Unicycler with long read input (#1041)

* Unicycler with long read input

* tests and md5sums

* remove unstable md5sums

* Update modules/unicycler/main.nf

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

modules/unicycler/main.nf

@@ -19,26 +19,30 @@ process UNICYCLER {
     }
 
     input:
-    tuple val(meta), path(reads)
+    tuple val(meta), path(shortreads), path(longreads)
 
     output:
-    tuple val(meta), path('*.scaffolds.fa'), emit: scaffolds
-    tuple val(meta), path('*.assembly.gfa'), emit: gfa
-    tuple val(meta), path('*.log')         , emit: log
-    path  "versions.yml"                   , emit: versions
+    tuple val(meta), path('*.scaffolds.fa.gz'), emit: scaffolds
+    tuple val(meta), path('*.assembly.gfa.gz'), emit: gfa
+    tuple val(meta), path('*.log')            , emit: log
+    path  "versions.yml"                      , emit: versions
 
     script:
     def prefix      = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
-    def input_reads = meta.single_end ? "-s $reads" : "-1 ${reads[0]} -2 ${reads[1]}"
+    def short_reads = shortreads ? ( meta.single_end ? "-s $shortreads" : "-1 ${shortreads[0]} -2 ${shortreads[1]}" ) : ""
+    def long_reads  = longreads ? "-l $longreads" : ""
     """
     unicycler \\
         --threads $task.cpus \\
         $options.args \\
-        $input_reads \\
+        $short_reads \\
+        $long_reads \\
         --out ./
 
     mv assembly.fasta ${prefix}.scaffolds.fa
+    gzip -n ${prefix}.scaffolds.fa
     mv assembly.gfa ${prefix}.assembly.gfa
+    gzip -n ${prefix}.assembly.gfa
     mv unicycler.log ${prefix}.unicycler.log
 
     cat <<-END_VERSIONS > versions.yml

modules/unicycler/meta.yml

@@ -19,11 +19,15 @@ input:
       description: |
         Groovy Map containing sample information
         e.g. [ id:'test', single_end:false ]
-  - reads:
+  - shortreads:
           type: file
           description: |
-              List of input FastQ files of size 1 and 2 for single-end and paired-end data,
+              List of input Illumina FastQ files of size 1 and 2 for single-end and paired-end data,
               respectively.
+  - longreads:
+          type: file
+          description: |
+              List of input FastQ files of size 1, PacBio or Nanopore long reads.
 output:
   - meta:
       type: map
@@ -37,11 +41,11 @@ output:
   - scaffolds:
       type: file
       description: Fasta file containing scaffolds
-      pattern: "*.{scaffolds.fa}"
+      pattern: "*.{scaffolds.fa.gz}"
   - gfa:
       type: file
       description: gfa file containing assembly
-      pattern: "*.{assembly.gfa}"
+      pattern: "*.{assembly.gfa.gz}"
   - log:
       type: file
       description: unicycler log file
@@ -53,3 +57,4 @@ output:
 authors:
   - "@JoseEspinosa"
   - "@drpatelh"
+  - "@d4straub"

tests/modules/unicycler/main.nf

@@ -6,7 +6,8 @@ include { UNICYCLER } from '../../../modules/unicycler/main.nf' addParams( optio
 
 workflow test_unicycler_single_end {
     input = [ [ id:'test', single_end:true ], // meta map
-              [ file(params.test_data['sarscov2']['nanopore']['test_fastq_gz'], checkIfExists: true) ]
+              [ file(params.test_data['bacteroides_fragilis']['illumina']['test1_1_fastq_gz'], checkIfExists: true) ],
+              []
             ]
 
     UNICYCLER ( input )
@@ -14,8 +15,19 @@ workflow test_unicycler_single_end {
 
 workflow test_unicycler_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) ]
+              [ file(params.test_data['bacteroides_fragilis']['illumina']['test1_1_fastq_gz'], checkIfExists: true),
+                file(params.test_data['bacteroides_fragilis']['illumina']['test1_2_fastq_gz'], checkIfExists: true) ],
+              []
+            ]
+
+    UNICYCLER ( input )
+}
+
+workflow test_unicycler_shortreads_longreads {
+    input = [ [ id:'test', single_end:false ], // meta map
+              [ file(params.test_data['bacteroides_fragilis']['illumina']['test1_1_fastq_gz'], checkIfExists: true),
+                file(params.test_data['bacteroides_fragilis']['illumina']['test1_2_fastq_gz'], checkIfExists: true) ],
+              [ file(params.test_data['bacteroides_fragilis']['nanopore']['test_fastq_gz'], checkIfExists: true) ]
             ]
 
     UNICYCLER ( input )

tests/modules/unicycler/test.yml

@@ -1,21 +1,32 @@
-- name: unicycler single-end
-  command: nextflow run ./tests/modules/unicycler -entry test_unicycler_single_end -c tests/config/nextflow.config
+- name: unicycler test_unicycler_single_end
+  command: nextflow run tests/modules/unicycler -entry test_unicycler_single_end -c tests/config/nextflow.config
   tags:
     - unicycler
   files:
-    - path: output/unicycler/test.scaffolds.fa
-    - path: output/unicycler/test.assembly.gfa
+    - path: output/unicycler/test.assembly.gfa.gz
+    - path: output/unicycler/test.scaffolds.fa.gz
     - path: output/unicycler/test.unicycler.log
       contains:
         - "Assembly complete"
 
-- name: unicycler paired-end
-  command: nextflow run ./tests/modules/unicycler -entry test_unicycler_paired_end -c tests/config/nextflow.config
+- name: unicycler test_unicycler_paired_end
+  command: nextflow run tests/modules/unicycler -entry test_unicycler_paired_end -c tests/config/nextflow.config
   tags:
     - unicycler
   files:
-    - path: output/unicycler/test.scaffolds.fa
-    - path: output/unicycler/test.assembly.gfa
+    - path: output/unicycler/test.assembly.gfa.gz
+    - path: output/unicycler/test.scaffolds.fa.gz
+    - path: output/unicycler/test.unicycler.log
+      contains:
+        - "Assembly complete"
+
+- name: unicycler test_unicycler_shortreads_longreads
+  command: nextflow run tests/modules/unicycler -entry test_unicycler_shortreads_longreads -c tests/config/nextflow.config
+  tags:
+    - unicycler
+  files:
+    - path: output/unicycler/test.assembly.gfa.gz
+    - path: output/unicycler/test.scaffolds.fa.gz
     - path: output/unicycler/test.unicycler.log
       contains:
         - "Assembly complete"