strandRule unnecessary/unused in process rseqc

[andreas-wilm]

Hello, looks like strandRule is computed unnecessarily or is at least unused in process rseqc

Best,
Andreas

[ewels]

Hah, correct. It looks like we used to use it for the RPKM_saturation.py command:

rnaseq/main.nf

Line 644 in 5d65867

RPKM_saturation.py -i $bam_rseqc -r $bed12 $strandRule -o ${bam_rseqc.baseName}.RPKM_saturation

We then later removed this call: f6f6669#diff-28b7bea57b04f15beeb4c1de0ee78382L725

I'm not totally sure whether this removal was intentional or not. The rest of the commit is where we first moved geneBody_coverage.py out into its own process, so we were definitely thinking about performance issues at the time. My guess is that I said that we should remove that command because it was super slow and performance-hogging and we already had very similar QC metrics from dupRadar.

Anyway, yes I think you're right that we can remove these lines without any effect:

rnaseq/main.nf

Lines 732 to 737 in 1cd5ab7

	def strandRule = ''
	if (forward_stranded && !unstranded){
	strandRule = params.singleEnd ? '-d ++,--' : '-d 1++,1--,2+-,2-+'
	} else if (reverse_stranded && !unstranded){
	strandRule = params.singleEnd ? '-d +-,-+' : '-d 1+-,1-+,2++,2--'
	}

Note that it won't change the way that the pipeline runs or is configured though, as the strandedness configuration is still used in other processes.

Phil

[ewels]

Oops, accidentally auto-closed the issue from my commit. Should wait for the PR to be merged really.

[apeltzer]

Done :-)