Initial commit mummer module (#940)

This commit is contained in:
Michael J Cipriano 2021-11-15 10:53:41 -05:00 committed by GitHub
parent 1a4c7cec1b
commit 0754d49db8
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
6 changed files with 206 additions and 0 deletions

View file

@ -0,0 +1,78 @@
//
// 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()
}
//
// Extract name of module from process name using $task.process
//
def getProcessName(task_process) {
return task_process.tokenize(':')[-1]
}
//
// Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
//
def initOptions(Map args) {
def Map options = [:]
options.args = args.args ?: ''
options.args2 = args.args2 ?: ''
options.args3 = args.args3 ?: ''
options.publish_by_meta = args.publish_by_meta ?: []
options.publish_dir = args.publish_dir ?: ''
options.publish_files = args.publish_files
options.suffix = args.suffix ?: ''
return options
}
//
// Tidy up and join elements of a list to return a path string
//
def getPathFromList(path_list) {
def paths = path_list.findAll { item -> !item?.trim().isEmpty() } // Remove empty entries
paths = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
return paths.join('/')
}
//
// Function to save/publish module results
//
def saveFiles(Map args) {
def ioptions = initOptions(args.options)
def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
// Do not publish versions.yml unless running from pytest workflow
if (args.filename.equals('versions.yml') && !System.getenv("NF_CORE_MODULES_TEST")) {
return null
}
if (ioptions.publish_by_meta) {
def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
for (key in key_list) {
if (args.meta && key instanceof String) {
def path = key
if (args.meta.containsKey(key)) {
path = args.meta[key] instanceof Boolean ? "${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"
}
}

55
modules/mummer/main.nf Normal file
View file

@ -0,0 +1,55 @@
// Import generic module functions
include { initOptions; saveFiles; getSoftwareName; getProcessName } from './functions'
params.options = [:]
options = initOptions(params.options)
def VERSION = '3.23'
process MUMMER {
tag "$meta.id"
label 'process_low'
publishDir "${params.outdir}",
mode: params.publish_dir_mode,
saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), meta:meta, publish_by_meta:['id']) }
conda (params.enable_conda ? "bioconda::mummer=3.23" : null)
if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
container "https://depot.galaxyproject.org/singularity/mummer:3.23--pl5262h1b792b2_12"
} else {
container "quay.io/biocontainers/mummer:3.23--pl5262h1b792b2_12"
}
input:
tuple val(meta), path(ref), path(query)
output:
tuple val(meta), path("*.coords"), emit: coords
path "versions.yml" , emit: versions
script:
def prefix = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
def is_compressed_ref = ref.getName().endsWith(".gz") ? true : false
def fasta_name_ref = ref.getName().replace(".gz", "")
def is_compressed_query = query.getName().endsWith(".gz") ? true : false
def fasta_name_query = query.getName().replace(".gz", "")
"""
if [ "$is_compressed_ref" == "true" ]; then
gzip -c -d $ref > $fasta_name_ref
fi
if [ "$is_compressed_query" == "true" ]; then
gzip -c -d $query > $fasta_name_query
fi
mummer \\
$options.args \\
$fasta_name_ref \\
$fasta_name_query \\
> ${prefix}.coords
cat <<-END_VERSIONS > versions.yml
${getProcessName(task.process)}:
${getSoftwareName(task.process)}: \$( echo $VERSION )
END_VERSIONS
"""
}

48
modules/mummer/meta.yml Normal file
View file

@ -0,0 +1,48 @@
name: mummer
description: MUMmer is a system for rapidly aligning entire genomes
keywords:
- align
- genome
- fasta
tools:
- mummer:
description: MUMmer is a system for rapidly aligning entire genomes
homepage: http://mummer.sourceforge.net/
documentation: http://mummer.sourceforge.net/
tool_dev_url: http://mummer.sourceforge.net/
doi: https://doi.org/10.1186/gb-2004-5-2-r12
licence: ['The Artistic License']
input:
- meta:
type: map
description: |
Groovy Map containing sample information
e.g. [ id:'test', single_end:false ]
- ref:
type: file
description: FASTA file of the reference sequence
pattern: "*.{fasta,fasta.gz,fa,fa.gz,fna,fna.gz}"
- query:
type: file
description: FASTA file of the query sequence
pattern: "*.{fasta,fasta.gz,fa,fa.gz,fna,fna.gz}"
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"
- coords:
type: file
description: File containing coordinates of matches between reference and query sequence
pattern: "*.coords"
authors:
- "@mjcipriano"
- "@sateeshperi"

View file

@ -867,6 +867,10 @@ multiqc:
- modules/multiqc/**
- tests/modules/multiqc/**
mummer:
- modules/mummer/**
- tests/modules/mummer/**
muscle:
- modules/muscle/**
- tests/modules/muscle/**

View file

@ -0,0 +1,14 @@
#!/usr/bin/env nextflow
nextflow.enable.dsl = 2
include { MUMMER } from '../../../modules/mummer/main.nf' addParams( options: [:] )
workflow test_mummer {
input = [ [ id:'test', single_end:false ], // meta map
file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true),
file(params.test_data['sarscov2']['genome']['transcriptome_fasta'], checkIfExists: true) ]
MUMMER ( input )
}

View file

@ -0,0 +1,7 @@
- name: mummer test_mummer
command: nextflow run tests/modules/mummer -entry test_mummer -c tests/config/nextflow.config
tags:
- mummer
files:
- path: output/mummer/test.coords
md5sum: 6084fe43c7cb2eca8b96d674560bdefc