PRBR

PRBR predicts RNA-binding residues from protein amino-acid sequences using hybrid feature representations and an enriched random forest classifier.

Key Features:

• Hybrid feature set: Combines predicted secondary-structure descriptors, evolutionary/conservation signals linked to amino-acid physicochemical properties, and dependency features capturing polarity–charge and hydrophobicity relationships along the protein sequence.
• Enriched random forest (ERF) algorithm: Uses an ERF classifier to model non-linear interactions among features and improve robustness on complex sequence datasets for residue-level RNA-binding prediction.

Scientific Applications:

• Protein function analysis: Mapping putative RNA-interacting residues aids interpretation of protein roles and interaction mechanisms.
• Post-transcriptional regulation studies: Characterization of RNA–protein interfaces informs models of RNA processing, stability, transport, and translation control.
• Drug design and development: Localization of RNA-binding sites can guide design of molecules that modulate RNA–protein interactions for therapeutic purposes.

Methodology:

Analyzes amino-acid sequences using a composite feature design that couples predicted secondary-structure descriptors with evolutionary conservation of amino-acid physicochemical properties and dependency measures (polarity–charge and hydrophobicity), and applies an enriched random forest classifier to model non-linear feature interactions for residue-level RNA-binding prediction.