CNVKit on targeted amplicons with UMIs

Question

[Sophie]

I understand that if running CNVkit with targeted amplicon data, the BAM file prior to duplication marking/removal should be used. However, my sequencing data has UMIs which are shown to be better than duplicate marking/removal. I'm running UMI-tools to de-duplicate - is this BAM file suitable for use with CNVKit or should I use the BAM prior to that?

Answer 1

[Matthias Zepper]

In my opinion, you can safely use the files that you have deduplicated with UMI-tools.

Indeed, targeted amplicon sequencing data should not be deduplicated with traditional duplication callers based on sequence identity respectively mapping-position alone, but the UMIs allow distinguishing biological from technical duplicates. See the method description of UMI-tools for details.

Answer 2

[Eric T.]

Yes, I agree UMI-tools is a good approach here.