About UMI impact on CNVkit results

[tetedange13]

Hi,

Regarding CNVkit's documentation about marking read duplicates or not (and also this discussion), I was wondering:
=> Is it correct enough to run CNVkit on BAM files that are UMI-deduplicated (+ a bit error-corrected in the same process) ?
=> In my particular case, I work on a hotspots panel of FFPE samples, obtained through hybrid-capture and I use CNVkit with a flat reference (+ HAAR segmentation most of the time)

Thanks for your answer.
Kind regards.
Felix.