Pathogens are difficult to fight off with antibiotics because they resemble bacteria and can be successfully taken up by bacteria. But scientists at Los Alamos National Laboratory have come up with a new way to tackle this problem. They are using machine learning to fight pathogens.
The study was published in the journal Communications Chemistry. It primarily focuses on identifying specific molecular properties, which may help in the discovery of new antibiotics. It turns out that new antibiotics can be effective against bacteria that are becoming resistant to current drugs.
Gnana Gnanakaran, a scientist at Los Alamos National Laboratory, said some bacteria are resistant to antibiotics and it is very difficult to find compounds that can penetrate and stop these bacteria. Their approach helps drill down into the molecular details of bacteria, which is critical to successful drug development.
It is true that bacterial defenses are tough. Gram-negative bacteria have an outer layer that makes it difficult for antibiotics to penetrate. It also has the ability to push out compounds that have gotten inside. This will weaken the effectiveness of the antibiotic.
The research team is turning to machine learning to solve this problem. They have developed a model that they claim can pinpoint which properties of a particular compound help it break through defenses and stay inside.
The study primarily focused on a bacterium called Pseudomonas aeruginosa, which is commonly found in infections. They analyzed over 1000 different compounds with the help of machine learning and learned how they interact with the outer layer of bacteria.
The discovery reveals the properties that make certain compounds effective against Pseudomonas aeruginosa and also paves the way for similar studies on other bacteria.
