Add bcftools reheader (#585) (#608)

* local tests and linting passing (#585)

* fix: picard filtersamreads input (#610)

* Move readlist into same input channel as bam

* Update test reflecting input restructuring

* Update tests/modules/picard/filtersamreads/main.nf

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* fix test

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Added module arriba (#611)

* Updated the version of STAR in align and genomegenerate modules

* Changes in test.yml

* Changes in test.yml

* Added module arriba

* Changes in test configs

* Added module Arriba for fusion detection

* Fixed review comments

* Added an output option for discarded fusions

* Resolved some conflits

* conflicts

* Apply suggestions from code review

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* added test for new header

* enhance module fastp: add `save_merged` (#598) (#614)

* enhance module fastp: add `save_merged` (#598)

* removed md5sum checks from log and json

* Apply suggestions from code review

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>
Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>
Co-authored-by: praveenraj2018 <43108054+praveenraj2018@users.noreply.github.com>

modules/bcftools/reheader/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"
+        }
+    }
+}

modules/bcftools/reheader/main.nf

@@ -0,0 +1,47 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process BCFTOOLS_REHEADER {
+    tag "$meta.id"
+    label 'process_low'
+    publishDir "${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::bcftools=1.13" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/bcftools:1.13--h3a49de5_0"
+    } else {
+        container "quay.io/biocontainers/bcftools:1.13--h3a49de5_0"
+    }
+
+    input:
+    tuple val(meta), path(vcf)
+    path fai
+    path header
+
+    output:
+    tuple val(meta), path("*.vcf.gz"), emit: vcf
+    path "*.version.txt"             , emit: version
+
+    script:
+    def software         = getSoftwareName(task.process)
+    def prefix           = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    def update_sequences = fai ? "-f $fai" : ""
+    def new_header       = header ? "-h $header" : ""
+    """
+    bcftools \\
+        reheader \\
+        $update_sequences \\
+        $new_header \\
+        $options.args \\
+        --threads $task.cpus \\
+        -o ${prefix}.vcf.gz \\
+        $vcf
+
+    echo \$(bcftools --version 2>&1) | sed 's/^.*bcftools //; s/ .*\$//' > ${software}.version.txt
+    """
+}

modules/bcftools/reheader/meta.yml

@@ -0,0 +1,51 @@
+name: bcftools_reheader
+description: Reheader a VCF file
+keywords:
+  - reheader
+  - vcf
+  - update header
+tools:
+  - reheader:
+      description: |
+        Modify header of VCF/BCF files, change sample names.
+      homepage: http://samtools.github.io/bcftools/bcftools.html
+      documentation: http://samtools.github.io/bcftools/bcftools.html#reheader
+      doi: 10.1093/gigascience/giab008
+      licence: ['GPL']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - vcf:
+      type: file
+      description: VCF/BCF file
+      pattern: "*.{vcf.gz,vcf,bcf}"
+  - fai:
+      type: file
+      description: Fasta index to update header sequences with
+      pattern: "*.{fai}"
+  - header:
+      type: file
+      description: New header to add to the VCF
+      pattern: "*.{header.txt}"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+  - vcf:
+      type: file
+      description: VCF with updated header
+      pattern: "*.{vcf.gz}"
+
+authors:
+  - "@bjohnnyd"

tests/config/pytest_modules.yml

@@ -50,6 +50,10 @@ bcftools/mpileup:
   - modules/bcftools/mpileup/**
   - tests/modules/bcftools/mpileup/**
 
+bcftools/reheader:
+  - modules/bcftools/reheader/**
+  - tests/modules/bcftools/reheader/**
+
 bcftools/stats:
   - modules/bcftools/stats/**
   - tests/modules/bcftools/stats/**

tests/modules/bcftools/reheader/main.nf

@@ -0,0 +1,40 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { BCFTOOLS_REHEADER } from '../../../../modules/bcftools/reheader/main.nf' addParams( options: [suffix: '.updated'] )
+
+workflow test_bcftools_reheader_update_sequences {
+
+    input = [
+        [ id:'test', single_end:false ],
+        file(params.test_data['sarscov2']['illumina']['test_vcf_gz'], checkIfExists: true)
+    ]
+    fai    = file(params.test_data['sarscov2']['genome']['genome_fasta_fai'], checkIfExists: true)
+    header = []
+    BCFTOOLS_REHEADER ( input, fai, header )
+}
+
+workflow test_bcftools_reheader_new_header {
+
+    input = [
+        [ id:'test', single_end:false ],
+        file(params.test_data['sarscov2']['illumina']['test_vcf_gz'], checkIfExists: true)
+    ]
+    fai    = []
+    header = file(params.test_data['sarscov2']['illumina']['test_vcf'], checkIfExists: true)
+
+    BCFTOOLS_REHEADER ( input, fai, header )
+}
+
+workflow test_bcftools_reheader_new_header_update_sequences {
+
+    input = [
+        [ id:'test', single_end:false ],
+        file(params.test_data['sarscov2']['illumina']['test_vcf_gz'], checkIfExists: true)
+    ]
+    fai    = file(params.test_data['sarscov2']['genome']['genome_fasta_fai'], checkIfExists: true)
+    header = file(params.test_data['sarscov2']['illumina']['test_vcf'], checkIfExists: true)
+
+    BCFTOOLS_REHEADER ( input, fai, header )
+}

tests/modules/bcftools/reheader/test.yml

@@ -0,0 +1,26 @@
+- name: bcftools reheader test_bcftools_reheader_update_sequences
+  command: nextflow run tests/modules/bcftools/reheader -entry test_bcftools_reheader_update_sequences -c tests/config/nextflow.config
+  tags:
+    - bcftools/reheader
+    - bcftools
+  files:
+    - path: output/bcftools/test.updated.vcf.gz
+      md5sum: 9e29f28038bfce77ee00022627209ed6
+
+- name: bcftools reheader test_bcftools_reheader_new_header
+  command: nextflow run tests/modules/bcftools/reheader -entry test_bcftools_reheader_new_header -c tests/config/nextflow.config
+  tags:
+    - bcftools/reheader
+    - bcftools
+  files:
+    - path: output/bcftools/test.updated.vcf.gz
+      md5sum: f7f536d889bbf5be40243252c394ee1f
+
+- name: bcftools reheader test_bcftools_reheader_new_header_update_sequences
+  command: nextflow run tests/modules/bcftools/reheader -entry test_bcftools_reheader_new_header_update_sequences -c tests/config/nextflow.config
+  tags:
+    - bcftools/reheader
+    - bcftools
+  files:
+    - path: output/bcftools/test.updated.vcf.gz
+      md5sum: 9e29f28038bfce77ee00022627209ed6