View Single Post
Old 12-06-2010, 03:52 AM   #30
Senior Member
Location: Purdue University, West Lafayette, Indiana

Join Date: Aug 2008
Posts: 2,313

Hi Nate,
Currently we don't normalize when we construct libraries.

I take your point as to the utility of allowing the double-stranded RNA-DNA 1st strand to block internal priming of the GGG tail of the 2nd strand synthesis primer. (My guess would be that if the adapter is at a much higher molar concentration than the cDNA strands, that adapter concatamers would be more likely to form.) And, as I mentioned previously, we don't attempt to decouple 1st strand and 2nd strand synthesis to avoid adapter concatamers.

If you were keen to try decoupling out, but did not want to risk internal 2nd strand synthesis, you could remove the MMLV RT used by the SMART kit after 1st strand synthesis (eg, using phenol, etc.) Then add the 2nd strand primer + some other polymerase capable of displacing the RNA strand. I could even be a RT -- just one that doesn't have MMLV's tendency to add oligo C tails.

If we get enough total RNA and our polyA yield is high enough, we use the rapid cDNA method. If not, we use the SMART kit in the Metz/Colbourne/Mokaitis/Buchanan-Carter method. The latter has lots of tricks embedded in it to allow construction of libraries (including normalized libraries, if you are so inclined) optimized for 454 sequence generation with limited amounts of total RNA.

In principle, I see no reason the Rapid cDNA method could not be modified to normalize. Just use PCR (with the pA/pB primers) to generate enough cDNA to treat with a double-stranded nuclease.
pmiguel is offline   Reply With Quote