RBscore&NBench: a high-level web server for nucleic acid binding residues prediction with a large-scale benchmarking...

Area of Science:

• Structural Biology
• Bioinformatics
• Computational Biology

Background:

• Identifying protein residues involved in RNA and DNA binding is crucial for understanding molecular mechanisms.
• Existing prediction tools often suffer from biases related to datasets, binding site definitions, and assessment metrics.

Purpose of the Study:

• To develop and validate a robust method for predicting RNA/DNA-binding residues in proteins.
• To create a comprehensive benchmark database for evaluating prediction methods.

Main Methods:

• RBscore utilizes a scoring scheme linking feature values to nucleic acid binding probabilities.
• Predictions are visualized on protein structures, illustrating binding energy landscapes.
• RBscore was rigorously compared against 18 web servers and 3 standalone programs across 41 diverse datasets.

Main Results:

• RBscore demonstrated high and stable accuracy in predicting RNA/DNA-binding residues.
• The comprehensive comparison highlighted the limitations of existing methods and the need for standardized benchmarks.
• NBench was developed as a benchmark database to mitigate biases in method evaluation.

Conclusions:

• RBscore provides a reliable and accurate tool for predicting nucleic acid-binding residues.
• The NBench database serves as a valuable resource for the community to assess prediction tools objectively.
• The integrated RBscore & NBench web server facilitates research in nucleic acid-protein interactions.