Using sequence signatures and kink-turn motifs in knowledge-based statistical potentials for RNA structure prediction

Area of Science:

• Computational Biology
• Structural Biology
• RNA Biology

Background:

• Kink turns (k-turns) are prevalent RNA motifs in internal loops, essential for biological processes like translation, regulation, and splicing.
• These motifs exhibit conserved sequence patterns (3-nt bulge, G-A/A-G base pairs) and unique helical bending (∼50°) due to A-minor interactions.

Purpose of the Study:

• To develop computational folding rules for predicting k-turn RNA topologies based on sequence and geometrical features.
• To create and validate knowledge-based potentials for RNA structure prediction, specifically incorporating k-turn motifs.

Main Methods:

• Annotation of k-turn motifs in an RNA dataset using sequence and geometric signatures.
• Analysis of bending and torsion angles to define k-turn specific potentials.
• Application of updated scoring potentials within the RAGTOP (RNA-As-Graph-Topologies) graph sampling protocol for RNA structure prediction.

Main Results:

• Successfully identified and analyzed k-turn motifs in 35 RNA samples (12 k-turn, 23 non-k-turn).
• RAGTOP protocol with k-turn potentials showed significant improvements in predicting RNA structures compared to k-turn-free potentials.
• Demonstrated the utility of sequence and geometric features for motif-based RNA structure prediction.

Conclusions:

• Kink turns serve as a model for incorporating sequence-structure motifs into RNA structure prediction algorithms.
• The developed computational approach enhances the accuracy of RNA structure prediction and design.
• This strategy can be extended to other RNA structural motifs with distinct sequence and geometric properties.