Allow Salmon to infer strandedness

[kathleencheung]

It would be great if there was an option to let Salmon automatically infer the library type instead of taking this from the input file.

E.g. --libType A

This is already inferred by RSeQC infer_experiment.py and the user can re-run the pipeline with the correct strandedness if it was incorrect, however, allowing Salmon to infer it will save this extra step when the user does not know the strandedness.

[drpatelh]

Hi @kathleencheung! Apologies for the late response. This seems like a decent proposition as is described here. We could possibly add a parameter to the pipeline such as --salmon_quant_libtype that allows users to override the default behaviour of the pipeline.

I am a bit hesitant to make --libType A the default because getting the strandedness correct will affect the results generated by other QC-based tools used in the pipeline too and this sort of thing can be easily overlooked and the wrong inferences will be made without re-running the pipeline with the correct strandedness e.g. qualimap, stringtie and featurecounts

How does that sound?

[kathleencheung]

Thanks @drpatelh that sounds great!

[drpatelh]

FIxed in #659 - using --salmon_quant_libtype A will now be able to give you this behaviour and in fact the ability to override any valid --libType setting. Just beware that some of the downstream QC metrics from other tools may not be entirely accurate if there is a discrepancy in the strandedness provided in the samplesheet compared to that auto-inferred by salmon quant.