Here is the natural continuation of Brourd’s 1-1 loop strategy. I think Brourd’s idea explains why the GU/UG puzzle trick for the 2-2 loop, don’t work in lab.
G and U are attracted, if opposite to each other. Even if in a 2-2 loop. Which means attraction between structures, we want to stay apart.
This is Wisdaves design WisTest from Shape test lab:
On the shapedata you can see, that the 2-2 loop joins basepairs, as there is no yellow nucleotides in the loop ring.
So far none of the GU/UG 2-2 loops have worked.
The Eterna energy model don’t allow for GC-pairs to be opposite each other in the 2-2 loop, so those we don’t have to worry about. The same appear to be the case with AU-pairs across each other.
I would like a strategy that says:
Give -1 pr. each GU pair opposite each other in a 2-2 multiloop.
So for the first picture shown, give all in all -2.