Overview
- A peer‑reviewed study published in June 2026 reports researchers trained a graph neural network on lab tests of 38,650 molecules and then virtually screened about six million compounds to find new anti‑gonorrhea chemicals.
- Computational filtering produced a prioritized set of candidates (reported as 213 in some accounts and 83 in others) that the team narrowed to two lead molecules named A1 and MP20.
- Proteomics and genetic tests showed A1 binds and inhibits alanine racemase, an enzyme needed for bacterial cell‑wall building, while MP20 disrupts the bacterial membrane and damages DNA.
- Both compounds rapidly killed multidrug‑resistant Neisseria gonorrhoeae in a human Vagina Organ Chip and in a mouse vaginal infection model and did not induce resistance in the limited laboratory tests reported.
- Researchers and public‑health experts caution these are early preclinical hits that will need medicinal‑chemistry optimization, expanded safety studies, and clinical trials before they can become treatments, underscoring the broader need for new antibiotics as resistance to current drugs rises.