A group of scientists in Canada and the United States, with the help of artificial intelligence (AI), discovered an antibiotic that can kill one of the deadliest species in the world of superbugs.
AI helped them reduce the number of chemical compounds they had to analyze to get to the ones they needed, which were later analyzed in a laboratory. until they found an important antibiotic substance they called abaucin, and although it promises to be very useful to counteract some diseases, it is still in a trial period.
For these researchers, AI could become an important tool to speed up the discovery process for new drugs.
But what are superbugs?
When bacteria become immune to some antibiotics, it is in those moments that we are dealing with a superbug. Some scientific studies estimate that more than a million people a year die from infections caused precisely by agents that have become resistant to antibacterial agents.
For this reason, the researchers decided to focus their study on one of the most problematic bacteria recorded to date: Acinetobacter baumannii. This usually causes pneumonia and infects wounds, and It is one of the three superbugs that the World Health Organization (WHO) has classified as “critical” because of its effects on the respiratory system and skin.
This bacterium is feared by medical personnel, as it can easily ignore multiple antibiotics that become a problem in hospitals or nursing homes. The bacteria can survive on any surface and in surgical equipment.
Jonathan Stokes, a researcher at McMaster University in Canada, told BBC Mundo that he would classify this bacterium as “public enemy number one”, since it has become “really common”, and that has generated its resistance “to almost all antibiotics”.
Artificial Intelligence (AI) in science
The credits for the discovery of the new antibiotic could not be taken only by the AI, the researchers trained it. Several drugs were taken, and they were manually tested on Acinetobacter baumannii in order to identify which ones slowed down or could kill the bacteria to a large extent, and then put them at the service of the AI so that through the identification of compounds, I could identify new ones.
From a list of 6,680 compounds, in an hour and a half the tool provided scientists with a short list of antibiotics, made up of 240, which were later tested in the laboratory, reducing the list to nine, where the powerful antibiotic abaucin finally came from. The results have already been published in the scientific journal Nature Chemical Biology.
It was possible to prove that it could treat infected wounds in mice, a finding that is only the beginning for now. It has not yet been possible to establish whether it would have the same effects in the human body. Then, the next step, the experts explain, will be to perfect the drug to go to clinical trials.