Tuesday, June 1, 2010

RNA Helix Junctions

To understand RNA we need first to split it into two main regions, that is:

1) Single strand.
2) Double-strand helical regions.

This has been known for quite a while now, and so people have done quite a good job at describing double-stranded helical regions where there is canonical Watson-Crick base-pairing, that is, G pairs with C forming 3 hydrogen-bonds, and A pairs to U, forming 2 hydrogen-bonds. Now, the part that is always left a bit out of the picture, or which at least is less commonly looked at is the single stranded part of RNA.
Aalberts and Nandagopal from Williams College in Massachusetts are pointing out this problem, specifically in relation to secondary structure prediction algorithms.
They have refined mfold to take better account of hairpin loops and "multi-branch" loops. By including some type of self-avoiding freely-jointed chain model called FJC2 they give a polymer theory description of loops and claim to decrease the mfold error for multibranches 10 times.

Yet another paper which is also concerned with the interplay between single-stranded and helical regions is another one of the unstoppable paper making machine of the Czechs.

Besseova, Reblova, Leontis and Sponer publish in NAR ahead of print on May 27 a summary of MD simulation performed on three-way junctions picked from the ribosome. They characterize their main normal modes of motion in a fashion similar to what has been carried on by their group for other parts of the ribosome. Strangely they don't cite Christian Laing's latest papers on junctions, nor Dave Lilley's, not even on passing.

Here's a picture of one of their simulation snapshots: