The reason I ask is that I did a bake-off between same samples same library prep but with and without UMI and I have a very large difference in total sequences before deduplication. Has anyone run into a similar issue?