RNA Binding Protein Specific RNA Dinucleotide Conformers

Update

Including better resolution dataset (<=2.5 Å)

Summary

• We have utilised the non-redundant protein-RNA docking benchmark version 3.1 (PRDBv3.1), comprising 263 protein-RNA complexes from PDB with ≤35% sequence identity.
• Analysed 7,607 dinucleotides from these protein-RNA complexes using twelve geometric parameters to study RNA conformational states.
• Classified dinucleotides by interface location: 3,912 (52.42%) at interface and 3,695 (47.58%) non-interface.
• Also analysed 15,549 dinucleotides from 145 non-redundant RNA-only structures to provide a baseline for RNA geometric parameters.
• Applied Density-Based Spatial Clustering of Applications with Noise (DBSCAN) on backbone and side chain torsion angles to identify 29 unique RNA dinucleotide conformer clusters.

Cluster and Parameter Description

1. Cluster: DBSCAN-derived clusters representing unique dinucleotide conformations.
2. PDB ID: Source complex identifier from Protein Data Bank.
3. Chain: Chain containing the cluster representative.
4. NtC: Dinucleotide conformer type (new conformers highlighted).
   5-20. Various torsion angles (δ, ε, ζ, α, β, γ, χ) for the ith and (i+1)th nucleotides, base plane torsion (μ), Sugar Pucker, Glycosidic Bond rotation, secondary structure (SS), and interface status.

Transformed Confal Score (scaled 0–100)

[ \text{Confal Score} = \max\left(0, 100 - \sum_{i=1}^{n} \frac{|\theta_i - \mu_i|}{\sigma_i^2} \right) ]

• θᵢ: Observed torsion angle (0°–360°)
• μᵢ: Cluster mean torsion angle (0°–360°)
• σᵢ: Standard deviation in the reference conformer class
• n: Number of torsion angles considered

Keywords

protein-RNA recognition, torsion angle, RNA dinucleotide conformation, RNA conformational landscape, RNA conformers library.

Citation

Shri Kant, Savan Masipeddi, and Ranjit Prasad Bahadur. Deciphering conformational preferences of RNA in protein-RNA recognition. BioRxiv: https://doi.org/10.64898/2026.05.14.725147