Add leehom module (#1052)

* Specify more guidelines on input channels * Linting * Updates based on code review * Update README.md * Fix broken sentence * feat: add megahit module, currently decompressed output * Update main.nf * Update tests/modules/megahit/test.yml Co-authored-by: Maxime Borry <maxibor@users.noreply.github.com> * Apply suggestions from code review Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com> * feat: compress all outputs, remove md5sums due to gz stochasicity * fix: wrong conda channel for pigz * fix: broken singleend tests and update meta.yml * Missed one * Apply suggestions from code review Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com> * fix: pigz formatting * Apply suggestions from code review Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com> * Apply suggestions from code review * Add leeHom module * Update modules/leehom/main.nf Co-authored-by: FriederikeHanssen <Friederike.hanssen@qbic.uni-tuebingen.de> * Update modules/leehom/main.nf Co-authored-by: FriederikeHanssen <Friederike.hanssen@qbic.uni-tuebingen.de> * Update modules/leehom/main.nf Co-authored-by: FriederikeHanssen <Friederike.hanssen@qbic.uni-tuebingen.de> * Update modules/leehom/main.nf Co-authored-by: FriederikeHanssen <Friederike.hanssen@qbic.uni-tuebingen.de> Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com> Co-authored-by: Maxime Borry <maxibor@users.noreply.github.com> Co-authored-by: FriederikeHanssen <Friederike.hanssen@qbic.uni-tuebingen.de>
2024-11-10 20:23:10 +00:00 · 2021-11-15 15:26:06 +01:00 · 2021-11-15 15:26:06 +01:00 · 2c3c87a10f
commit 2c3c87a10f
parent b012b349c1
6 changed files with 331 additions and 9 deletions
--- a/modules/leehom/functions.nf
+++ b/modules/leehom/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)}/$args.filename"
+    }
+}
--- a/modules/leehom/main.nf
+++ b/modules/leehom/main.nf
@ -0,0 +1,85 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName; getProcessName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+def VERSION="1.2.15"
+
+process LEEHOM {
+    tag "$meta.id"
+    label 'process_low'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), meta:meta, publish_by_meta:['id']) }
+
+    conda (params.enable_conda ? "bioconda::leehom=1.2.15" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/leehom:1.2.15--h29e30f7_1"
+    } else {
+        container "quay.io/biocontainers/leehom:1.2.15--h29e30f7_1"
+    }
+
+    input:
+    tuple val(meta), path(reads)
+
+    output:
+    tuple val(meta), path("${prefix}.bam")          , optional: true, emit: bam
+    tuple val(meta), path("${prefix}.fq.gz")        , optional: true, emit: fq_pass
+    tuple val(meta), path("${prefix}.fail.fq.gz")   , optional: true, emit: fq_fail
+    tuple val(meta), path("${prefix}_r1.fq.gz")     , optional: true, emit: unmerged_r1_fq_pass
+    tuple val(meta), path("${prefix}_r1.fail.fq.gz"), optional: true, emit: unmerged_r1_fq_fail
+    tuple val(meta), path("${prefix}_r2.fq.gz")     , optional: true, emit: unmerged_r2_fq_pass
+    tuple val(meta), path("${prefix}_r2.fail.fq.gz"), optional: true, emit: unmerged_r2_fq_fail
+    tuple val(meta), path("*.log")                                  , emit: log
+
+    path "versions.yml"                                             , emit: versions
+
+    script:
+    prefix = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+
+    if ( reads.toString().endsWith('.bam') ) {
+            """
+            leeHom \\
+                $options.args \\
+                -t $task.cpus \\
+                -o ${prefix}.bam \\
+                --log ${prefix}.log \\
+                $reads
+
+            cat <<-END_VERSIONS > versions.yml
+            ${getProcessName(task.process)}:
+                ${getSoftwareName(task.process)}: \$( echo $VERSION )
+            END_VERSIONS
+            """
+    } else if ( meta.single_end ) {
+            """
+            leeHom \\
+                $options.args \\
+                -t $task.cpus \\
+                -fq1 $reads \\
+                -fqo ${prefix} \\
+                --log ${prefix}.log
+
+            cat <<-END_VERSIONS > versions.yml
+            ${getProcessName(task.process)}:
+                ${getSoftwareName(task.process)}: \$( echo $VERSION )
+            END_VERSIONS
+            """
+    } else {
+            """
+            leeHom \\
+                $options.args \\
+                -t $task.cpus \\
+                -fq1 ${reads[0]} \\
+                -fq2 ${reads[1]} \\
+                -fqo ${prefix} \\
+                --log ${prefix}.log
+
+            cat <<-END_VERSIONS > versions.yml
+            ${getProcessName(task.process)}:
+                ${getSoftwareName(task.process)}: \$( echo $VERSION )
+            END_VERSIONS
+            """
+    }
+}
--- a/modules/leehom/meta.yml
+++ b/modules/leehom/meta.yml
@ -0,0 +1,77 @@
+name: leehom
+description: Bayesian reconstruction of ancient DNA fragments
+keywords:
+  - ancient DNA
+  - adapter removal
+  - clipping
+  - trimming
+  - merging
+  - collapsing
+  - preprocessing
+  - bayesian
+tools:
+  - leehom:
+      description: Bayesian reconstruction of ancient DNA fragments
+      homepage: "https://grenaud.github.io/leeHom/"
+      documentation: "https://github.com/grenaud/leeHom"
+      tool_dev_url: "https://github.com/grenaud/leeHom"
+      doi: "10.1093/nar/gku699"
+      licence: ['GPL v3']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - reads:
+      type: file
+      description: Unaligned BAM or one or two gzipped FASTQ file(s)
+      pattern: "*.{bam,fq.gz,fastq.gz}"
+
+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"
+  - bam:
+      type: file
+      description: BAM file
+      pattern: "*.bam"
+  - fq_pass:
+      type: file
+      description: Trimmed and merged FASTQ
+      pattern: "*.fq.gz"
+  - fq_fail:
+      type: file
+      description: Failed trimmed and merged FASTQs
+      pattern: "*.fail.fq.gz"
+  - unmerged_r1_fq_pass:
+      type: file
+      description: Passed unmerged R1 FASTQs
+      pattern: "*.r1.fq.gz"
+  - unmerged_r1_fq_fail:
+      type: file
+      description: Failed unmerged R1 FASTQs
+      pattern: "*.r1.fail.fq.gz"
+  - unmerged_r2_fq_pass:
+      type: file
+      description: Passed unmerged R1 FASTQs
+      pattern: "*.r2.fq.gz"
+  - unmerged_r2_fq_pass:
+      type: file
+      description: Failed unmerged R1 FASTQs
+      pattern: "*.r2.fail.fq.gz"
+  - log:
+      type: file
+      description: Log file of command
+      pattern: "*.log"
+
+
+authors:
+  - "@jfy133"
--- a/tests/config/pytest_modules.yml
+++ b/tests/config/pytest_modules.yml
@ -621,14 +621,14 @@ homer/makeucscfile:
  - modules/homer/makeucscfile/**
  - tests/modules/homer/makeucscfile/**

-imputeme/vcftoprs:
-  - modules/imputeme/vcftoprs/**
-  - tests/modules/imputeme/vcftoprs/**
-
 idr:
  - modules/idr/**
  - tests/modules/idr/**

+imputeme/vcftoprs:
+  - modules/imputeme/vcftoprs/**
+  - tests/modules/imputeme/vcftoprs/**
+
 iqtree:
  - modules/iqtree/**
  - tests/modules/iqtree/**
@ -718,6 +718,10 @@ last/train:
  - modules/last/train/**
  - tests/modules/last/train/**

+leehom:
+  - modules/leehom/**
+  - tests/modules/leehom/**
+
 lima:
  - modules/lima/**
  - tests/modules/lima/**
@ -786,6 +790,10 @@ megahit:
  - modules/megahit/**
  - tests/modules/megahit/**

+meningotype:
+  - modules/meningotype/**
+  - tests/modules/meningotype/**
+
 metabat2/jgisummarizebamcontigdepths:
  - modules/metabat2/jgisummarizebamcontigdepths/**
  - tests/modules/metabat2/jgisummarizebamcontigdepths/**
@ -794,11 +802,6 @@ metabat2/metabat2:
  - modules/metabat2/metabat2/**
  - tests/modules/metabat2/metabat2/**

-meningotype:
-  - modules/meningotype/**
-  - tests/modules/meningotype/**
-
-
 metaphlan3:
  - modules/metaphlan3/**
  - tests/modules/metaphlan3/**
--- a/tests/modules/leehom/main.nf
+++ b/tests/modules/leehom/main.nf
@ -0,0 +1,36 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { LEEHOM } from '../../../modules/leehom/main.nf' addParams( options: [:] )
+include { SAMTOOLS_VIEW } from '../../../modules/samtools/view/main.nf'  addParams( options: [args: "-f4 -b"] )
+
+workflow test_leehom_bam {
+
+    input = [ [ id:'test', single_end:false ], // meta map
+              file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_bam'], checkIfExists: true) ]
+
+    fasta = []
+
+    SAMTOOLS_VIEW ( input, fasta )
+    LEEHOM ( SAMTOOLS_VIEW.out.bam )
+}
+
+workflow test_leehom_se_fq {
+
+    input = [ [ id:'test', single_end:true ], // meta map
+              file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true) ]
+
+    LEEHOM ( input )
+}
+
+workflow test_leehom_pe_fq {
+
+    input = [ [ id:'test', single_end:false ], // meta map
+              [
+                  file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true),
+                  file(params.test_data['sarscov2']['illumina']['test_2_fastq_gz'], checkIfExists: true)
+              ] ]
+
+    LEEHOM ( input )
+}
--- a/tests/modules/leehom/test.yml
+++ b/tests/modules/leehom/test.yml
@ -0,0 +1,43 @@
+- name: leehom test_leehom_bam
+  command: nextflow run tests/modules/leehom -entry test_leehom_bam -c tests/config/nextflow.config
+  tags:
+    - leehom
+  files:
+    - path: output/leehom/test.bam
+      md5sum: 19a1bf95714523868791f1d4d3aaee73
+    - path: output/leehom/test.log
+      md5sum: d1f5da273eb69f41babda510797c7671
+    - path: output/samtools/test.bam
+      md5sum: 25d13b3b31b147bb3836dea9932c38dd
+
+- name: leehom test_leehom_se_fq
+  command: nextflow run tests/modules/leehom -entry test_leehom_se_fq -c tests/config/nextflow.config
+  tags:
+    - leehom
+  files:
+    - path: output/leehom/test.fail.fq.gz
+      md5sum: 7029066c27ac6f5ef18d660d5741979a
+    - path: output/leehom/test.fq.gz
+      md5sum: ed10c4bbf5c3082ca68823535b91e1e2
+    - path: output/leehom/test.log
+      md5sum: 59aa280cb72dfbea05ba913cb89db143
+
+- name: leehom test_leehom_pe_fq
+  command: nextflow run tests/modules/leehom -entry test_leehom_pe_fq -c tests/config/nextflow.config
+  tags:
+    - leehom
+  files:
+    - path: output/leehom/test.fail.fq.gz
+      md5sum: 7029066c27ac6f5ef18d660d5741979a
+    - path: output/leehom/test.fq.gz
+      md5sum: 84929b78e3f89371ecd3b4c915b9ec33
+    - path: output/leehom/test.log
+      md5sum: 800b5a88dc0822886bfbb271029e2a4a
+    - path: output/leehom/test_r1.fail.fq.gz
+      md5sum: 7029066c27ac6f5ef18d660d5741979a
+    - path: output/leehom/test_r1.fq.gz
+      md5sum: e9258420fa712e8536106995a7d1d97a
+    - path: output/leehom/test_r2.fail.fq.gz
+      md5sum: 7029066c27ac6f5ef18d660d5741979a
+    - path: output/leehom/test_r2.fq.gz
+      md5sum: 27230bcc5eae81ec5c1701798d39c1af