ribominus vs DSN

[daughart]

Has anyone used the DSN method (available on Illumina website - Duplex-specific thermostable nuclease)? I assume since it degrades highly-abundant transcripts, that it will prevent accurate quantitation, but I'm not sure to what extent. I have used ribominus in the past and it sucks (long protocol, not very efficient, low efficiency).

[kmcarr]

I don't have the data handy but according to Illumina it does not significant effect abundance of mRNA species. They have a plot comparing mRNA hits quantitated from a standard library prep (polyA selection) and a DSN library prep. They observed only a slight change for the most abundant transcripts.

The theory is that the conditions are tuned to hit the re-annealing curve before any mRNA derived cDNAs have had sufficient time to form duplexes.

Update:

Found what I was looking for. Slides 35-46 discuss the DSN normalization and slide #44 specifically addressed your question.

[ScottC]

The other main benefit being that you can analyse degraded RNA samples because you don't have to use a poly-A selection step (they discuss using RNA extracted from FFPE tissue).

[steven]

I don't have the data handy but according to Illumina it does not significant effect abundance of mRNA species. They have a plot comparing mRNA hits quantitated from a standard library prep (polyA selection) and a DSN library prep. They observed only a slight change for the most abundant transcripts.

The theory is that the conditions are tuned to hit the re-annealing curve before any mRNA derived cDNAs have had sufficient time to form duplexes.

Update:

Found what I was looking for. Slides 35-46 discuss the DSN normalization and slide #44 specifically addressed your question.

Thanks kmcarr for providing this document. I had seen this slide too, but now i realize that on the y-axis there is still a polyA selection going on. I am curious about what would give polyA (-DSN) vs. total (+DSN).
The following slide (#45) may show this, but does the "Total RNA" really include DSN? If so, why is the effect on the most abundant transcripts the opposite of what is on the previous slide?

As you said, the assumption is that only the most abundant transcripts are affected, which is not a problem anyway: who cares about housekeeping genes? I am still wondering if this would not affect a quartile-based normalization method for instance.