TriRNASP: A knowledge-based potential with three-body effects for accurate RNA structure evaluation.

Area of Science:

• Computational Biology
• Structural Biology
• Bioinformatics

Background:

• Accurate RNA 3D structure evaluation is crucial for predicting RNA structures.
• Existing knowledge-based potentials and scoring functions have limited performance on challenging datasets from RNA 3D structure prediction methods.

Purpose of the Study:

• To develop TriRNASP, an efficient statistical potential incorporating three-body effects for enhanced RNA 3D structure evaluation.
• To improve the accuracy and efficiency of RNA 3D structure assessment.

Main Methods:

• Developed TriRNASP, integrating coarse-grained three-body correlations with Kullback-Leibler divergence and an atom clash penalty.
• Benchmarked TriRNASP against state-of-the-art statistical potentials and deep-learning scoring functions on diverse RNA 3D prediction datasets.

Main Results:

• TriRNASP demonstrated superior performance in identifying native and near-native RNA structures across extensive test datasets.
• Performance gains were particularly notable on recent CASP15 and CASP16 assessment datasets.
• TriRNASP achieved exceptional computational efficiency, suitable for large-scale evaluations.

Conclusions:

• TriRNASP offers a significant advancement in RNA 3D structure evaluation, outperforming current methods.
• Its efficiency and accuracy make it a valuable tool for RNA structure prediction and analysis.
• The TriRNASP potential is publicly available for research use.