tron-bioinformatics / tronflow-bam-preprocessing Goto Github PK

Do something like this

process DUPLICATION_RATE {
    cpus params.cpus
    memory params.memory
    tag "${name}"
    //publishDir "${params.output}/${name}/metrics/coverage", mode: "copy"
    //publishDir "${params.output}/${name}/", mode: "copy", pattern: "software_versions.*"

    conda (params.enable_conda ? "bioconda::samtools=1.15.1" : null)

    input:
        tuple val(name), file(bam), file(bai)

    output:
        file("${name}.duplication_rate.tsv")
        
    script:
    """
    samtools flagstat -@ ${task.cpus} ${bam} > output.txt
    parse_duplication_rate.py output.txt 
    """
}

Skipping the metrics step results in an error:

"Argument METRICS_FILE was missing: Argument 'METRICS_FILE' is required"

I would suggest to add a new flag, with which you can select the tool used for marking or removing duplicate reads. The choices would be then

1. picard (in particular MarkDuplicates) or
2. sambamba (in particular markdup).

With this feature, the pipeline remains backwards compatible to users who want to run Picard's MarkDuplicates for any reason, but who similarily want new updates or features introduced in later versions than v1.9.1.

The depth of coverage file contains the vertical coverage at 1 bp resolution. This file can become large. Also, unindexed it is difficult to use.