AI · 8 min read · April 17, 2026
Machine Learning Maps Drug Binding to Viral RNA Pseudoknot
Spectral map analysis reveals how small-molecule inhibitors distort SARS-CoV-2 RNA structure in topology-dependent ways, with protonation state determining mechanism.
Machine learning identifies how antiviral drugs bind and distort SARS-CoV-2 RNA pseudoknot structure based on topology and chemical form.
- — Spectral map extracts slow conformational modes from molecular dynamics trajectories without manual feature engineering.
- — Merafloxacin destabilizes different RNA stem regions depending on whether pseudoknot adopts threaded or unthreaded fold.
- — Zwitterionic form of merafloxacin induces slow dynamics in unthreaded pseudoknot; neutral form does not.
- — Ligand protonation state at physiological pH fundamentally changes how drug molecules interact with RNA target.
- — Free-energy landscapes show ligand-induced distortion is selective by topology, not uniform across all conformations.
- — Study clarifies mechanism of -1 programmed ribosomal frameshifting inhibition, a viral protein synthesis pathway.
- — Thermodynamics-driven machine learning bridges gap between unbiased simulation and interpretable drug-binding mechanisms.
Frequently asked
- A pseudoknot is a three-dimensional RNA structure formed by two helical stems that fold back on themselves. In SARS-CoV-2, the pseudoknot controls programmed ribosomal frameshifting, a process the virus uses to synthesize proteins. By distorting the pseudoknot, small-molecule drugs can block this process and inhibit viral replication, making it an attractive drug target.