Open AccessVfold2D-MC: A Physics-Based Hybrid Model for Predicting RNA Secondary Structure Folding
Author(s) -
Yi-Sung Cheng,
Sicheng Zhang,
Xiaojun Xu,
Shijie Chen
Publication year - 2021
Publication title -
the journal of physical chemistry. b
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.864
H-Index - 392
eISSN - 1520-6106
pISSN - 1520-5207
DOI - 10.1021/acs.jpcb.1c04731
Subject(s) - rna , intramolecular force , monte carlo method , statistical physics , folding (dsp implementation) , nucleic acid secondary structure , physics , algorithm , chemistry , biological system , mathematics , biology , engineering , statistics , quantum mechanics , biochemistry , electrical engineering , gene
Accurate prediction of RNA structure and folding stability has a far-reaching impact on our understanding of RNA functions. Here we develop Vfold2D-MC, a new physics-based model, to predict RNA structure and folding thermodynamics from the sequence. The model employs virtual bond-based coarse-graining of RNA backbone conformation and generates RNA conformations through Monte Carlo sampling of the bond angles and torsional angles of the virtual bonds. Using a coarse-grained statistical potential derived from the known structures, we assign each conformation with a statistical weight. The weighted average over the conformational ensemble gives the entropy and free energy parameters for the hairpin, bulge, and internal loops, and multiway junctions. From the thermodynamic parameters, we predict RNA structures, melting curves, and structural changes from the sequence. Theory-experiment comparisons indicate that Vfold2D-MC not only gives improved structure predictions but also enables the interpretation of thermodynamic results for different RNA structures, including multibranched junctions. This new model sets a promising framework to treat more complicated RNA structures, such as pseudoknotted and intramolecular kissing loops, for which experimental thermodynamic parameters are often unavailable.