[husamia]

I have seen duplicate percentage of 33% in targeted high coverage experiment and based on my assessment of false positive and false negative I am happy not removing duplicates.

[Michael.James.Clark]

I have seen duplicate percentage of 33% in targeted high coverage experiment and based on my assessment of false positive and false negative I am happy not removing duplicates.

It depends on what your experiment is. In some cases it has very little effect, in some it can be detrimental. In no case that I've heard of does leaving PCR duplicates in your data improve your results, however.

[panos_ed]

It is possible to dedup before mapping. You may hash the first 14bp of each end and discard a pair if the 14+14bp coincides another pair. This method is not as good as deduping after mapping, but should be good enough. On the other hand, I do not think deduping is quite necessary for assembly.

Why 14bp? Why not something else?

[transforu]

Hi. everybody
I have a problem about remove duplicates,
In this study , http://www.ncbi.nlm.nih.gov/Traces/sra/?study=ERP000603
It's have 2 Experiments, and 13 RUNs,
How to do remove duplicates? by study? by Experiments? or by one runs?
Thanks !!

[sarvidsson]

Remove PCR duplicates by library (and this looks like one big "library" mixed from 4 PCR pools). Optical duplicates are removed by lane (but I guess Picard would detect the within lane reads, given proper read group labelling).