Add module: Clustalo align (#3734)

* Add modules - test and linting to be fied when guidetree gets merged to main

* Fix tests

* Remove trailing whitespace

* Update modules/nf-core/clustalo/align/main.nf

Co-authored-by: Sateesh_Peri <33637490+sateeshperi@users.noreply.github.com>

* Update modules/nf-core/clustalo/align/meta.yml

Co-authored-by: Sateesh_Peri <33637490+sateeshperi@users.noreply.github.com>

* Add options tree directly in args

---------

Co-authored-by: Sateesh_Peri <33637490+sateeshperi@users.noreply.github.com>

modules/nf-core/clustalo/align/main.nf

@@ -0,0 +1,48 @@
+process CLUSTALO_ALIGN {
+    tag "$meta.id"
+    label 'process_medium'
+
+    conda "bioconda::clustalo=1.2.4"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/clustalo:1.2.4--h87f3376_5':
+        'biocontainers/clustalo:1.2.4--h87f3376_5' }"
+
+    input:
+    tuple val(meta),  path(fasta)
+    tuple val(meta2), path(tree)
+
+    output:
+    tuple val(meta), path("*.aln"), emit: alignment
+    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}"
+    """
+    clustalo \\
+        -i ${fasta} \\
+        --threads=${task.cpus} \\
+        $args \\
+        -o ${prefix}.aln
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        clustalo: \$( clustalo --version )
+    END_VERSIONS
+    """
+
+    stub:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    touch ${prefix}.aln
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        clustalo: \$( clustalo --version )
+    END_VERSIONS
+    """
+}

modules/nf-core/clustalo/align/meta.yml

@@ -0,0 +1,57 @@
+name: "clustalo_align"
+description: Align sequences using Clustal Omega
+keywords:
+  - alignment
+  - MSA
+  - genomics
+tools:
+  - "clustalo":
+      description: "Latest version of Clustal: a multiple sequence alignment program for DNA or proteins"
+      homepage: "http://www.clustal.org/omega/"
+      documentation: "http://www.clustal.org/omega/"
+      tool_dev_url: "http://www.clustal.org/omega/"
+      doi: "10.1038/msb.2011.75"
+      licence: "['GPL v2']"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'test']`
+
+  - fasta:
+      type: file
+      description: Input sequences in FASTA format
+      pattern: "*.{fa,fasta}"
+
+  - meta2:
+      type: map
+      description: |
+        Groovy Map containing tree information
+        e.g. `[ id:'test_tree']`
+
+  - tree:
+      type: file
+      description: Input guide tree in Newick format
+      pattern: "*.{dnd}"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'test']`
+
+  - alignment:
+      type: file
+      description: Alignment file.
+      pattern: "*.{aln}"
+
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+
+authors:
+  - "@luisas"

tests/config/pytest_modules.yml

@@ -730,6 +730,10 @@ clonalframeml:
   - modules/nf-core/clonalframeml/**
   - tests/modules/nf-core/clonalframeml/**
 
+clustalo/align:
+  - modules/nf-core/clustalo/align/**
+  - tests/modules/nf-core/clustalo/align/**
+
 clustalo/guidetree:
   - modules/nf-core/clustalo/guidetree/**
   - tests/modules/nf-core/clustalo/guidetree/**

tests/modules/nf-core/clustalo/align/main.nf

@@ -0,0 +1,30 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { CLUSTALO_GUIDETREE } from '../../../../../modules/nf-core/clustalo/guidetree/main.nf'
+include { CLUSTALO_ALIGN     } from '../../../../../modules/nf-core/clustalo/align/main.nf'
+
+workflow test_clustalo_align {
+
+    input = [
+        [ id:'test' ], // meta map
+        file(params.test_data['sarscov2']['illumina']['contigs_fasta'], checkIfExists: true)
+    ]
+
+    CLUSTALO_ALIGN ( input, [[:],[]] )
+}
+
+
+workflow test_clustalo_align_with_tree {
+
+    input = [
+        [ id:'test' ], // meta map
+        file(params.test_data['sarscov2']['illumina']['contigs_fasta'], checkIfExists: true)
+    ]
+
+    ch_tree = CLUSTALO_GUIDETREE ( input ).tree
+
+    CLUSTALO_ALIGN ( input , ch_tree )
+
+}

tests/modules/nf-core/clustalo/align/nextflow.config

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

tests/modules/nf-core/clustalo/align/test.yml

@@ -0,0 +1,21 @@
+- name: clustalo align test_clustalo_align
+  command: nextflow run ./tests/modules/nf-core/clustalo/align -entry test_clustalo_align -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/clustalo/align/nextflow.config
+  tags:
+    - clustalo/align
+    - clustalo
+  files:
+    - path: output/clustalo/test.aln
+      md5sum: 74bb9a2820a91cf68db94dbd46787722
+    - path: output/clustalo/versions.yml
+
+- name: clustalo align test_clustalo_align_with_tree
+  command: nextflow run ./tests/modules/nf-core/clustalo/align -entry test_clustalo_align_with_tree -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/clustalo/align/nextflow.config
+  tags:
+    - clustalo/align
+    - clustalo
+  files:
+    - path: output/clustalo/test.aln
+      md5sum: 74bb9a2820a91cf68db94dbd46787722
+    - path: output/clustalo/test.dnd
+      md5sum: 5428bad500a0a0bd985744bec1a12a70
+    - path: output/clustalo/versions.yml