gpertea / stringtie

Transcript assembly and quantification for RNA-Seq
MIT License
361 stars 76 forks source link

Non-canonical splice sites in long-read data #422

Open wenzelm opened 2 months ago

wenzelm commented 2 months ago

Hello,

I was wondering if you could clarify how best to deal with non-canonical splice sites when assembling long-read ONT data, please. My organism uses a wide range of splice sites, both canonical and non-canonical. I've aligned my ONT reads with minimap2 without penalizing non-canonical splice sites:

minimap2 -a -x splice -un -C0 --junc-bed

I am unsure whether I should be using the --viral option in stringtie2 or not. Does the default mode ignore non-canonical splice sites altogether? Or, in other words, is the --viral option designed for cases where all splice sites are expected to be non canonical? I would be grateful for clarification on how the --viral option operates, please.

Thank you very much!

Best wishes, Marius