AnnapuRNA

AnnapuRNA evaluates RNA–small molecule complex structures using a knowledge-based scoring function and machine learning trained on statistical data from experimentally determined RNA–ligand complexes.

Key Features:

• Knowledge-Based Scoring Function: Employs a scoring function derived from statistical data of experimentally determined RNA–small molecule complexes to assess the quality of RNA–ligand structures generated by any computational docking method.
• Machine Learning Integration: Uses machine learning models trained on interaction statistics from known RNA–ligand complexes to distinguish naturally occurring interactions from theoretically predicted ones and improve prediction accuracy.
• Comprehensive Evaluation Factors: Accounts for the ligand starting conformer, the docking program used, and the scoring function applied when evaluating predicted RNA–ligand structures.

Scientific Applications:

• RNA-Targeted Drug Discovery: Facilitates in silico identification and ranking of small molecules for RNA targets to support RNA-targeted drug discovery efforts.
• FMN Riboswitch Analysis: Demonstrated in a post-hoc study of FMN riboswitch structures to predict preferred ligands and elucidate RNA–small molecule interaction mechanisms.

Methodology:

Collects and analyzes statistical data from experimentally validated RNA–ligand complexes, trains machine learning models on these statistics, and applies the resulting knowledge-based scoring function to evaluate docking-generated RNA–ligand structures.