High levels of duplicated reads Illumina from PCR-free libraries

Question

[grey]

Checking fastqc results from HiSeq 4000 run of PCR-free libraries and came across really high Sequence Duplication Levels.

Note: We were trying to sequence at high depth to detect somatic mutations (~240X).

Can't be PCR duplicates since these are PCR-free so it's a bit mysterious to me, though others might have ideas. Optical duplicates? Insufficient DNA in the sample? Should we be ok just removing duplicates or does this indicate something systemically wrong?

[GenoMax]

You can use clumpify.sh from BBMap suite to check on types/numbers of duplicates you have without doing alignments : Introducing Clumpify: Create 30% Smaller, Faster Gzipped Fastq Files. And remove duplicates.

While it is possible that there are some cluster/optical duplicates it may also simply be the characteristic of this library prep. I am not sure if your kit used tagmentation so there is a possibility that similar fragments were generated,