ITMO Scientists' New Screening Platform Predicts Antibiotic-Nanoparticle Combinations

Antibiotics are losing the battle against infections. A study of mortality rates in 204 countries in 1990-2021 has shown that, on an annual basis, over a million people died of infections that are resistant to medication, and by 2050, this number could increase to almost ten million. Developing new medications against antibiotic-resistant bacteria is expensive and takes much time, so scientists aim to improve the efficiency of existing solutions. One of the ways to do that is to mix antibiotics with gold or silver nanoparticles. Thanks to their size, nanoparticles can easily enter a bacteria and destroy it; in combination with antibiotics, they also increase the efficiency and decrease the dosage and side effects of therapy.

Testing the antibacterial effect of nanoparticle-antibiotic combinations takes a lot of time and effort. First, scientists have to choose a promising combination, then synthesize it and test it experimentally. The entire process can take from several months to a year.

Researchers at ITMO ChemBio Cluster have already learned to use AI to identify particles that are selectively toxic to pathogenic bacteria but safe for useful microorganisms. This time, they’ve presented their new solution: the world’s first screening platform that predicts nanoparticle-antibiotic combinations that are efficient against antibiotic-resistant bacteria. It is based on machine learning models and genetic algorithms that make it possible to find effective combinations in mere days. Now, scientists no longer have to waste time and materials on hundreds of tests and can concentrate on testing the best options.

“In this new study, we combined nanoparticles with antibiotics in order to enhance the antibacterial effect. The platform allows us to use smaller dosages of both medications and nanoparticles, hence the lesser risk of side effects. What’s more, this makes it harder for bacteria to develop antibiotic resistance, as they have to adapt to the effects of nanoparticles and antibiotics at the same time. Our research will help choose efficient combinations against harmful multiresistant pathogens faster,” comments Susan Jyakhwo, the article’s author and a third-year PhD student at ITMO’s ChemBio Cluster.

For the new screening platform, the researchers collected data from over a hundred articles, published in the past ten years, on the effects of antibiotics and nanoparticles on various bacteria, both individually and combined. This data was processed and used to train the machine learning model and the genetic algorithms. The software platform also accounts for a multitude of factors: the shape and size of particles, the properties of antibiotics, and the bacterial species.

As a result, the screening platform has identified several new combinations that can potentially eliminate bacteria that are dangerous to humans and resistant to antibiotics. For example, Salmonella typhimurium, which causes typhoid fever, can be destroyed by a combination of gold nanoparticles and chloramphenicol. A combination of silver nanoparticles and amikacin can help against Klebsiella pneumoniae, which causes pneumonia and other respiratory tract diseases. What’s more, both these pathogens can be destroyed with a lesser amount of the combined treatment than if the nanoparticles and antibiotics were used separately, thereby reducing the side effects.

“We will verify the platform’s predictions experimentally in the lab, analyze how other large language models cope with the same tasks, and finalize our platform. In the future, we are planning to present it to companies that synthesize drugs and gain their support. I would also like to add more functions to the program. As of now, we’ve collected data on resistant bacteria that are harmful to humans. If we add data on pathogenic bacteria that are harmful to animals and crops, we’ll be able to use the platform to search for combinations of nanoparticles and antibiotics against more diseases,” notes Susan Jyakhwo.