No new classes have been developed for decades, and in 2019 alone, resistance to existing drugs contributed to 1.27 million deaths worldwide – describes the website. However, scientists used it to test millions of compounds for antibiotic activity and managed to discover a completely new class of antibiotics. The entire study was described in the scientific journal “”.
Artificial intelligence has helped discover new classes of antibiotics
The effect of over 39,000 compounds on the pathogenic bacterium Staphylococcus aureus and three types of human cells – from the liver, skeletal muscles and lungs – was tested. Artificial intelligence examined their relationships: antibacterial and toxicity. Trained artificial intelligence systems checked 12 million compounds.
This way, 3,646 were found with drug-like properties. Additional calculations then identified substructures that could explain the properties of each compound. Thus, two non-toxic compounds were discovered, capable of killing Staphylococcus aureus and vancomycin-resistant enterococci. New classes of potential antibiotics have been identified. Subsequent tests in mice confirmed that they could be effective in both cases.
Groundbreaking research
“Our approach enables the discovery of structural classes of antibiotics using deep learning and demonstrates that machine learning models in drug discovery can be used to provide insight into the chemical substructures underlying the selective action of antibiotics,” the authors of the study boast in the paper. Their work allowed the identification of one of the few new classes of antibiotics in 60 years. Moreover, only a few of those discovered so far are effective against both MRSA and vancomycin-resistant bacteria.
Source: Gazeta
Mabel is a talented author and journalist with a passion for all things technology. As an experienced writer for the 247 News Agency, she has established a reputation for her in-depth reporting and expert analysis on the latest developments in the tech industry.
