Add GeneScopeFK (#1781)

modules/genescopefk/main.nf

@@ -0,0 +1,42 @@
+process GENESCOPEFK {
+    tag "$meta.id"
+    label 'process_low'
+
+    if (params.enable_conda) {
+        error "Conda environments cannot be used when using the GeneScope tool. Please use docker or singularity containers."
+    }
+    container 'ghcr.io/nbisweden/fastk_genescopefk_merquryfk:1.0'
+
+    input:
+    tuple val(meta), path(fastk_histex_histogram)
+
+    output:
+    tuple val(meta), path("*_linear_plot.png")            , emit: linear_plot
+    tuple val(meta), path("*_log_plot.png")               , emit: log_plot
+    tuple val(meta), path("*_model.txt")                  , emit: model
+    tuple val(meta), path("*_summary.txt")                , emit: summary
+    tuple val(meta), path("*_transformed_linear_plot.png"), emit: transformed_linear_plot
+    tuple val(meta), path("*_transformed_log_plot.png")   , emit: transformed_log_plot
+    path "versions.yml"                                   , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    def GENESCOPE_VERSION = '380815c420f50171f9234a0fd1ff426b39829b91' // Git commit id is used instead of GeneScopeFK.R -v as software is not release versioned.
+    """
+    GeneScopeFK.R \\
+        $args \\
+        --input $fastk_histex_histogram \\
+        --output . \\
+        --name_prefix ${prefix}
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        genescope: $GENESCOPE_VERSION
+        r: \$( R --version | sed '1!d; s/.*version //; s/ .*//' )
+    END_VERSIONS
+    """
+}

modules/genescopefk/meta.yml

@@ -0,0 +1,63 @@
+name: "genescopefk"
+description: A derivative of GenomeScope2.0 modified to work with FastK
+keywords:
+  - k-mer
+  - genome profile
+  - histogram
+tools:
+  - "genescopefk":
+      description: "A derivative of GenomeScope2.0 modified to work with FastK"
+      homepage: "https://github.com/thegenemyers/GENESCOPE.FK"
+      documentation: ""
+      tool_dev_url: "https://github.com/thegenemyers/GENESCOPE.FK"
+      doi: ""
+      licence: "https://github.com/thegenemyers/GENESCOPE.FK/blob/main/LICENSE"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - fastk_histex_histogram:
+      type: file
+      description: A histogram formatted for GeneScope using the -G parameter from Fastk Histex
+      pattern: "*.hist"
+
+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"
+  - linear_plot:
+      type: file
+      description: A GeneScope linear plot in PNG format
+      pattern: "*_linear_plot.png"
+  - transformed_linear_plot:
+      type: file
+      description: A GeneScope transformed linear plot in PNG format
+      pattern: "*_transformed_linear_plot.png"
+  - log_plot:
+      type: file
+      description: A GeneScope log plot in PNG format
+      pattern: "*_log_plot.png"
+  - transformed_log_plot:
+      type: file
+      description: A GeneScope transformed log plot in PNG format
+      pattern: "*_transformed_log_plot.png"
+  - model:
+      type: file
+      description: GeneScope model fit summary
+      pattern: "*_model.txt"
+  - summary:
+      type: file
+      description: GeneScope histogram summary
+      pattern: "*_summary.txt"
+
+authors:
+  - "@mahesh-panchal"

tests/config/pytest_modules.yml

@@ -907,6 +907,10 @@ gatk4/variantrecalibrator:
   - modules/gatk4/variantrecalibrator/**
   - tests/modules/gatk4/variantrecalibrator/**
 
+genescopefk:
+  - modules/genescopefk/**
+  - tests/modules/genescopefk/**
+
 genmap/index:
   - modules/genmap/index/**
   - tests/modules/genmap/index/**

tests/modules/genescopefk/main.nf

@@ -0,0 +1,19 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { FASTK_FASTK  } from '../../../modules/fastk/fastk/main.nf'
+include { FASTK_HISTEX } from '../../../modules/fastk/histex/main.nf'
+include { GENESCOPEFK  } from '../../../modules/genescopefk/main.nf'
+
+workflow test_genescopefk {
+
+    input = [
+        [ id:'test' , single_end: true ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test_1_fastq_gz'], checkIfExists: true)
+    ]
+
+    FASTK_FASTK ( input )
+    FASTK_HISTEX ( FASTK_FASTK.out.hist )
+    GENESCOPEFK ( FASTK_HISTEX.out.hist )
+}

tests/modules/genescopefk/nextflow.config

@@ -0,0 +1,15 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+    withName: 'FASTK_.*' {
+        publishDir = [ enabled: false ]
+    }
+    withName: 'FASTK_HISTEX' {
+        ext.args = '-G'
+    }
+    withName: 'GENESCOPEFK' {
+        ext.args = '--kmer_length 40 --ploidy 1'
+    }
+
+}

tests/modules/genescopefk/test.yml

@@ -0,0 +1,17 @@
+- name: genescopefk test_genescopefk
+  command: nextflow run ./tests/modules/genescopefk -entry test_genescopefk -c ./tests/config/nextflow.config  -c ./tests/modules/genescopefk/nextflow.config
+  tags:
+    - genescopefk
+  files:
+    - path: output/genescopefk/test_linear_plot.png
+      md5sum: bb221e2e21a6c0d04dcb6ec8f1bf1cd7
+    - path: output/genescopefk/test_log_plot.png
+      md5sum: d71ad0c963775992043c8f3c55d27fc4
+    - path: output/genescopefk/test_model.txt
+      md5sum: 49e5a66cf02b53a3b20880bfabef2ee8
+    - path: output/genescopefk/test_summary.txt
+      md5sum: 34540dcc2ba4ab9530f8878bbdc0921c
+    - path: output/genescopefk/test_transformed_linear_plot.png
+      md5sum: 3c8e3ad315e6c15cdf51c87de277ec1f
+    - path: output/genescopefk/test_transformed_log_plot.png
+      md5sum: 90d983c6faa9b8a02a1d26b777feefc6