Hi all, first of all thanks for the great pipeline you provide here. I have a conceptual question. Just got libraries where the RNASeq part has UMIs embedded in the reads while the RiboSeq part doesn't have UMIs. Is this something the pipeline can deal with or would I have to extract UMIs first, add them to the read name and add a pseudo-UMI to the riboseq names that is just the same for all reads and then run with the

--with_umi --skip_umi_extract --umitools_umi_separator ":"

parameters?

For the RNAseq, there is the #CE8SSJV3N pipeline. If the UMI options that come with that don't meet your needs, have a look at #C0453Q2SFCM to preprocess your reads, it has more options AFAIK. Don't know how reliable RiboSeq will be without UMIs, but you can run #C05785S3TMX with

--with_umi false

.

Ok, that's what I thought, thanks for your feedback!

Technically the riboseq pipeline should process both rnaseq and riboseq, and ideally you'd run these together if you want TE measurements etc. The question you are asking is legitimate, atm umi options are applied globally regardless of dataset type. So indeed if you have a mismatch in how the libraries for the two modalities are made, currently this isn't supported. I am not sure how common this situation is to require implementing a different configuration in the pipeline vs using a workaround

I think if you use #C0453Q2SFCM to preprocess the data set with UMI (and get a consensus), then you could run all of your data together through #C05785S3TMX with

--with_umi false

.

Thank you guys for your help, I'll try my luck with #C0453Q2SFCM. Idk if it's worth the effort, but decoupling the UMI processing for RNA and riboseq might help other people as well. I'll let you be the judge on that. If I find another solution I'll let you know.