Estimating RNA Loop Entropies Using a New Nucleobase Model and Sequential Monte Carlo Method

Abstract

We report a new scheme that is designed to accurately and efficiently compute the entropy of RNA loops. The scheme is based on a new RNA nucleobase discrete state (RNAnbds) model and a Sequential Monte Carlo (SMC) method. The novelty of the RNAnbds model is that it directly represents the conformation of the RNA nucleobases, instead of the RNA backbones. To test the performance of this new scheme, we calculate the entropies for RNA hairpin loops and compare the results with the exact computational values obtained by an enumeration strategy and with the experimental data. It is found that the SMC method gives almost indistinguishable results from enumerations for short loops. For long hairpin loops, it also provides a good estimation that agrees with experiments.