Using AI to develop super strong antibiotics that can kill antibiotic-resistant bacteria

To do so, scientists have spent a lot of effort creating a computer model that synthesizes information on millions of chemical compounds and uses machine learning algorithms to pick out what one can develop into an effective new antibiotic strain, then select a few specific compounds for development and in-depth testing. The results showed that the new antibiotic could easily kill E. coli and many other common bacteria when tested on mice.

Most of the newly developed antibiotics today are variations of existing drugs, using similar mechanisms. However, this AI-developed antibiotic has a completely different mechanism, meaning it has the ability to treat infections that current drugs cannot.

Using AI to develop super strong antibiotics that can kill antibiotic-resistant bacteria Picture 1

'We want to develop a platform that paves the way for harnessing the power of artificial intelligence to discover antibiotics,' said James Collins, Professor of Medical Science and Engineering at MIT.

This new approach could help identify the different compounds that make bacteria resistant to drugs faster, at significantly lower costs than other methods. The researchers trained their AI model with 2,500 existing molecules, and then tested it on a data library of 6,000 compounds to identify molecules that could kill E.coli bacteria. . The model identified a specific molecule as a potential target, and when researchers tested it in the lab, the results showed that it could also destroy dangerous and common antibiotic-resistant bacteria. Currently such as Clostridium difficile, Acinetobacter baumannii and Mycobacterium tuberculosis.

The problem of antibiotic resistant bacteria is really becoming a threat to human health. MIT scientists' new method of antibiotic research opens up prospects for solving this problem. In the future, the team plans to use this model to optimize existing antibiotics as well as develop new antibiotics. In particular, the application of additional deep learning technology at all stages of antibiotic development, from discovery to improvement of efficacy and toxicity through modification of medicinal ingredients and pharmaceutical chemistry, will play a role. The game is extremely important.