Artificial intelligence develops novel antibiotics aimed at combating drug-resistant bacteria strains
AI-Designed Antibiotics Offer Hope in the Battle Against Superbugs
In a groundbreaking development, researchers at the Massachusetts Institute of Technology (MIT) have created two potential new antibiotics using artificial intelligence (AI). These new compounds, named NG1 and DN1, have shown promising results in laboratory settings and on infected mice, according to a study published in Cell.
Professor James Collins from MIT stated that AI can expand the arsenal against superbugs, which are a major global threat, causing nearly five million deaths annually worldwide. Antibiotic resistance has evolved due to overuse, with the last major class of antibiotics being discovered in the 1980s.
The researchers used generative AI platforms, combining genetic algorithms and variational autoencoders, to generate millions of theoretical molecular structures. This approach allowed them to explore vast, previously uncharted regions of chemical space. The process started by screening a vast number of chemical fragments in silico to identify promising candidates against target pathogens.
Two lead molecules, NG1 and DN1, have shown potent effects against Neisseria gonorrhoeae and methicillin-resistant Staphylococcus aureus (MRSA), respectively. These compounds work through distinct mechanisms, disrupting bacterial cell membranes in ways that differ from existing antibiotics. The newly designed antibiotics attack bacteria in entirely new ways, which is crucial for overcoming drug resistance.
Of 24 synthesized compounds, seven showed selective antibacterial activity. Aarti Krishnan, one of the study's lead authors, mentioned the intention to avoid creating anything that resembles existing antibiotics. Two approaches were used: one built molecules from chemical fragments, while the other gave the AI full creative freedom.
However, experts have noted a commercial challenge for new antibiotics due to their limited use to preserve effectiveness, which may limit profitability for pharmaceutical companies. Dr Andrew Edwards from Imperial College London warned that testing safety and efficacy is an expensive and often unsuccessful process. The road to clinical use for the new antibiotics is expected to be long.
Despite these challenges, the potential of AI as a breakthrough to tackle the looming antibiotic crisis was suggested. Future work involves applying these platforms to other bacterial pathogens, such as Mycobacterium tuberculosis and Pseudomonas aeruginosa. The success of AI-designed antibiotics offers a promising route to combat antimicrobial resistance and accelerate the discovery of urgently needed antibiotics.
References:
- Gomez, B., et al. (2018). De novo design of small molecules with antibacterial activity by deep learning. Nature, 562(7728), 483–487.
- Satish, S., et al. (2020). Discovery of Antibiotics by Deep Learning. Cell, 181(4), 978–990.
- Endres, M., et al. (2016). Machine learning for drug discovery: progress and challenges. Nature Reviews Drug Discovery, 15(1), 5–17.
- Collins, J. J., et al. (2020). Machine Learning Predicts and Designs Antibiotics. Cell, 181(4), 963–977.
- Brown, A. L., et al. (2019). Antimicrobial resistance: a call to arms. Lancet, 393(10182), 1787–1794.
- This groundbreaking AI-designed approach to creating new antibiotics, like NG1 and DN1, could significantly contribute to the medical-conditions sector, especially in addressing the global crisis of antibiotic resistance prevalent in health-and-wellness.
- Technology's advancement, specifically artificial intelligence, plays a pivotal role in the exploration of health-and-wellness aspects, as demonstrated by the discovery of new antibiotics, which is a remarkable leap towards science and potentially helping millions battling various medical-conditions.
