Merge pull request #130 from KevinMenden/master

Adding a QUAST module

.github/workflows/quast.yml

@@ -0,0 +1,40 @@
+name: quast
+on:
+  push:
+    paths:
+      - software/quast/**
+      - .github/workflows/quast.yml
+      - tests
+  pull_request:
+    paths:
+      - software/pquast/**
+      - .github/workflows/quest.yml
+      - tests
+
+jobs:
+  ci_test:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        nxf_version: [20.11.0-edge]
+    env:
+      NXF_ANSI_LOG: false
+    steps:
+      - uses: actions/checkout@v2
+
+      - name: Install Nextflow
+        env:
+          NXF_VER: ${{ matrix.nxf_version }}
+        run: |
+          wget -qO- get.nextflow.io | bash
+          sudo mv nextflow /usr/local/bin/
+
+      - name: Set up Python
+        uses: actions/setup-python@v2
+        with:
+          python-version: "3.x"
+      - name: Install dependencies
+        run: python -m pip install --upgrade pip pytest-workflow
+
+      # Test the module
+      - run: pytest --tag quast --symlink --wt 2

software/quast/functions.nf

@@ -0,0 +1,59 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args          = args.args ?: ''
+    options.args2         = args.args2 ?: ''
+    options.publish_by_id = args.publish_by_id ?: false
+    options.publish_dir   = args.publish_dir ?: ''
+    options.publish_files = args.publish_files
+    options.suffix        = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }  // Remove empty entries
+    paths = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_id) {
+            path_list.add(args.publish_id)
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

software/quast/main.nf

@@ -0,0 +1,48 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+def options    = initOptions(params.options)
+
+process QUAST {
+    label 'process_medium'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), publish_id:'') }
+
+    conda (params.enable_conda ? 'bioconda::quast=5.0.2' : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container 'https://depot.galaxyproject.org/singularity/quast:5.0.2--py37pl526hb5aa323_2'
+    } else {
+        container 'quay.io/biocontainers/quast:5.0.2--py37pl526hb5aa323_2'
+    }
+
+    input:
+    path consensus
+    path fasta
+    path gff
+    val use_fasta
+    val use_gff
+
+    output:
+    path "${prefix}"    , emit: results
+    path '*.tsv'        , emit: tsv
+    path '*.version.txt', emit: version
+
+    script:
+    def software  = getSoftwareName(task.process)
+    prefix        = options.suffix ?: software
+    def features  = use_gff ? "--features $gff" : ''
+    def reference = use_fasta ? "-r $fasta" : ''
+    """
+    quast.py \\
+        --output-dir $prefix \\
+        $reference \\
+        $features \\
+        --threads $task.cpus \\
+        $options.args \\
+        ${consensus.join(' ')}
+    ln -s ${prefix}/report.tsv
+    echo \$(quast.py --version 2>&1) | sed 's/^.*QUAST v//; s/ .*\$//' > ${software}.version.txt
+    """
+}

software/quast/meta.yml

@@ -0,0 +1,68 @@
+name: quast
+description: Quality Assessment Tool for Genome Assemblies
+keywords:
+  - quast
+  - assembly
+  - quality
+tools:
+  - quast:
+      description: |
+        QUAST calculates quality metrics for genome assemblies
+      homepage: http://bioinf.spbau.ru/quast
+      doi:
+params:
+  - outdir:
+      type: string
+      description: |
+        The pipeline's output directory. By default, the module will
+        output files into `$params.outdir/<SOFTWARE>`
+  - publish_dir_mode:
+      type: string
+      description: |
+        Value for the Nextflow `publishDir` mode parameter.
+        Available: symlink, rellink, link, copy, copyNoFollow, move.
+  - enable_conda:
+      type: boolean
+      description: |
+        Run the module with Conda using the software specified
+        via the `conda` directive
+  - singularity_pull_docker_container:
+      type: boolean
+      description: |
+        Instead of directly downloading Singularity images for use with Singularity,
+        force the workflow to pull and convert Docker containers instead.
+
+input:
+  - consensus:
+      type: file
+      description: |
+        Fasta file containing the assembly of interest
+  - fasta:
+      type: file
+      description: |
+        The genome assembly to be evaluated. Has to contain at least a non-empty string dummy value.
+  - use_fasta:
+      type: boolean
+      description: Whether to use the provided fasta reference genome file
+  - gff:
+      type: file
+      description: The genome GFF file. Has to contain at least a non-empty string dummy value.
+  - use_gff:
+      type: boolean
+      description: Whether to use the provided gff reference annotation file
+
+output:
+  - quast:
+      type: directory
+      description: Directory containing complete quast report
+      pattern: "{prefix}.lineage_report.csv"
+  - report:
+
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+
+authors:
+  - "@drpatelh"
+  - "@kevinmenden"

tests/software/quast/main.nf

@@ -0,0 +1,25 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { QUAST }   from '../../../software/quast/main.nf'   addParams(options: [:])
+
+workflow test_quast_ref {
+    consensus = file("${launchDir}/tests/data/fasta/sarscov2/GCA_011545545.1_ASM1154554v1_cds_from_genomic.fna", checkIfExists: true)
+    gff = file("${launchDir}/tests/data/gff/sarscov2/GCA_011545545.1_ASM1154554v1_genomic.gtf", checkIfExists: true)
+    fasta = file("${launchDir}/tests/data/fasta/sarscov2/GCA_011545545.1_ASM1154554v1_genomic.fna", checkIfExists: true)
+    def use_fasta = true
+    def use_gtf = true
+
+    QUAST( consensus, fasta, gff, use_fasta, use_gtf )
+}
+
+workflow test_quast_noref {
+    consensus = file("${launchDir}/tests/data/fasta/sarscov2/GCA_011545545.1_ASM1154554v1_cds_from_genomic.fna", checkIfExists: true)
+    gff = file('gff_dummy')
+    fasta = file('fasta_dummy')
+    def use_fasta = false
+    def use_gtf = false
+
+    QUAST( consensus, fasta, gff, use_fasta, use_gtf )
+}

tests/software/quast/test.yml

@@ -0,0 +1,120 @@
+- name: quast_ref
+  command: nextflow run ./tests/software/quast -profile docker -entry test_quast_ref -c ./tests/config/nextflow.config
+  tags:
+    - quast
+    - quast_reference
+  files:
+    - path: ./output/quast/quast/basic_stats/NGx_plot.pdf
+    - path: ./output/quast/quast/contigs_reports/minimap_output/GCA_011545545-1_ASM1154554v1_cds_from_genomic.coords.filtered
+      md5sum: ec9191d0acb5d5bce56b4842551a8598
+    - path: ./output/quast/quast/aligned_stats/NGAx_plot.pdf
+    - path: ./output/quast/quast/icarus_viewers/contig_size_viewer.html
+      md5sum: c42838b1553759fb0460a9854e39da36
+    - path: ./output/quast/report.tsv
+      md5sum: b203a9d2c995bd66552eeb5cb0a98762
+    - path: ./output/quast/quast/report.tex
+      md5sum: 8168f197dfb2f781accdefaa0d90fc87
+    - path: ./output/quast/quast/contigs_reports/misassemblies_frcurve_plot.pdf
+    - path: ./output/quast/quast/transposed_report.tex
+      md5sum: 0b8fd00649a5758024c8093d5919f71c
+    - path: ./output/quast/quast/contigs_reports/minimap_output/GCA_011545545-1_ASM1154554v1_cds_from_genomic.coords
+      md5sum: dda3fc0addc41ecc0d5183dee6f95886
+    - path: ./output/quast/quast/transposed_report.txt
+      md5sum: 73216314e65ef91c24422d5bdd0795e4
+    - path: ./output/quast/quast/report.html
+    - path: ./output/quast/quast/report.pdf
+    - path: ./output/quast/quast/contigs_reports/all_alignments_GCA_011545545-1_ASM1154554v1_cds_from_genomic.tsv
+      md5sum: c247152eb82b361106492642fd796e2c
+    - path: ./output/quast/quast/contigs_reports/minimap_output/GCA_011545545-1_ASM1154554v1_cds_from_genomic.used_snps.gz
+      md5sum: 7b1db1b433cd95243a949bcb72e7e3a6
+    - path: ./output/quast/quast/contigs_reports/unaligned_report.tex
+      md5sum: 48562e30903fcf33946e13b602b0e324
+    - path: ./output/quast/quast/contigs_reports/minimap_output/GCA_011545545-1_ASM1154554v1_cds_from_genomic.sf
+    - path: ./output/quast/quast/contigs_reports/transposed_report_misassemblies.txt
+      md5sum: 53c481fe8b3ef18280ea86cbded78b31
+    - path: ./output/quast/quast/contigs_reports/contigs_report_GCA_011545545-1_ASM1154554v1_cds_from_genomic.mis_contigs.info
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: ./output/quast/quast/contigs_reports/unaligned_report.tsv
+      md5sum: 769f2d9d7deb5e7a63e238d8ba0cfe13
+    - path: ./output/quast/quast/basic_stats/GCA_011545545.1_ASM1154554v1_cds_from_genomic_GC_content_plot.pdf
+    - path: ./output/quast/quast/genome_stats/features_cumulative_plot.pdf
+    - path: ./output/quast/quast/basic_stats/GC_content_plot.pdf
+    - path: ./output/quast/quast/contigs_reports/transposed_report_misassemblies.tex
+      md5sum: 85614257e0bb62ac750f7872886b2fbf
+    - path: ./output/quast/quast/quast.log
+    - path: ./output/quast/quast/contigs_reports/misassemblies_report.tsv
+      md5sum: 33144f8d064782864e8400e3a08b2a3a
+    - path: ./output/quast/quast/contigs_reports/contigs_report_GCA_011545545-1_ASM1154554v1_cds_from_genomic.unaligned.info
+      md5sum: a8505cf206bf53ca369f7e3073fee587
+    - path: ./output/quast/quast/genome_stats/GCA_011545545-1_ASM1154554v1_cds_from_genomic_genomic_features_any.txt
+      md5sum: 307b3bb1f42fc2f11a60a2e9846021d7
+    - path: ./output/quast/quast/icarus.html
+      md5sum: 6e57f5f7d07528496fc2f35b449951fe
+    - path: ./output/quast/quast/contigs_reports/transposed_report_misassemblies.tsv
+      md5sum: 4d24cc0da823a100a076aef36f9fd75e
+    - path: ./output/quast/quast/contigs_reports/contigs_report_GCA_011545545-1_ASM1154554v1_cds_from_genomic.stderr
+    - path: ./output/quast/quast/aligned_stats/cumulative_plot.pdf
+    - path: ./output/quast/quast/genome_stats/genome_info.txt
+      md5sum: bc3df27ad7356c045c9a108f9d4bcfbe
+    - path: ./output/quast/quast/genome_stats/features_frcurve_plot.pdf
+    - path: ./output/quast/quast/contigs_reports/misassemblies_report.txt
+      md5sum: 0784e0d31e9f2a165730cc0029023427
+    - path: ./output/quast/quast/contigs_reports/minimap_output/GCA_011545545-1_ASM1154554v1_cds_from_genomic.unaligned
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: ./output/quast/quast/contigs_reports/contigs_report_GCA_011545545-1_ASM1154554v1_cds_from_genomic.stdout
+    - path: ./output/quast/quast/contigs_reports/unaligned_report.txt
+      md5sum: c1c26797084ea743ec5a224ab97adb1f
+    - path: ./output/quast/quast/genome_stats/GCA_011545545-1_ASM1154554v1_cds_from_genomic_gaps.txt
+      md5sum: c52381f09ea40b6141be5232494727b6
+    - path: ./output/quast/quast/icarus_viewers/alignment_viewer.html
+      md5sum: ad5176f21d3c4809a9a8f9c27bb8d70a
+    - path: ./output/quast/quast/contigs_reports/misassemblies_report.tex
+      md5sum: ffe19cf4d5332952f0627e9cd82e3375
+    - path: ./output/quast/quast/contigs_reports/minimap_output/GCA_011545545-1_ASM1154554v1_cds_from_genomic.coords_tmp
+      md5sum: ce66eaeb99fdc11e4d50efadc1816e04
+    - path: ./output/quast/quast/basic_stats/gc.icarus.txt
+      md5sum: b3a770802ff9b2dd4ee8e47bddb2df3e
+    - path: ./output/quast/quast/basic_stats/Nx_plot.pdf
+    - path: ./output/quast/quast/transposed_report.tsv
+      md5sum: 1754bd6c6fa8a0172a342a48dd3ca505
+    - path: ./output/quast/quast/basic_stats/cumulative_plot.pdf
+    - path: ./output/quast/quast/report.txt
+      md5sum: e334b3390a5028cbf88fe4ccab2b6499
+    - path: ./output/quast/quast/contigs_reports/misassemblies_plot.pdf
+    - path: ./output/quast/quast/report.tsv
+      md5sum: b203a9d2c995bd66552eeb5cb0a98762
+    - path: ./output/quast/quast/contigs_reports/GCA_011545545_1_ASM1154554v1_cds_from_genomic.mis_contigs.fa
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: ./output/quast/quast/aligned_stats/NAx_plot.pdf
+
+- name: quast_noref
+  command: nextflow run ./tests/software/quast -profile docker -entry test_quast_noref -c ./tests/config/nextflow.config
+  tags:
+    - quast
+    - quast_no_reference
+  files:
+    - path: ./output/quast/quast/icarus_viewers/contig_size_viewer.html
+      md5sum: d2680b8083139e62baf7af0f852d52b8
+    - path: ./output/quast/quast/transposed_report.txt
+      md5sum: b6b126e9dcfd8c8e7bd36dff9220caf0
+    - path: ./output/quast/report.tsv
+      md5sum: 55d2938fcdb678e97d441cf32b9c740e
+    - path: ./output/quast/quast/icarus.html
+      md5sum: dcb11b4c7b6c21190c1344c46288d3f4
+    - path: ./output/quast/quast/report.txt
+      md5sum: c8e87f40e888a0395be04f78532d9deb
+    - path: ./output/quast/quast/transposed_report.tsv
+      md5sum: 27b6c755be5d70e3eed22e1bf3f68ce5
+    - path: ./output/quast/quast/quast.log
+    - path: ./output/quast/quast/report.tex
+      md5sum: 69c31172ae6257663b22bb192ba84682
+    - path: ./output/quast/quast/report.pdf
+    - path: ./output/quast/quast/basic_stats/GCA_011545545.1_ASM1154554v1_cds_from_genomic_GC_content_plot.pdf
+    - path: ./output/quast/quast/basic_stats/cumulative_plot.pdf
+    - path: ./output/quast/quast/report.tsv
+      md5sum: 55d2938fcdb678e97d441cf32b9c740e
+    - path: ./output/quast/quast/transposed_report.tex
+      md5sum: 213fb615afe21904e52d670e602c1f62
+    - path: ./output/quast/quast/basic_stats/GC_content_plot.pdf
+    - path: ./output/quast/quast/basic_stats/Nx_plot.pdf
+    - path: ./output/quast/quast/report.html