Massachusetts Institute of Technology (MIT) researchers used artificial intelligence to develop antibiotics capable of fighting certain bacteria. They aim particularly the most resistant, such as multi -resistant gonorrhea and SARM (staphylococcus golden resistant to methicillin).
Advertisement
It is respectively a sexually transmitted infection which causes unusual pain and losses, while the SRM is a skin or blood infection. According to Aarti Krishnan, MIT researcher, “We wanted to explore completely new paths, far from any existing antibiotic, in order to fight against resistance to antimicrobials in a radically different way.”
An anti bacteria created by the MIT
The team has exploited generative algorithms without specifying its name, to design more than 36 million hypothetical molecules, then selecting those presenting the greatest antibacterial potential. Among them, two compounds have been successfully synthesized and tested in the laboratory.
One, called NG1, has shown remarkable effectiveness against Gonorrhea, both in Petri boxes and in mice. Similarly, six other molecules have made it possible to neutralize the SARM, which is very encouraging for the future.
Thanks to these discoveries, scientists will be able to create new drugs against other resistant bacteria. As a reminder, antimicrobial resistance, that is to say when bacteria, viruses or fungi become capable of surviving drugs, is already the source of millions of deaths each year. It could continue to progress if new solutions are not found.
Advertisement
In 2021, she contributed to nearly 4.71 million dead around the world. The use of AI to generate unpublished antibiotics thus makes it possible to find new approaches still little explored in the fight against these infections. In addition, it accelerates the discovery. As MIT researchers point out, “By exploring unpublished territories of the chemical space, we hope to identify compounds capable of surpassing the limits of existing antibiotics.”
Scientist holding a bike plaque for bacterial or microorganism diagnosis, blurred microscopic background in the laboratory. Selective petri with colonies of bacteria under the lens of a microscope. © Komsan Loonprom Shutterstock
When AI helps medicine
This is not the first time that the AI has helped doctors. It has made it possible to discover several promising antibiotics in recent years. In 2019, MIT researchers developed halicine, a molecule capable of fighting resistant bacteria such as Clostridium difficile, Acinetobacter baumannii et Mycobacterium tuberculosis. This discovery has marked a major advance, being the first new antibiotic identified thanks to AI for over 60 years.
More recently, in 2023, scientists used AI models to identify a new class of effective antibiotics against golden staphylococcus, resistant to methicillin (STM).
