"Bug" in how input bams channel is formed

[soccin]

I am still new to all this nextflow/channel stuff but I think there is a bug here:

#LINE 44 of realignment.nf
ch_indexed_bams = normal_index.out.bam.join(tumor_index.out.bam)

should be

ch_indexed_bams = tumor_index.out.bam.join(normal_index.out.bam)

because in the gatk_bqsr the first input tuple read:

tuple val(meta), path(tumor_bam), path(tumor_bam_index), path(normal_bam), path(normal_bam_index)

tumor first, normal second. I do not think it matters though for this step as those two bams are treated equally but it is confusing.

https://github.com/mskcc/Arrakis/blob/a13d58c4cbd1f332759b04ef1717753f57fc2181/subworkflows/local/reallignment.nf#L44