removing duplicates in small (si)RNA data

[Kennels]

Hello,

I hopefully have found most of the threads regarding duplicate removal in this forum.
(e.g. http://seqanswers.com/forums/showthr...uplicate+reads)

I am wondering if anyone has some input regarding this topic on small RNA data sets - how large an effect (if any) does real PCR duplicates have on real sRNA reads?
An sRNA read distribution profile typically displays many reads originating from an exact 5' position, which can be observed as 'hotspots' of read stacks across a reference.

I am analyzing miRNA/siRNA distribution profiles on viral genomes and hpRNAs, so the read depth is typically very high in very localized hotspots, but I am wondering if duplicates should be considered in my analysis, and if so, how does one distinguish real from duplicates reads?

thanks

[ron128]

Hello. There is no real way to differentiate between pcr duplicates and "real" reads. The only way out is if you use tags in the sample preparation. We had tried something similar to what kcarr has suggested, and were able to differentiate between the two.

Duplicates IMHO would not matter much in your case unless you are studying rnaseq data or analysing expression profiles.