Add FastK/Merge (#1828)

* Add FastK/Merge

* Update modules/fastk/merge/main.nf

* Update modules/fastk/merge/meta.yml

Co-authored-by: Sébastien Guizard <sguizard@ed.ac.uk>

Co-authored-by: Sébastien Guizard <sguizard@ed.ac.uk>

modules/fastk/fastk/main.nf

@@ -12,10 +12,10 @@ process FASTK_FASTK {
     tuple val(meta), path(reads)
 
     output:
-    tuple val(meta), path("*.hist")               , emit: hist
+    tuple val(meta), path("*.hist")                      , emit: hist
-    tuple val(meta), path("*.ktab*", hidden: true), emit: ktab, optional: true
+    tuple val(meta), path("*.ktab*", hidden: true)       , emit: ktab, optional: true
-    tuple val(meta), path("*.prof*", hidden: true), emit: prof, optional: true
+    tuple val(meta), path("*.{prof,pidx}*", hidden: true), emit: prof, optional: true
-    path "versions.yml"                           , emit: versions
+    path "versions.yml"                                  , emit: versions
 
     when:
     task.ext.when == null || task.ext.when

modules/fastk/merge/main.nf

@@ -0,0 +1,39 @@
+process FASTK_MERGE {
+    tag "$meta.id"
+    label 'process_medium'
+
+    if (params.enable_conda) {
+        error "Conda environments cannot be used when using the FastK tool. Please use docker or singularity containers."
+    }
+    // WARN: Version information not provided by tool on CLI. Please update version string below when bumping container versions.
+    container 'ghcr.io/nbisweden/fastk_genescopefk_merquryfk:1.0'
+
+    input:
+    tuple val(meta), path(hist), path(ktab), path(prof)
+
+    output:
+    tuple val(meta), path("*.hist")                      , emit: hist
+    tuple val(meta), path("*.ktab*", hidden: true)       , emit: ktab, optional: true
+    tuple val(meta), path("*.{prof,pidx}*", hidden: true), emit: prof, optional: true
+    path "versions.yml"                                  , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    def FASTK_VERSION = 'f18a4e6d2207539f7b84461daebc54530a9559b0' // WARN: Version information not provided by tool on CLI. Please update this string when bumping container versions.
+    """
+    Fastmerge \\
+        $args \\
+        -T$task.cpus \\
+        ${prefix} \\
+        $hist
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        fastk: $FASTK_VERSION
+    END_VERSIONS
+    """
+}

modules/fastk/merge/meta.yml

@@ -0,0 +1,60 @@
+name: "fastk_merge"
+description: A tool to merge FastK histograms
+keywords:
+  - merge
+  - k-mer
+  - histogram
+  - fastk
+tools:
+  - fastk:
+      description: "A fast K-mer counter for high-fidelity shotgun datasets"
+      homepage: "https://github.com/thegenemyers/FASTK"
+      documentation: ""
+      tool_dev_url: "https://github.com/thegenemyers/FASTK"
+      doi: ""
+      licence: "https://github.com/thegenemyers/FASTK/blob/master/LICENSE"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - fastk_hist:
+      type: file
+      description: A histogram files from the program FastK
+      pattern: "*.hist"
+  - fastk_ktab:
+      type: file
+      description: Histogram ktab files from the program FastK (option -t)
+      pattern: "*.ktab*"
+  - fastk_prof:
+      type: file
+      description: Histogram profile files from the program FastK (option -p)
+      pattern: "*.{prof,pidx}*"
+
+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"
+  - fastk_hist:
+      type: file
+      description: A histogram files from the program FastK
+      pattern: "*.hist"
+  - fastk_ktab:
+      type: file
+      description: Histogram ktab files from the program FastK (option -t)
+      pattern: "*.ktab*"
+  - fastk_prof:
+      type: file
+      description: Histogram profile files from the program FastK (option -p)
+      pattern: "*.{prof,pidx}*"
+
+authors:
+  - "@mahesh-panchal"

tests/config/pytest_modules.yml

@@ -695,6 +695,10 @@ fastk/histex:
   - modules/fastk/histex/**
   - tests/modules/fastk/histex/**
 
+fastk/merge:
+  - modules/fastk/merge/**
+  - tests/modules/fastk/merge/**
+
 fastp:
   - modules/fastp/**
   - tests/modules/fastp/**

tests/modules/fastk/fastk/test.yml

@@ -8,6 +8,8 @@
       md5sum: ceeacd0cb3aa69bf9b2a402830b40e26
     - path: output/fastk/.test_fk.ktab.2
       md5sum: f2629fd15b285aed3dc2d5fe546edf3f
+    - path: output/fastk/.test_fk.pidx.1
+      md5sum: 90bc384f61d2ecdb4586ab52ab04fddf
     - path: output/fastk/.test_fk.prof.1
       md5sum: ebd48923a724cf79934f0b2ed42ba73d
     - path: output/fastk/test_fk.hist
@@ -27,6 +29,10 @@
       md5sum: 7f28fb44940fda799797e3069f5d7263
     - path: output/fastk/.test_fk.ktab.2
       md5sum: c14a85c128926ace78372f09029977b1
+    - path: output/fastk/.test_fk.pidx.1
+      md5sum: e7e760f714070a4afefb38ffff559684
+    - path: output/fastk/.test_fk.pidx.2
+      md5sum: a549612bbdba2506eb3311237638c4b0
     - path: output/fastk/.test_fk.prof.1
       md5sum: 46a5fd9e297262b058f8c1fd062fcf56
     - path: output/fastk/.test_fk.prof.2

tests/modules/fastk/merge/main.nf

@@ -0,0 +1,58 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { FASTK_FASTK } from '../../../../modules/fastk/fastk/main.nf'
+include { FASTK_MERGE } from '../../../../modules/fastk/merge/main.nf'
+
+workflow test_fastk_merge_hist_only {
+
+    input1 = [
+        [ 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)
+        ]
+    ]
+    input2= [
+        [ 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)
+        ]
+    ]
+
+    FASTK_FASTK ( Channel.of( input1, input2 ) )
+    FASTK_MERGE (
+        FASTK_FASTK.out.hist.groupTuple()
+            .join( FASTK_FASTK.out.ktab.groupTuple(), remainder: true )
+            .join( FASTK_FASTK.out.prof.groupTuple(), remainder: true )
+            .map { meta, hist, ktab, prof -> [meta, hist, ktab ? ktab.flatten() : [] , prof ? prof.flatten() : [] ] }
+    )
+}
+
+workflow test_fastk_merge_all_files {
+
+    input1 = [
+        [ 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)
+        ]
+    ]
+    input2= [
+        [ 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)
+        ]
+    ]
+
+    FASTK_FASTK ( Channel.of( input1, input2 ) )
+    FASTK_MERGE (
+        FASTK_FASTK.out.hist.groupTuple()
+            .join( FASTK_FASTK.out.ktab.groupTuple(), remainder: true )
+            .join( FASTK_FASTK.out.prof.groupTuple(), remainder: true )
+            .map { meta, hist, ktab, prof -> [meta, hist, ktab ? ktab.flatten() : [] , prof ? prof.flatten() : [] ] }
+    )
+}

tests/modules/fastk/merge/nextflow.config

@@ -0,0 +1,14 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+    withName: 'test_fastk_merge_all_files:FASTK_FASTK' {
+        ext.args   = '-t1 -p'
+    }
+    withName: 'FASTK_FASTK' {
+        ext.args   = '-t1'
+        ext.prefix = { "${meta.id}_${task.index}" }
+        publishDir = [ enabled: false ]
+    }
+
+}

tests/modules/fastk/merge/test.yml

@@ -0,0 +1,39 @@
+- name: fastk merge test_fastk_merge_hist_only
+  command: nextflow run ./tests/modules/fastk/merge -entry test_fastk_merge_hist_only -c ./tests/config/nextflow.config  -c ./tests/modules/fastk/merge/nextflow.config
+  tags:
+    - fastk/merge
+    - fastk
+  files:
+    - path: output/fastk/.test.ktab.1
+      md5sum: dff1e9d326aea87778645235cfa3380f
+    - path: output/fastk/.test.ktab.2
+      md5sum: 438e7807dcc2eb8120e1338838147600
+    - path: output/fastk/test.hist
+      md5sum: 75d41eb0e3f8af5456711a95966e2b00
+    - path: output/fastk/test.ktab
+      md5sum: 7d882f4e3542df1a2f5cb0858b12dc03
+
+- name: fastk merge test_fastk_merge_all_files
+  command: nextflow run ./tests/modules/fastk/merge -entry test_fastk_merge_all_files -c ./tests/config/nextflow.config  -c ./tests/modules/fastk/merge/nextflow.config
+  tags:
+    - fastk/merge
+    - fastk
+  files:
+    - path: output/fastk/.test.ktab.1
+      md5sum: dff1e9d326aea87778645235cfa3380f
+    - path: output/fastk/.test.ktab.2
+      md5sum: 438e7807dcc2eb8120e1338838147600
+    - path: output/fastk/.test.pidx.1
+      md5sum: 7553344778f6571192f04c7e659b967c
+    - path: output/fastk/.test.pidx.2
+      md5sum: 09bb65fb10db58d1fd5bb0a123fccfc6
+    - path: output/fastk/.test.prof.1
+      md5sum: 7f0f84a62c89063ec18c9b0e447afe93
+    - path: output/fastk/.test.prof.2
+      md5sum: 7f0f84a62c89063ec18c9b0e447afe93
+    - path: output/fastk/test.hist
+      md5sum: 75d41eb0e3f8af5456711a95966e2b00
+    - path: output/fastk/test.ktab
+      md5sum: 7d882f4e3542df1a2f5cb0858b12dc03
+    - path: output/fastk/test.prof
+      md5sum: d3c7d8decd4ea6e298291b8be0e2de85