Discovery of a structural class of antibiotics with explainable deep learning. (https://www.nature.com/articles/s41586-023-06887-8)

These scientists wanted to find new types of antibiotics to help fight against bacteria that are becoming resistant to the antibiotics we currently have. They used a special kind of computer program called a deep learning model to help them search for these new antibiotics.

The deep learning model learned about different chemical structures that are associated with antibiotic activity. It then used this knowledge to predict which new compounds might have antibiotic activity.

To test if the predictions were correct, the scientists did experiments on 39,312 different compounds to see if they had antibiotic activity and if they were harmful to human cells. They also used the deep learning model to predict the antibiotic activity and harm to human cells for 12,076,365 other compounds.

Using a special kind of algorithm, the scientists looked at the chemical structures of the compounds that the deep learning model predicted to have high antibiotic activity and low harm to human cells. They found patterns in these structures that could explain why these compounds might be good antibiotics.

To make sure their predictions were correct, the scientists tested 283 of these compounds in the lab. They found that some of the compounds were indeed effective at killing a type of bacteria called Staphylococcus aureus, including the ones that are resistant to other antibiotics.

One of these compounds was also tested in mice and it was able to reduce the amount of bacteria in both skin and systemic infections. This means that it could potentially be used to treat infections in humans.

This study shows that using deep learning models can help scientists find new types of antibiotics. It also shows that these models can explain why certain compounds might be good antibiotics by looking at their chemical structures.

Wong F., Zheng EJ., Valeri JA., Donghia NM., Anahtar MN., Omori S., Li A., Cubillos-Ruiz A., Krishnan A., Jin W., Manson AL., Friedrichs J., Helbig R., Hajian B., Fiejtek DK., Wagner FF., Soutter HH., Earl AM., Stokes JM., Renner LD., Collins JJ. Discovery of a structural class of antibiotics with explainable deep learning. Nature. 2024 Feb;626(7997):177-185. doi: 10.1038/s41586-023-06887-8. Epub 2023 Dec 20.