Logged-out Icon

AI Generates New Hope in the Fight Against Antibiotic-Resistant Bacteria

In a pioneering endeavor to address the pressing global threat of antibiotic resistance, researchers leverage the power of generative artificial intelligence to design and synthesize six new compounds targeting antibiotic-resistant strains, marking a significant milestone in the fight against drug-resistant bacteria

Ai drug

In a groundbreaking development, researchers at Stanford Medicine and McMaster University have harnessed the power of generative artificial intelligence to create novel antibiotics targeting antibiotic-resistant bacteria. With nearly 5 million deaths attributed to antibiotic resistance each year, the urgent need for new ways to combat these resistant strains has never been more pressing.

The researchers’ innovative model, named SyntheMol (short for synthesizing molecules), has successfully generated structures and chemical recipes for six new compounds aimed at killing resistant strains of Acinetobacter baumannii, a notorious pathogen responsible for a significant portion of antibacterial resistance-related deaths. The study, published on March 22 in the journal Nature Machine Intelligence, details the model’s development and the experimental validation of these novel compounds.

James Zou, Ph.D., an associate professor of biomedical data science and co-senior author of the study, emphasized the pressing public health need for rapid antibiotic development. The research team hypothesized that there are countless potential molecules that could be effective drugs but have yet to be discovered or tested. This realization led them to explore the use of AI to design entirely new molecules that have never been observed in nature.

Prior to the emergence of generative AI, researchers relied on computational approaches that sifted through existing drug libraries to identify compounds with the highest potential to act against specific pathogens. While this technique yielded results, it only scratched the surface of the vast chemical space, estimated to contain nearly 1060 possible drug-like molecules.

Generative AI’s tendency to “hallucinate,” or generate novel responses, proved to be a valuable asset in drug discovery. However, previous attempts to generate new drugs using this type of AI often resulted in compounds that were impossible to synthesize in the real world. To address this challenge, the researchers implemented guardrails around SyntheMol’s activity, ensuring that the generated molecules could be synthesized in a laboratory setting.

The model was trained using a library of over 130,000 molecular building blocks and a set of validated chemical reactions. SyntheMol not only generated the final compounds but also provided the step-by-step instructions for their synthesis. Additionally, the model was trained on existing data regarding the antibacterial activity of various chemicals against A. baumannii. With these guidelines in place, SyntheMol generated approximately 25,000 potential antibiotics and their corresponding recipes in less than nine hours.

To prevent the rapid development of bacterial resistance to the new compounds, the researchers filtered the generated molecules to include only those that were dissimilar from existing antibiotics. From the pool of generated compounds, the researchers selected the 70 most promising candidates and collaborated with the Ukrainian chemical company Enamine to synthesize them. Enamine successfully generated 58 of these compounds, six of which demonstrated effective antibacterial activity against a resistant strain of A. baumannii when tested in the lab. These new compounds also showed promise against other infectious bacteria prone to antibiotic resistance, such as E. coli, Klebsiella pneumoniae, and MRSA.

Further testing of two of the six compounds in mice indicated their apparent safety, with the next step being to evaluate their efficacy in mice infected with A. baumannii. The six compounds generated by SyntheMol are notably diverse and distinct from existing antibiotics. While the exact mechanisms of their antibacterial properties remain unknown, exploring these details could yield valuable insights for future antibiotic development.

The researchers emphasize that SyntheMol is not only designing new molecules but also expanding our understanding of previously unexplored chemical spaces. Zou and Swanson are currently refining the model and broadening its applications, collaborating with other research groups to utilize SyntheMol for drug discovery in heart disease and the creation of novel fluorescent molecules for laboratory research.

The development of SyntheMol and its successful generation of novel antibiotics marks a significant milestone in the fight against antibiotic resistance. By harnessing the power of generative AI, researchers are unlocking new frontiers in drug discovery, offering renewed hope in the battle against one of the most pressing public health challenges of our time.

This website uses cookies to ensure you get the best experience on our website