect2_tumor_single
+  command: nextflow run tests/modules/gatk4/mutect2 -entry test_gatk4_mutect2_tumor_single -c tests/config/nextflow.config
+  tags:
+    - gatk4
+    - gatk4/mutect2
+  files:
+    - path: output/gatk4/test.vcf.gz
+    - path: output/gatk4/test.vcf.gz.tbi
+
+- name: gatk4 mutect2 test_gatk4_mutect2_generate_pon
+  command: nextflow run tests/modules/gatk4/mutect2 -entry test_gatk4_mutect2_generate_pon -c tests/config/nextflow.config
+  tags:
+    - gatk4
+    - gatk4/mutect2
+  files:
+    - path: output/gatk4/test.vcf.gz
+    - path: output/gatk4/test.vcf.gz.tbi