Fig. 1: Nucleobase-centric statistical potential and sampling technique for RNA structure refinement.
A Six-dimensional base–base statistical potentials with relative positions defined by rij and rji vectors along with the rotational angle ωij. B A schematic illustration of the orientation dependence before and after Quantum Mechanical (QM) energy scaling as labeled. C This QM scaling is based on the correlation between the statistical energy scores of hydrogen-bonded base pairs (-ln Phb) and QM calculations. Phb is the probability of hydrogen-bonded base pairs. The QM energies for a few base pairs were illustrated in the insert. D The distribution of OP (red dots) around nucleobase C and the distribution of O4’(red dots) around the nucleobase A as labeled. E Backbone rotamers defined according to various torsion and improper angles that control the ribose (χ, ν) and phosphate (ε, ζ) backbone. Dihedral angles α, β, and γ, bond angles θPOC, θOCC and bond length of C5’O5’ were required to calculate the internal energy. F Nucleobase-centric fold-tree (NuTree) algorithm for refinement by defining bases as the nodes and locally or globally connected bases as the edges, illustrated by the GCAA tetraloop with the canonical and noncanonical base pairs shown in blue and green colors, respectively.
