Using Machine Learning to Predict the Antibacterial Activity of Ruthenium Complexes**
Problem
Antimicrobial resistance and a stalled antibiotic pipeline make new antimicrobial compounds urgently needed; metal complexes are promising but understudied.
Why it matters
MRSA and resistant pathogens threaten effective treatment of infections worldwide.
What the AI did
Machine learning models trained on 288 synthesized ruthenium arene Schiff-base complexes predicted antibacterial activity, guiding selection of 54 new compounds that were then tested.
Results
Predicted 54 new compounds showed 5.7x higher hit-rate (53.7%) against MRSA versus 9.4% for the original 288-compound library.
Impact
Experimentally validated ML-guided screening substantially improved the antibacterial hit-rate for metal complexes.
Record
| Venue | Angewandte Chemie International Edition |
|---|---|
| Published | 2023 |
| Work type | article |
| Authors | Markus Orsi, Boon Shing Loh, Cheng Weng, Wee Han Ang, Angelo Frei |
| Citations | 44 (OpenAlex, at import time) |
| References | 32 |
| Model or method | not stated in source (a series of ML models) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1002/anie.202317901 |
| Open access | bronze |
| Keywords | machine learning, ruthenium complexes, antibacterial, MRSA, drug discovery |
Importance rating
major — millions of lives, an industry or a field
Abstract frames rising AMR and stalled antibiotic pipeline as necessitating novel drug discovery; MRSA is named target. Major global health problem, though no mortality/scale figures given.