Researchers at Karolinska Institutet have contributed to an international study published in Science Translational Medicine demonstrating how artificial intelligence (AI) can accelerate the discovery of new antibiotics against Neisseria gonorrhoeae, the bacterium that causes gonorrhoea.
Antibiotic resistance in gonorrhoea is a growing global health concern. Many strains of Neisseria gonorrhoeae have developed resistance to commonly used antibiotics, raising concerns that infections may become harder to treat in the future.
In the new study, led by researchers at the Massachusetts Institute of Technology (MIT) with contributions from researchers at Karolinska Institutet, a deep‑learning model was trained using tens of thousands of experimentally tested molecules. The model was then used to screen around six million chemical compounds. From this large dataset, the AI identified a small number of promising candidates, including two that showed strong activity against both antibiotic‑sensitive and multidrug‑resistant strains.
One of the compounds rapidly killed the bacteria and cleared the infection in a laboratory model that mimics human tissue. The lead compound, A1, also reduced bacterial levels in animal models and showed a selective antibacterial effect.
Photo: N/A
Can speed up new treatments
Researchers at Karolinska Institutet contributed by identifying how the compound works inside bacterial cells. Using a proteomics‑based method known as the PISA assay, they found that A1 targets alanine racemase, an enzyme essential for building the bacterial cell wall.
“Understanding how new antibiotics work is crucial for their further development and clinical translation. Using advanced proteomics, we were able to identify alanine racemase as the target of A1 and provide insight into its antibacterial activity”, says Amir Ata Saei , Assistant Professor at the Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet.
The researchers combined proteomics, genetics, and biochemical experiments to confirm the mechanism. The results indicate that the compounds act differently from existing antibiotics, which may be important in addressing antibiotic resistance.
The study highlights how AI‑driven discovery and experimental biology can be combined to speed up the development of new treatments. The researchers hope that similar approaches can be applied to other pathogens where treatment options are becoming limited.
The study is a collaboration between researchers at Karolinska Institutet -Margaux Gaborieau, Edmund Loh, Marie-Stéphanie Aschtgen, and Amir Ata Saei – and colleagues at MIT, the Broad Institute, Harvard University, and several other international institutions.
Publication
Deep learning-enabled discovery of antibiotics effective against Neisseria gonorrhoeae.
Anahtar MN, Valeri JA, Modaresi SM, Krishnan A, Donghia NM, Palace SG, Zheng EJ, Gulati A, Jorgenson A, Junaid A, Bandyopadhyay P, Luttens A, Suresh K, Edwards P, Wong F, Zhang Y, Ritz D, Gaborieau M, Loh E, Gaetani M, Aschtgen MS, Saei AA, Grad YH, Ingber DE, Collins JJ
Sci Transl Med 2026 Jun;18(854):eads4699
