Add kaiju_kaiju module (#1448)

* Added kaiju_kaiju module

* Update modules/kaiju/kaiju/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/kaiju/kaiju/main.nf

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update modules/kaiju/kaiju/meta.yml

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update keywords

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

* Update output file naming

* Update output file naming

* update spacing for lint

* Update input file patterns

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>

modules/kaiju/kaiju/main.nf

@@ -0,0 +1,39 @@
+process KAIJU_KAIJU {
+    tag "$meta.id"
+    label 'process_high'
+
+    conda (params.enable_conda ? "bioconda::kaiju=1.8.2" : null)
+    container "${ 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--h5b5514e_1' }"
+
+    input:
+    tuple val(meta), path(reads)
+    tuple path(db), path(dbnodes)
+
+    output:
+    tuple val(meta), path('*.tsv'), emit: results
+    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 input = meta.single_end ? "-i ${reads}" : "-i ${reads[0]} -j ${reads[1]}"
+    """
+    kaiju \\
+        $args \\
+        -z $task.cpus \\
+        -t ${dbnodes} \\
+        -f ${db} \\
+        -o ${prefix}.tsv \\
+        $input
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        kaiju: \$(echo \$( kaiju -h 2>&1 | sed -n 1p | sed 's/^.*Kaiju //' ))
+    END_VERSIONS
+    """
+}

modules/kaiju/kaiju/meta.yml

@@ -0,0 +1,52 @@
+name: kaiju_kaiju
+description: Taxonomic classification of metagenomic sequence data using a protein reference database
+keywords:
+  - classify
+  - metagenomics
+  - fastq
+  - taxonomic profiling
+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 ]
+  - reads:
+      type: file
+      description: |
+        List of input fastq/fasta files of size 1 and 2 for single-end and paired-end data,
+        respectively.
+      pattern: "*.{fastq,fq,fasta,fa,fsa,fas,fna,fastq.gz,fq.gz,fasta.gz,fa.gz,fsa.gz,fas.gz,fna.gz}"
+  - db:
+      type: files
+      description: |
+        List containing the database and nodes files for Kaiju
+        e.g. [ 'database.fmi', 'nodes.dmp' ]
+
+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: Results with taxonomic classification of each read
+      pattern: "*.tsv"
+
+authors:
+  - "@talnor"
+  - "@sofstam"

tests/config/pytest_modules.yml

@@ -961,6 +961,10 @@ jupyternotebook:
   - modules/jupyternotebook/**
   - tests/modules/jupyternotebook/**
 
+kaiju/kaiju:
+  - modules/kaiju/kaiju/**
+  - tests/modules/kaiju/kaiju/**
+
 kallisto/index:
   - modules/kallisto/index/**
   - tests/modules/kallisto/index/**

tests/config/test_data.config

@@ -28,6 +28,10 @@ params {
                 kraken2_bracken                                = "${test_data_dir}/genomics/sarscov2/genome/db/kraken2_bracken"
                 kraken2_bracken_tar_gz                         = "${test_data_dir}/genomics/sarscov2/genome/db/kraken2_bracken.tar.gz"
 
+                kaiju_fmi                                      = "${test_data_dir}/genomics/sarscov2/genome/db/kaiju/proteins.fmi"
+                kaiju_nodes                                    = "${test_data_dir}/genomics/sarscov2/genome/db/kaiju/nodes.dmp"
+                kaiju_names                                    = "${test_data_dir}/genomics/sarscov2/genome/db/kaiju/names.dmp"
+
                 ncbi_taxmap_zip                                = "${test_data_dir}/genomics/sarscov2/genome/db/maltextract/ncbi_taxmap.zip"
                 taxon_list_txt                                 = "${test_data_dir}/genomics/sarscov2/genome/db/maltextract/taxon_list.txt"
 

tests/modules/kaiju/kaiju/main.nf

@@ -0,0 +1,34 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { KAIJU_KAIJU } from '../../../../modules/kaiju/kaiju/main.nf'
+
+workflow test_kaiju_kaiju_single_end {
+
+    input = [
+        [ id:'test', single_end:true ], // meta map
+        file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true)
+    ]
+    db    = [
+        file(params.test_data['sarscov2']['genome']['kaiju_fmi'], checkIfExists: true), // database
+        file(params.test_data['sarscov2']['genome']['kaiju_nodes'], checkIfExists: true) // taxon nodes
+    ]
+
+    KAIJU_KAIJU ( input, db )
+}
+
+workflow test_kaiju_kaiju_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) ]
+    ]
+    db    = [
+        file(params.test_data['sarscov2']['genome']['kaiju_fmi'], checkIfExists: true), // database
+        file(params.test_data['sarscov2']['genome']['kaiju_nodes'], checkIfExists: true) // taxon nodes
+    ]
+
+    KAIJU_KAIJU ( input, db )
+}

tests/modules/kaiju/kaiju/nextflow.config

@@ -0,0 +1,5 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+    
+}

tests/modules/kaiju/kaiju/test.yml

@@ -0,0 +1,21 @@
+- name: kaiju kaiju test_kaiju_kaiju_single_end
+  command: nextflow run tests/modules/kaiju/kaiju -entry test_kaiju_kaiju_single_end -c tests/config/nextflow.config
+  tags:
+    - kaiju/kaiju
+    - kaiju
+  files:
+    - path: output/kaiju/test.tsv
+      contains: ["C\tERR5069949.2257580\t2697049"]
+    - path: output/kaiju/versions.yml
+      md5sum: 7e218c0ea00a71dd3a5ec5aaf28804f4
+
+- name: kaiju kaiju test_kaiju_kaiju_paired_end
+  command: nextflow run tests/modules/kaiju/kaiju -entry test_kaiju_kaiju_paired_end -c tests/config/nextflow.config
+  tags:
+    - kaiju/kaiju
+    - kaiju
+  files:
+    - path: output/kaiju/test.tsv
+      contains: ["C\tERR5069949.2257580\t2697049"]
+    - path: output/kaiju/versions.yml
+      md5sum: a74215f6f69979ae046fb1d65c56ac67