Merge pull request #65 from genomic-medicine-sweden/classification_kreport

Add centrifuge kreport in taxprofiler

conf/modules.config

@@ -254,6 +254,16 @@ process {
         ext.prefix = params.perform_runmerging ? { "${meta.id}-${meta.db_name}" } : { "${meta.id}-${meta.run_accession}-${meta.db_name}" }
     }
 
+    withName: CENTRIFUGE_KREPORT {
+        ext.args = { "${meta.db_params}" }
+        ext.prefix = params.perform_runmerging ? { "${meta.id}-${meta.db_name}" } : { "${meta.id}-${meta.run_accession}-${meta.db_name}" }
+        publishDir = [
+            path: { "${params.outdir}/centrifuge/${meta.db_name}" },
+            mode: params.publish_dir_mode,
+            pattern: '*.{txt}'
+        ]
+    }
+
     withName: CUSTOM_DUMPSOFTWAREVERSIONS {
         publishDir = [
             path: { "${params.outdir}/pipeline_info" },

modules.json

@@ -21,6 +21,9 @@
             "centrifuge/centrifuge": {
                 "git_sha": "d2726fcf75063960f06b36d2229a4c0966614108"
             },
+            "centrifuge/kreport": {
+                "git_sha": "be4ae28c3c95b3c4047a7d9fb4cb0ed749631cea"
+            },
             "custom/dumpsoftwareversions": {
                 "git_sha": "e745e167c1020928ef20ea1397b6b4d230681b4d"
             },
@@ -30,6 +33,9 @@
             "fastqc": {
                 "git_sha": "e745e167c1020928ef20ea1397b6b4d230681b4d"
             },
+            "kaiju/kaiju": {
+                "git_sha": "8856f127c58f6af479128be8b8df4d42e442ddbe"
+            },
             "kraken2/kraken2": {
                 "git_sha": "e745e167c1020928ef20ea1397b6b4d230681b4d"
             },
@@ -53,9 +59,6 @@
             },
             "untar": {
                 "git_sha": "e080f4c8acf5760039ed12ec1f206170f3f9a918"
-            },
-            "kaiju/kaiju": {
-                "git_sha": "8856f127c58f6af479128be8b8df4d42e442ddbe"
             }
         }
     }

modules/nf-core/modules/centrifuge/kreport/main.nf

@@ -0,0 +1,33 @@
+process CENTRIFUGE_KREPORT {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda (params.enable_conda ? "bioconda::centrifuge=1.0.4_beta" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/centrifuge:1.0.4_beta--h9a82719_6':
+        'quay.io/biocontainers/centrifuge:1.0.4_beta--h9a82719_6' }"
+
+    input:
+    tuple val(meta), path(results)
+    path db
+
+    output:
+    tuple val(meta), path('*.txt')                , emit: kreport
+    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}"
+    """
+    db_name=`find -L ${db} -name "*.1.cf" -not -name "._*"  | sed 's/.1.cf//'`
+    centrifuge-kreport -x \$db_name ${results} > ${prefix}.txt
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        centrifuge: \$( centrifuge --version  | sed -n 1p | sed 's/^.*centrifuge-class version //')
+    END_VERSIONS
+    """
+}

modules/nf-core/modules/centrifuge/kreport/meta.yml

@@ -0,0 +1,41 @@
+name: "centrifuge_kreport"
+description: Creates Kraken-style reports from centrifuge out files
+keywords:
+  - metagenomics
+tools:
+  - centrifuge:
+      description: Centrifuge is a classifier for metagenomic sequences.
+      homepage: https://ccb.jhu.edu/software/centrifuge/
+      documentation: https://ccb.jhu.edu/software/centrifuge/manual.shtml
+      doi: 10.1101/gr.210641.116
+      licence: ["GPL v3"]
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - results:
+      type: file
+      description: File containing the centrifuge classification results
+      pattern: "*.{txt}"
+
+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"
+  - kreport:
+      type: file
+      description: |
+        File containing kraken-style report from centrifuge
+        out files.
+      pattern: "*.{txt}"
+authors:
+  - "@sofstam"
+  - "@jfy133"

subworkflows/local/profiling.nf

@@ -6,6 +6,7 @@ include { MALT_RUN                    } from '../../modules/nf-core/modules/malt
 include { MEGAN_RMA2INFO              } from '../../modules/nf-core/modules/megan/rma2info/main'
 include { KRAKEN2_KRAKEN2             } from '../../modules/nf-core/modules/kraken2/kraken2/main'
 include { CENTRIFUGE_CENTRIFUGE       } from '../../modules/nf-core/modules/centrifuge/centrifuge/main'
+include { CENTRIFUGE_KREPORT          } from '../../modules/nf-core/modules/centrifuge/kreport/main'
 include { METAPHLAN3                  } from '../../modules/nf-core/modules/metaphlan3/main'
 include { KAIJU_KAIJU                 } from '../../modules/nf-core/modules/kaiju/kaiju/main'
 
@@ -142,8 +143,9 @@ workflow PROFILING {
 
     if ( params.run_centrifuge ) {
         CENTRIFUGE_CENTRIFUGE ( ch_input_for_centrifuge.reads, ch_input_for_centrifuge.db, params.centrifuge_save_unaligned, params.centrifuge_save_aligned, params.centrifuge_sam_format  )
+        CENTRIFUGE_KREPORT (CENTRIFUGE_CENTRIFUGE.out.results, ch_input_for_centrifuge.db)
         ch_versions        = ch_versions.mix( CENTRIFUGE_CENTRIFUGE.out.versions.first() )
-        ch_raw_profiles    = ch_raw_profiles.mix( CENTRIFUGE_CENTRIFUGE.out.report )
+        ch_raw_profiles    = ch_raw_profiles.mix( CENTRIFUGE_KREPORT.out.kreport )
     }
 
     if ( params.run_metaphlan3 ) {