nf-core_modules/software/hisat2/build/main.nf
Harshil Patel d1c6082a66
Update modules required for rnaseq pipeline (#449)
* Update HISAT2 build module

* Bump preseq version

* Fix tests

* Add meta.yml for preseq to fix linting

* Auto-detect --genomeSAindexNbases for smaller genomes

* Add placeholder to use human data for the tests

* Add CSI output option to samtools/index

* Fix samtools/index tests
2021-04-16 08:56:47 +01:00

70 lines
2.4 KiB
Text

// Import generic module functions
include { initOptions; saveFiles; getSoftwareName } from './functions'
params.options = [:]
options = initOptions(params.options)
def VERSION = '2.2.0'
process HISAT2_BUILD {
tag "$fasta"
label 'process_high'
label 'process_high_memory'
publishDir "${params.outdir}",
mode: params.publish_dir_mode,
saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:'index', meta:[:], publish_by_meta:[]) }
conda (params.enable_conda ? "bioconda::hisat2=2.2.0" : null)
if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
container "https://depot.galaxyproject.org/singularity/hisat2:2.2.0--py37hfa133b6_4"
} else {
container "quay.io/biocontainers/hisat2:2.2.0--py37hfa133b6_4"
}
input:
path fasta
path gtf
path splicesites
output:
path "hisat2" , emit: index
path "*.version.txt", emit: version
script:
def avail_mem = 0
if (!task.memory) {
log.info "[HISAT2 index build] Available memory not known - defaulting to 0. Specify process memory requirements to change this."
} else {
log.info "[HISAT2 index build] Available memory: ${task.memory}"
avail_mem = task.memory.toGiga()
}
def ss = ''
def exon = ''
def extract_exons = ''
def hisat2_build_memory = params.hisat2_build_memory ? (params.hisat2_build_memory as nextflow.util.MemoryUnit).toGiga() : 0
if (avail_mem >= hisat2_build_memory) {
log.info "[HISAT2 index build] At least ${hisat2_build_memory} GB available, so using splice sites and exons to build HISAT2 index"
extract_exons = "hisat2_extract_exons.py $gtf > ${gtf.baseName}.exons.txt"
ss = "--ss $splicesites"
exon = "--exon ${gtf.baseName}.exons.txt"
} else {
log.info "[HISAT2 index build] Less than ${hisat2_build_memory} GB available, so NOT using splice sites and exons to build HISAT2 index."
log.info "[HISAT2 index build] Use --hisat2_build_memory [small number] to skip this check."
}
def software = getSoftwareName(task.process)
"""
mkdir hisat2
$extract_exons
hisat2-build \\
-p $task.cpus \\
$ss \\
$exon \\
$options.args \\
$fasta \\
hisat2/${fasta.baseName}
echo $VERSION > ${software}.version.txt
"""
}