Distinguish technical duplicates from biological duplicates

Question

[BioStar22]

Hey Biostars, I'm asking for your expertise: I am handling 90x WGS data (with PCR) and I observe many different reads starting exactly at the same positions in the genome. Interestingly, the reads show different variants. Can these variants be true (=biological duplicates) or is to assume that they are false positives (=technical duplicates)?

In other words: how likely is the same start and end position of reads from different cells? Is it common? Seems unlikely to me.

Any opinions on this will be appreciated!! Thanks!

[ATpoint]

It is common practice to remove duplicate reads after alignment, did you do that?

[BioStar22]

Thanks, yes I used gatk MarkDuplicates I think it doesnt recognize those as duplicates, as they still show up when I choose not to show duplicates in IGV. Do you think something went wrong with this in my case?