RBscore

RBscore predicts RNA and DNA binding sites on protein structures, providing residue-level probabilistic assessments of nucleic-acid binding to support analysis of protein–nucleic acid interactions.

Key Features:

• Prediction accuracy and stability: Evaluated against 18 web servers and three standalone programs across 41 diverse datasets, demonstrating high accuracy and consistent performance in predicting nucleic acid-binding sites on proteins.
• Scoring scheme: Employs a scoring scheme that directly correlates feature values with probabilities of RNA/DNA binding and maps the binding energy landscape across the protein surface as an energy funnel.
• Visualization capabilities: Displays prediction scores and relevant features on protein structures to interpret how specific residues contribute to nucleic acid interactions.
• Benchmark database (NBench): Provides NBench to address biases in dataset selection, binding site definitions, and assessment metrics and to benchmark and validate nucleic acid-binding predictions across computational tools.

Scientific Applications:

• Mapping protein–nucleic acid interactions: Identifies residues likely to interact with RNA or DNA to support studies of gene regulation, RNA processing, and DNA replication.
• Mechanistic studies: Aids elucidation of molecular mechanisms underlying protein–nucleic acid interactions at the residue level.
• Drug discovery: Facilitates discovery efforts targeting nucleic-acid-binding proteins by pinpointing candidate binding residues.
• Method benchmarking: Uses NBench for benchmarking and validation of computational predictors of nucleic acid-binding sites.

Methodology:

Performs detailed analysis of protein features indicative of RNA/DNA binding, integrates structural data with predictive algorithms, and applies a scoring scheme that maps feature values to binding probabilities and an energy-funnel visualization for residue-level probabilistic assessment.