Exploring Secondary Structure Predictions for RNA-Targeted Drug Discovery: Power and Challenges

Area of Science:

• Biochemistry
• Molecular Biology
• Drug Discovery

Background:

• Ribonucleic acids (RNAs) are crucial regulators in biological processes and emerging drug targets.
• A gap exists between known RNA sequences and solved structures, impeding RNA-targeted drug discovery.
• RNA secondary structure prediction can identify potential druggable sites on RNA molecules.

Purpose of the Study:

• To benchmark RNA secondary structure prediction tools.
• To assess tool performance on a curated dataset of ligand-bound RNA structures.
• To evaluate the accuracy of predicting RNA secondary structures, especially within ligand binding sites.

Main Methods:

• Curated a dataset of ligand-bound RNA structures.
• Benchmarked widely used RNA secondary structure prediction tools.
• Assessed prediction accuracy for varying RNA lengths, motifs, and ligand binding sites.

Main Results:

• Most tools perform well on short, simple RNAs but decline for longer RNAs and pseudoknots.
• Prediction accuracy is significantly reduced within ligand binding sites.
• Noncanonical base pairs and complex structures in binding sites are poorly recognized.
• RNA ligand binding sites are inadequately reconstructed by current secondary structure predictions.

Conclusions:

• Existing RNA secondary structure prediction tools have limitations for RNAs involved in ligand binding.
• The accuracy gap hinders the integration of these tools into RNA-targeted drug discovery pipelines.
• Further development is needed to improve prediction accuracy for complex RNA structures and binding sites.