Merge pull request #67 from genomic-medicine-sweden/classification_kaiju2table

Add kaiju2table to taxprofiler

conf/modules.config

@@ -289,4 +289,14 @@ process {
         ext.args = { "${meta.db_params}" }
         ext.prefix = params.perform_runmerging ? { "${meta.id}-${meta.db_name}" } : { "${meta.id}-${meta.run_accession}-${meta.db_name}" }
     }
+
+        withName: KAIJU_KAIJU2TABLE {
+        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: { "${params.outdir}/kaiju/${meta.db_name}" },
+            mode: params.publish_dir_mode,
+            pattern: '*.{txt}'
+        ]
+    }
 }

modules.json

@@ -36,6 +36,9 @@
             "kaiju/kaiju": {
                 "git_sha": "8856f127c58f6af479128be8b8df4d42e442ddbe"
             },
+            "kaiju/kaiju2table": {
+                "git_sha": "538dbac98ba9c8f799536cd5a617195501439457"
+            },
             "kraken2/kraken2": {
                 "git_sha": "e745e167c1020928ef20ea1397b6b4d230681b4d"
             },

modules/nf-core/modules/kaiju/kaiju2table/main.nf

@@ -0,0 +1,40 @@
+process KAIJU_KAIJU2TABLE {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda (params.enable_conda ? "bioconda::kaiju=1.8.2" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/kaiju:1.8.2--h5b5514e_1':
+        'quay.io/biocontainers/kaiju:1.8.2--h2e03b76_0' }"
+
+    input:
+    tuple val(meta), path(results)
+    path db
+    val taxon_rank
+
+    output:
+    tuple val(meta), path('*.txt'), emit: summary
+    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}"
+    """
+    dbnodes=`find -L ${db} -name "*nodes.dmp"`
+    dbname=`find -L ${db} -name "*.fmi" -not -name "._*"`
+    kaiju2table   $args \\
+        -t \$dbnodes \\
+        -n \$dbname \\
+        -r ${taxon_rank} \\
+        -o ${prefix}.txt \\
+        ${results}
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        kaiju: \$(echo \$( kaiju -h 2>&1 | sed -n 1p | sed 's/^.*Kaiju //' ))
+    END_VERSIONS
+    """
+}

modules/nf-core/modules/kaiju/kaiju2table/meta.yml

@@ -0,0 +1,50 @@
+name: "kaiju_kaiju2table"
+description: write your description here
+keywords:
+  - classify
+  - metagenomics
+tools:
+  - kaiju:
+      description: Fast and sensitive taxonomic classification for metagenomics
+      homepage: https://kaiju.binf.ku.dk/
+      documentation: https://github.com/bioinformatics-centre/kaiju/blob/master/README.md
+      tool_dev_url: https://github.com/bioinformatics-centre/kaiju
+      doi: "10.1038/ncomms11257"
+      licence: ["GNU 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 kaiju classification results
+      pattern: "*.{txt}"
+  - taxon_rank:
+      type: string
+      description: |
+        Taxonomic rank to display in report
+        pattern: "phylum|class|order|family|genus|species"
+
+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"
+  - results:
+      type: file
+      description: |
+        Summary table for a given taxonomic rank
+      pattern: "*.{tsv}"
+
+authors:
+  - "@sofstam"
+  - "@talnor"
+  - "@jfy133"

nextflow.config

@@ -107,6 +107,7 @@ params {
 
     // kaiju
     run_kaiju                  = false
+    kaiju_taxon_name           = 'species'
 }
 
 // Load base.config by default for all pipelines

nextflow_schema.json

@@ -384,6 +384,11 @@
         },
         "malt_generatemegansummary": {
             "type": "boolean"
+        },
+        "kaiju_taxon_name": {
+            "type": "string",
+            "default": "species",
+            "enum": ["phylum", "class", "order", "family", "genus", "species"]
         }
     }
 }

subworkflows/local/profiling.nf

@@ -9,6 +9,7 @@ include { CENTRIFUGE_CENTRIFUGE       } from '../../modules/nf-core/modules/cent
 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'
+include { KAIJU_KAIJU2TABLE           } from '../../modules/nf-core/modules/kaiju/kaiju2table/main'
 
 workflow PROFILING {
     take:
@@ -155,8 +156,11 @@ workflow PROFILING {
     }
 
     if ( params.run_kaiju ) {
-        KAIJU_KAIJU ( ch_input_for_kaiju.reads, ch_input_for_kaiju.db )
+        KAIJU_KAIJU ( ch_input_for_kaiju.reads, ch_input_for_kaiju.db)
+        KAIJU_KAIJU2TABLE (KAIJU_KAIJU.out.results, ch_input_for_kaiju.db, params.kaiju_taxon_name)
+        ch_multiqc_files = ch_multiqc_files.mix( KAIJU_KAIJU2TABLE.out.summary.collect{it[1]}.ifEmpty([])  )
         ch_versions = ch_versions.mix( KAIJU_KAIJU.out.versions.first() )
+        ch_raw_profiles = ch_raw_profiles.mix( KAIJU_KAIJU2TABLE.out.summary )
     }
 
     emit: