Add option for HTSeq-count

[olgabot]

Hello!
We use HTSeq-count for quantifying gene expression, as featureCounts' combined exon model has been shown to not be very good (ref: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0141910). Is anyone else interested in this feature?

I noticed the feature to run Salmon/Kallisto (#171), which would be welcomed as well!

Warmest,
Olga

[ewels]

Also related: #70 (add RSEM).

Could definitely add the two new tools and have an option to choose which one is used.. 👍

[ewels]

Note that we used to use HTSeq in our previous pipeline way back when. We switched to using featureCounts because it was significantly faster and less computationally intensive. I'm not sure if this is still the case though.

[olgabot]

I think the reason for HTSeq was that the meta-exon model of featurecounts was inaccurate compared to the summation of transcripts model by HTSeq (citation: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0141910)

I've implemented HTSeq-count here and could add it to nf-core/rnaseq as well.