Trimmomatic (#1757)

* Trimmomatic main only first draft

* Add test files

* SE PE Adjustment

* Remove extra reads input

* chore: Remove TODOs

* Apply suggestions from code review

Co-authored-by: Edmund Miller <edmund.a.miller@gmail.com>

* fix(trimmomatic): Handle SE output correctly

Since there's never going to be unpaired reads for SE reads we can get
away with it for SE

* fix(trimmomatic): Use correct elvis operator to handle logic

* fix(trimmomatic): Add hack to work with SE and PE reads

* Update test.yml

* use the PE and SE trimming correctly

* Made user set adaptors

* Add documentation

* test(trimmomatic): Add files to pytest_modules

* test(trimmomatic): Update name of failing test

Co-authored-by: Edmund Miller <edmund.a.miller@protonmail.com>
Co-authored-by: Edmund Miller <edmund.a.miller@gmail.com>

modules/trimmomatic/main.nf

@@ -0,0 +1,46 @@
+process TRIMMOMATIC {
+    tag "$meta.id"
+    label 'process_medium'
+
+    conda (params.enable_conda ? "bioconda::trimmomatic=0.39" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/trimmomatic:0.39--hdfd78af_2':
+        'quay.io/biocontainers/trimmomatic:0.39--hdfd78af_2' }"
+
+    input:
+    tuple val(meta), path(reads)
+
+    output:
+    tuple val(meta), path("*.paired.trim*.fastq.gz")   , emit: trimmed_reads
+    tuple val(meta), path("*.unpaired.trim_*.fastq.gz"), optional:true, emit: unpaired_reads
+    tuple val(meta), path("*.log")                     , emit: log
+    path "versions.yml"                                , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    def trimmed = meta.single_end ? "SE" : "PE"
+    def output = meta.single_end ?
+        "${prefix}.SE.paired.trim.fastq.gz" // HACK to avoid unpaired and paired in the trimmed_reads output
+        : "${prefix}.paired.trim_1.fastq.gz ${prefix}.unpaired.trim_1.fastq.gz ${prefix}.paired.trim_2.fastq.gz ${prefix}.unpaired.trim_2.fastq.gz"
+    // TODO Give better error output
+    def qual_trim = task.ext.args2 ?: ''
+    """
+    trimmomatic \\
+        $trimmed \\
+        -threads $task.cpus \\
+        -trimlog ${prefix}.log \\
+        $reads \\
+        $output \\
+        $qual_trim \\
+        $args
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        trimmomatic: \$(trimmomatic -version)
+    END_VERSIONS
+    """
+}

modules/trimmomatic/meta.yml

@@ -0,0 +1,53 @@
+name: "trimmomatic"
+description: Performs quality and adapter trimming on paired end and single end reads
+keywords:
+  - trimming
+  - adapter trimming
+  - quality trimming
+
+tools:
+  - "trimmomatic":
+      description: "A flexible read trimming tool for Illumina NGS data"
+      homepage: "http://www.usadellab.org/cms/?page=trimmomatic"
+      documentation: "https://github.com/usadellab/Trimmomatic"
+      tool_dev_url: "None"
+      doi: "10.1093/bioinformatics/btu170"
+      licence: "['GPL v3']"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - reads:
+      type: file
+      description: |
+        Input FastQ files of size 1 or 2 for single-end and paired-end data, respectively.
+      pattern: "*.fastq.gz"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - trimmed_reads:
+      type: file
+      description: The trimmed/modified paired end fastq reads
+      pattern: "*.paired.trim*.fastq.gz"
+  - unpaired_reads:
+      type: file
+      description: The trimmed/modified unpaired end fastq reads
+      pattern: "*.unpaired.trim_*.fastq.gz"
+  - log:
+      type: file
+      description: trimmomatic log file
+      pattern: "*.log"
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+
+authors:
+  - "@alyssa-ab"

tests/config/pytest_modules.yml

@@ -2079,6 +2079,10 @@ trimgalore:
   - modules/trimgalore/**
   - tests/modules/trimgalore/**
 
+trimmomatic:
+  - modules/trimmomatic/**
+  - tests/modules/trimmomatic/**
+
 ucsc/bed12tobigbed:
   - modules/ucsc/bed12tobigbed/**
   - tests/modules/ucsc/bed12tobigbed/**

tests/modules/trimmomatic/main.nf

@@ -0,0 +1,42 @@
+nextflow.enable.dsl = 2
+
+include {
+    TRIMMOMATIC as TRIMMOMATIC_SE
+    TRIMMOMATIC as TRIMMOMATIC_PE
+    TRIMMOMATIC
+} from '../../../modules/trimmomatic/main.nf'
+
+//
+// Test with single-end data
+//
+workflow test_trimmomatic_single_end {
+    input = [ [ id:'test', single_end:true ], // meta map
+              [ file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true) ]
+            ]
+
+    TRIMMOMATIC_SE ( input )
+}
+
+//
+// Test with paired-end data
+//
+workflow test_trimmomatic_paired_end {
+    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) ]
+            ]
+
+    TRIMMOMATIC_PE ( input )
+}
+
+//
+// Failing test with no adaptor
+//
+workflow test_trimmomatic_no_adaptor {
+    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) ]
+            ]
+
+    TRIMMOMATIC ( input )
+}

tests/modules/trimmomatic/nextflow.config

@@ -0,0 +1,12 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+    withName: TRIMMOMATIC_SE {
+        ext.args2 = 'ILLUMINACLIP:TruSeq3-SE:2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:36'
+    }
+
+    withName: TRIMMOMATIC_PE {
+        ext.args2 = 'ILLUMINACLIP:TruSeq3-PE.fa:2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:36'
+    }
+}

tests/modules/trimmomatic/test.yml

@@ -0,0 +1,26 @@
+- name: trimmomatic single-end
+  command: nextflow run ./tests/modules/trimmomatic -entry test_trimmomatic_single_end -c ./tests/config/nextflow.config -c ./tests/modules/trimmomatic/nextflow.config
+  tags:
+    - "trimmomatic"
+  files:
+    - path: "output/trimmomatic/test.SE.paired.trim.fastq.gz"
+    - path: "output/trimmomatic/test.log"
+      md5sum: e4c3f619e9b0e26847f8f3e3d9af319b
+
+- name: trimmomatic paired-end
+  command: nextflow run ./tests/modules/trimmomatic -entry test_trimmomatic_paired_end -c ./tests/config/nextflow.config -c ./tests/modules/trimmomatic/nextflow.config
+  tags:
+    - "trimmomatic"
+  files:
+    - path: "output/trimmomatic/test.log"
+      md5sum: 9629761761a34576b3484bf4174f681f
+    - path: "output/trimmomatic/test.paired.trim_1.fastq.gz"
+    - path: "output/trimmomatic/test.unpaired.trim_1.fastq.gz"
+    - path: "output/trimmomatic/test.paired.trim_2.fastq.gz"
+    - path: "output/trimmomatic/test.unpaired.trim_2.fastq.gz"
+
+- name: trimmomatic no adapter specified
+  command: nextflow run ./tests/modules/trimmomatic -entry test_trimmomatic_no_adaptor -c ./tests/config/nextflow.config -c ./tests/modules/trimmomatic/nextflow.config
+  tags:
+    - "trimmomatic"
+  exit_code: 1