Raxmlng (#443)

* new raxml module

* new raxml module

* pass in args for bootstrap and add test for support file

* remove unnecessary tag

* ensure tags meet guidleines

* Apply suggestions from code review

* Update to latest functions file

Co-authored-by: avantonder <avt@sanger.ac.uk>
Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

software/raxmlng/functions.nf

@@ -0,0 +1,70 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args            = args.args ?: ''
+    options.args2           = args.args2 ?: ''
+    options.args3           = args.args3 ?: ''
+    options.publish_by_meta = args.publish_by_meta ?: []
+    options.publish_dir     = args.publish_dir ?: ''
+    options.publish_files   = args.publish_files
+    options.suffix          = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
+    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions  = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_meta) {
+            def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
+            for (key in key_list) {
+                if (args.meta && key instanceof String) {
+                    def path = key
+                    if (args.meta.containsKey(key)) {
+                        path = args.meta[key] instanceof Boolean ? "${key}_${args.meta[key]}".toString() : args.meta[key]
+                    }
+                    path = path instanceof String ? path : ''
+                    path_list.add(path)
+                }
+            }
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

software/raxmlng/main.nf

@@ -0,0 +1,43 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process RAXMLNG {
+    label 'process_medium'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), meta:[:], publish_by_meta:[]) }
+
+    conda (params.enable_conda ? "bioconda::raxml-ng=1.0.2" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/raxml-ng:1.0.2--h7447c1b_0"
+    } else {
+        container "quay.io/biocontainers/raxml-ng:1.0.2--h7447c1b_0"
+    }
+
+    input:
+    path alignment
+
+    output:
+    path "*.raxml.bestTree", emit: phylogeny
+    path "*.raxml.support" , optional:true, emit: phylogeny_bootstrapped
+    path "*.version.txt"   , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+
+    if (options.args.contains('--bs-trees')) {
+        options.args = "--all ${options.args}"
+    }
+    """
+    raxml-ng \\
+        $options.args \\
+        --msa $alignment \\
+        --threads $task.cpus \\
+        --prefix output
+
+    echo \$(raxml-ng --version 2>&1) | sed 's/^.*RAxML-NG v. //; s/released.*\$//' > ${software}.version.txt
+    """
+}

software/raxmlng/meta.yml

@@ -0,0 +1,38 @@
+name: raxmlng
+description: RAxML-NG is a phylogenetic tree inference tool which uses maximum-likelihood (ML) optimality criterion.
+keywords:
+  - phylogeny
+  - newick
+  - maximum likelihood
+tools:
+  - raxmlng:
+      description: RAxML-NG is a phylogenetic tree inference tool which uses maximum-likelihood (ML) optimality criterion.
+      homepage: https://github.com/amkozlov/raxml-ng
+      documentation: https://github.com/amkozlov/raxml-ng/wiki
+      tool_dev_url: https://github.com/amkozlov/raxml-ng
+      doi: doi.org/10.1093/bioinformatics/btz305
+      licence: ['GPL v2-or-later']
+
+input:
+  - alignment:
+      type: file
+      description: A FASTA format multiple sequence alignment file
+      pattern: "*.{fasta,fas,fa,mfa}"
+
+output:
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+  - phylogeny:
+      type: file
+      description: A phylogeny in Newick format
+      pattern: "*.{raxml.bestTree}"
+  - phylogeny_bootstrapped:
+      type: file
+      description: A phylogeny in Newick format with bootstrap values
+      pattern: "*.{raxml.support}"
+
+authors:
+  - "@avantonder"
+  - "@aunderwo"

tests/config/pytest_software.yml

@@ -427,6 +427,10 @@ rasusa:
   - software/rasusa/**
   - tests/software/rasusa/**
 
+raxmlng:
+  - software/raxmlng/**
+  - tests/software/raxmlng/**
+
 rseqc/bamstat:
   - software/rseqc/bamstat/**
   - tests/software/rseqc/bamstat/**

tests/software/raxmlng/main.nf

@@ -0,0 +1,28 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { RAXMLNG as RAXMLNG_NO_BOOTSTRAP } from '../../../software/raxmlng/main.nf' addParams( options: [args:'--model GTR+G']                 )
+include { RAXMLNG as RAXMLNG_BOOTSTRAP    } from '../../../software/raxmlng/main.nf' addParams( options: [args:'--model GTR+G --bs-trees 1000'] )
+
+/*
+ * Test without bootstrapping
+ */
+
+workflow test_raxmlng_no_bootstrap {
+    
+    input = [ file(params.test_data['sarscov2']['genome']['informative_sites_fas'], checkIfExists: true) ]
+
+    RAXMLNG_NO_BOOTSTRAP ( input )
+}
+
+/*
+ * Test with bootstrapping
+ */
+
+workflow test_raxmlng_bootstrap {
+    
+    input = [ file(params.test_data['sarscov2']['genome']['informative_sites_fas'], checkIfExists: true) ]
+
+    RAXMLNG_BOOTSTRAP ( input )
+}

tests/software/raxmlng/test.yml

@@ -0,0 +1,31 @@
+- name: raxmlng no_bootstrap
+  command: nextflow run ./tests/software/raxmlng -entry test_raxmlng_no_bootstrap -c tests/config/nextflow.config
+  tags:
+    - raxmlng
+  files:
+    - path: output/raxmlng/output.raxml.bestTree
+      contains:
+        - 'sample1:0.359'
+        - 'sample2:1.50'
+        - 'sample3:0.000001'
+        - 'sample4:0.111'
+
+- name: raxmlng bootstrap
+  command: nextflow run ./tests/software/raxmlng -entry test_raxmlng_bootstrap -c tests/config/nextflow.config
+  tags:
+    - raxmlng
+  files:
+    - path: output/raxmlng/output.raxml.bestTree
+      contains:
+        - 'sample1:0.359'
+        - 'sample2:1.50'
+        - 'sample3:0.000001'
+        - 'sample4:0.111'
+    - path: output/raxmlng/output.raxml.support
+      contains:
+        - 'sample1:0.359'
+        - 'sample2:1.50'
+        - 'sample3:0.000001'
+        - 'sample4:0.111'
+      contains_regex:
+        - '\)[89]\d:'