Scientists have found a drug that could combat drug-resistant infections – and they did it using artificial intelligence.
Using a machine-learning algorithm, researchers at the Massachusetts Institute of Technology (MIT) and Canada’s McMaster University have identified a new antibiotic that can kill a type of bacteria responsible for many drug-resistant infections.
The compound kills Acinetobacter baumannii, which is a species of bacteria often found in hospitals. It can lead to pneumonia, meningitis and other serious infections.
The microbe is also a leading cause of infections in wounded soldiers in Iraq and Afghanistan.
In order to get training data for the model, they first exposed the bacteria grown in a lab dish to around 7,500 different chemical compounds in order to see which could inhibit growth of the microbe. They fed the structure of each molecule into their model and told it whether each structure could inhibit bacterial growth.
People walk through the Massachusetts Institute of Technology campus in Cambridge, Massachusetts, on Wednesday, June 2, 2021. (Photographer: Adam Glanzman/Bloomberg via Getty Images)
After the model was trained, it was used to analyze a set of 6,680 compounds it had not seen before, and researchers narrowed down 240 hits to test experimentally, focusing on compounds with structures that were different from those of existing antibiotics or molecules from the training data. That testing led to nine antibiotics, including one that was very strong.
A lab at McMaster University is now working for others to optimize the medicinal properties of the compound and hopefully develop it for eventual use in patients.
The study’s authors also plan to use their modeling approach to identify potential antibiotics for other types of drug-resistant infections.