[Strategy market] Reduce Mismatch using CUAG

I propose a CUAG strategy to help design stacks that minimize potential for mismatch. The strategy is motivated by looking at the free energy of stack “cells” formed by adjacent bonding pairs and the potential of certain nucleotide sequences to miss-pair.

GU bonds are weak and designs may be stronger if we avoid them (or at least use them very carefully and with close attention). Adjacent GG, GU, UG and UU have lots of potential for mismatch, so we try to avoid them by using a CUAG ring mnemonic for designing stacks. With no UUs in the design, AA can fairly safely be used to fill rings. With no UG and GU in the design, AC and CA can more safely be used to skip or add a nucleotide at a bulge. CC should be avoided.

The CUAG mnenomic is summarized by this image:

The details can be found in a Google Document here.

Note: I’m still thinking on how to “score” designs that fail to follow this strategy. Suggestions welcome.

I finally got what I think is a decent switch lab design using this strategy: CUAG the first.

It turns out that the lab constraints for the FMN2 lab require at least one exception to the CUAG rules. (Good thing I added the section on “Rules Were Made to be Broken”.) The synopsis is that the fixed pair at 20-34 combined with overlapping stack constraints would force 22 and 23 to form one of GG, GU, UG or GG. While I may try a design that goes with that at some point I instead made 21-22 a UG sequence and used an AC-GAU bulge. I like the fact that both target shapes show up fairly clearly in the dot plot as per hoglahoo’s idea.

So the rules can be used (with small exceptions) to design a lab candidate. Perhaps at some point one or more CUAG designs will get synthesized an we will get a chance to see if it helps.