I am trying to understand PCR duplicates in NGS analyses (actually whole-genome). I searched, and the best answer I found is in this blog.
However I don't understand if I understood how PCR duplicates arise correctly because I cannot see the problem of having them in the downstream analysis - aside from computational problems, i.e. unnecessary redundancy.
If I understood correctly, PCR duplicates arise during library preparation when PCR amplifies the fragments with adapters. In this case, if you have duplicate fragments, you'll end up amplifying some of them twice or more.
However, when you are doing the mapping, they should fall on the same region, most likely decreasing the quality of the mapping (since they increase the consensus on a specific sequence, which could have been subject to sequencing errors). But aside from that, you should have the same mapping with respect to the mapping where duplicates are removed, although with decreased quality.
Is the quality problem the actual reason for removing PCR duplicates or is there anything that I'm missing?