Adapterremoval glob fix (#1391)

* fix: remove left-over unnecessary code

* Partial fix for AR module output declarations

* Remove `def` for prefix so useable in output block

* Fix tests

* Add adapterlist support

* Fix tests after addition of adapter list

modules/adapterremoval/main.nf

@@ -9,48 +9,34 @@ process ADAPTERREMOVAL {
 
     input:
     tuple val(meta), path(reads)
+    path(adapterlist)
 
     output:
-    tuple val(meta), path('*.truncated.gz')         , optional: true, emit: singles_truncated
+    tuple val(meta), path("${prefix}.truncated.gz")            , optional: true, emit: singles_truncated
-    tuple val(meta), path('*.discarded.gz')         , optional: true, emit: discarded
+    tuple val(meta), path("${prefix}.discarded.gz")            , optional: true, emit: discarded
-    tuple val(meta), path('*.pair1.truncated.gz')   , optional: true, emit: pair1_truncated
+    tuple val(meta), path("${prefix}.pair1.truncated.gz")      , optional: true, emit: pair1_truncated
-    tuple val(meta), path('*.pair2.truncated.gz')   , optional: true, emit: pair2_truncated
+    tuple val(meta), path("${prefix}.pair2.truncated.gz")      , optional: true, emit: pair2_truncated
-    tuple val(meta), path('*.collapsed.gz')         , optional: true, emit: collapsed
+    tuple val(meta), path("${prefix}.collapsed.gz")            , optional: true, emit: collapsed
-    tuple val(meta), path('*.collapsed.truncated')  , optional: true, emit: collapsed_truncated
+    tuple val(meta), path("${prefix}.collapsed.truncated.gz")  , optional: true, emit: collapsed_truncated
-    tuple val(meta), path('*paired.gz')             , optional: true, emit: paired_interleaved
+    tuple val(meta), path("${prefix}.paired.gz")               , optional: true, emit: paired_interleaved
-    tuple val(meta), path('*.log')                  , emit: log
+    tuple val(meta), path('*.log')                             , emit: log
-    path "versions.yml"                             , emit: versions
+    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 list = adapterlist ? "--adapter-list ${adapterlist}" : ""
+    prefix = task.ext.prefix ?: "${meta.id}"
 
     if (meta.single_end) {
         """
         AdapterRemoval  \\
             --file1 $reads \\
             $args \\
-            --basename $prefix \\
+            $adapterlist \\
-            --threads ${task.cpus} \\
+            --basename ${prefix} \\
-            --settings ${prefix}.log \\
-            --seed 42 \\
-            --gzip
-
-        cat <<-END_VERSIONS > versions.yml
-        "${task.process}":
-            adapterremoval: \$(AdapterRemoval --version 2>&1 | sed -e "s/AdapterRemoval ver. //g")
-        END_VERSIONS
-        """
-    } else if (!meta.single_end ) {
-        """
-        AdapterRemoval  \\
-            --file1 ${reads[0]} \\
-            --file2 ${reads[1]} \\
-            $args \\
-            --basename $prefix \\
             --threads ${task.cpus} \\
             --settings ${prefix}.log \\
             --seed 42 \\
@@ -67,13 +53,13 @@ process ADAPTERREMOVAL {
             --file1 ${reads[0]} \\
             --file2 ${reads[1]} \\
             $args \\
-            --basename $prefix \\
+            $adapterlist \\
+            --basename ${prefix} \\
             --threads $task.cpus \\
             --settings ${prefix}.log \\
             --seed 42 \\
             --gzip
 
-        cat *.collapsed.gz *.collapsed.truncated.gz > ${prefix}.merged.fastq.gz
         cat <<-END_VERSIONS > versions.yml
         "${task.process}":
             adapterremoval: \$(AdapterRemoval --version 2>&1 | sed -e "s/AdapterRemoval ver. //g")

modules/adapterremoval/meta.yml

@@ -24,6 +24,13 @@ input:
         List of input FastQ files of size 1 and 2 for single-end and paired-end data,
         respectively.
       pattern: "*.{fq,fastq,fq.gz,fastq.gz}"
+  - adapterlist:
+      type: file
+      description:
+        Optional text file containing list of adapters to look for for removal
+        with one adapter per line. Otherwise will look for default adapters (see
+        AdapterRemoval man page), or can be modified to remove user-specified
+        adapters via ext.args.
 
 output:
   - meta:

tests/modules/adapterremoval/main.nf

@@ -2,14 +2,16 @@
 
 nextflow.enable.dsl = 2
 
-include { ADAPTERREMOVAL } from '../../../modules/adapterremoval/main.nf'
+include { ADAPTERREMOVAL                            } from '../../../modules/adapterremoval/main.nf'
+include { ADAPTERREMOVAL as ADAPTERREMOVAL_COLLAPSE } from '../../../modules/adapterremoval/main.nf'
+
 
 workflow test_adapterremoval_single_end {
     input = [ [ id:'test', single_end:true, collapse:false ], // meta map
               file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true)
             ]
 
-    ADAPTERREMOVAL ( input )
+    ADAPTERREMOVAL ( input, [] )
 }
 
 workflow test_adapterremoval_paired_end {
@@ -18,15 +20,15 @@ workflow test_adapterremoval_paired_end {
                 file(params.test_data['sarscov2']['illumina']['test_2_fastq_gz'], checkIfExists: true) ]
             ]
 
-    ADAPTERREMOVAL ( input )
+    ADAPTERREMOVAL ( input, [] )
 }
 
 workflow test_adapterremoval_paired_end_collapse {
-    input = [ [ id:'test', single_end:false, collapse:true ], // meta map
+    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) ]
             ]
 
-    ADAPTERREMOVAL ( input )
+    ADAPTERREMOVAL_COLLAPSE ( input, [] )
 }
 

tests/modules/adapterremoval/nextflow.config

@@ -2,4 +2,8 @@ process {
 
     publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
 
+    withName: ADAPTERREMOVAL_COLLAPSE {
+        ext.args = "--collapse"
+    }
+
 }

tests/modules/adapterremoval/test.yml

@@ -23,7 +23,6 @@
       md5sum: 294a6277f0139bd597e57c6fa31f39c7
     - path: output/adapterremoval/test.pair2.truncated.gz
       md5sum: de7b38e2c881bced8671acb1ab452d78
-    - path: output/adapterremoval/test.singleton.truncated.gz
     - path: output/adapterremoval/versions.yml
       md5sum: fa621c887897da5a379c719399c17db7
 
@@ -32,13 +31,15 @@
   tags:
     - adapterremoval
   files:
+    - path: output/adapterremoval/test.collapsed.gz
+      md5sum: ff956de3532599a56c3efe5369f0953f
+    - path: output/adapterremoval/test.collapsed.truncated.gz
     - path: output/adapterremoval/test.discarded.gz
     - path: output/adapterremoval/test.log
-      md5sum: b8a451d3981b327f3fdb44f40ba2d6d1
+      md5sum: 7f0b2328152226e46101a535cce718b3
     - path: output/adapterremoval/test.pair1.truncated.gz
-      md5sum: 294a6277f0139bd597e57c6fa31f39c7
+      md5sum: 683be19bc1c83008944b6b719bfa34e1
     - path: output/adapterremoval/test.pair2.truncated.gz
-      md5sum: de7b38e2c881bced8671acb1ab452d78
+      md5sum: e6548fe061f3ef86368b26da930174d0
-    - path: output/adapterremoval/test.singleton.truncated.gz
     - path: output/adapterremoval/versions.yml
-      md5sum: fd428f92a8446e0b34c5ae1c447215b8
+      md5sum: 78f589bb313c8da0147ca8ce77d7f3bf