Assessment Of Tertiary Structure And Statistical Analysis Of Torsion Angles Of Non-coding RNA

The tertiary structure of RNA is the basis of its biological function, so if we want to know the Regulatory functions of RNA molecules, especially non-coding RNAs, we must get their refined structures firstly. Consequently, more and more researchers turn their attentions to building or predicting RNA tertiary structures computationally. The successful RNA structure prediction methods not only need an effective algorithm for searching the conformational space to generate a large number of RNA models, but also a reasonable and sensitive scoring function to identify the appropriate tertiary structures from them. Recent years, the studied results show that the RNA structure prediction methods can get a lot of candidate RNA structures, however, the scoring function could not successfully select the near native conformation state from them. Thus, we proposed a new knowledge-based statistical potential for RNA tertiary structure evaluation in this paper. Further, in order to understand the structural features of the RNA molecule in detail, we analyzed the distribution of the dihedral angles of the RNA molecules in the PDB database.The main content of this article is summarized as follows:(1) This paper proposed a statistical potential based on the atom-shuffled reference state (3dRNAscoring) for the assessment of RNA tertiary structures. In different benchmarks, the success rate of3dRNAscoring method is5%higher than the RASP method and7.5%higher than KB method. For FARFAR test set with non-canonical base pair structures,3dNA scoring can pick the lowest RMSD structure more effectively than other existing scoring methods.(2) This paper also statistically analyzed the angular distribution of the RNA molecules in the PDB database. We counted all RNA monomers in the PDB database, as well as analyzed the similarities and differences of dihedral angle distribution for different kinds of RNAs, which is classified according to their functions. The in-depth understanding of the torsion angle distribution of RNA molecules may contribute to the RNA structure prediction and may improve the accuracy of the prediction.