An Open Problem in RNA Secondary Structure Prediction by the Comparative Approach

One approach to predict the secondary structure of RNA is the comparative approach. This approach is used when alignment of several homologous sequences of a RNA is available. The idea is to underline covarying residues which maintain the Watson-Crick complementarity. The main problem of this approach is the selection of homologous sequences to use for the prediction. Indeed, there are many possible combinations of these sequences and only a few number give correct structure predictions. This is due to the low quality of alignment in helices or to variability of the sequences. To select homologous sequences which are correctly aligned, we use evolutionary models under structure constraints. Indeed, constraints on secondary structure should deviate the evolutionary model in helices regions and sequences with correct alignment in helices show this deviation. We have set criteria to underline this deviation and differentiate sequences with correct helices alignment from other sequences.