Proper use of parameters

[mletexier-cnrgh]

Hello,

Can you confirm that when the --umi_read_structure parameter is used, the --skip-tools markduplicates and markduplicates_report parameters must also be added so that the gatk tool does not delete (or mark) additional duplicates without taking UMIs into account?

Sincerely
Mélanie

[tdanhorn]

Duplicates are not removed, just marked (but marked dups are skipped by variant callers). I haven't used UMIs with sarek, so I can't answer your question with confidence, but I would hope these parameters make no difference. You can just try it -- take a subsample, make sure there are some duplicates in there (you can just duplicate a read and change the UMI) and run the pipeline with and without --skip_tools markduplicates. You can skip BQSR and not specify any tools, so it would just give you CRAMs. Then run samtools view -cf 1024 /path/to/some.cram and see if you get more than 0 in any of them.
BTW, for questions about nf-core pipelines, the best place to ask is https://community.seqera.io/ or the nf-core Slack channel (in this case the #sarek channel). The GitHub is mostly read by the developers and used to track actual bugs, so your questions may get lost here.

[mletexier-cnrgh]

Thank you very much for taking the time to reply, I appreciate it.
Yes, I also hope that the tool doesn't add an extra layer with markup, but I have my doubts.
Thanks for the advice.

Mélanie