RNAcommender

RNAcommender predicts RNA targets for RNA-binding proteins (RBPs) using a recommender system that propagates known RBP–RNA interaction data and integrates protein and RNA structural features.

Key Features:

• RBP–RNA Interaction Prediction: Recommends candidate RNA targets for RBPs with limited or unknown interaction profiles based on existing interaction datasets.
• Feature Integration: Incorporates protein domain composition and predicted RNA secondary structure to improve prediction accuracy.
• Benchmark-Validated Performance: Achieves an average AUC ROC of 0.75 and significant enrichment of correct recommendations for 75% of evaluated human RBPs.

Scientific Applications:

• RNA Regulatory Network Analysis: Identifies candidate RBP–RNA interactions to support functional annotation of RBPs and prioritize experimental validation.

Methodology:

RNAcommender applies a recommender system framework to propagate known human RBP–RNA interaction data, integrating protein domain architecture and predicted RNA secondary structure features to generate interaction predictions evaluated by ROC analysis.