Researchers have utilized protein language models (PLMs) and ensemble learning to identify over 46,000 potential beta-lactamases from environmental metagenomes. These identified enzymes, distributed across various habitats and belonging to all four Ambler classes, confer resistance to multiple beta-lactam antibiotics when expressed in E. coli. Further analysis using inverse folding models revealed specific mutations that can alter resistance profiles, including enhancing resistance to carbapenems, and indicated that highly fit sites show greater natural sequence variability, particularly in wastewater-derived enzymes compared to those from the ocean. This work highlights the extensive, underappreciated diversity and mutational adaptability of beta-lactamases in environmental settings. AI
IMPACT This research demonstrates the power of AI in discovering novel biological entities and understanding mechanisms of antibiotic resistance.
RANK_REASON The cluster contains a scientific paper detailing a new research methodology and findings in bioinformatics. [lever_c_demoted from research: ic=1 ai=1.0]
- Ambler classes
- arXiv
- beta-lactamase
- carbapenem antibiotic
- Escherichia coli
- Ocean
- Protein language models
- wastewater
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
