AI finds new target for monkeypox ^{CEFR B1}

During the 2022 mpox outbreak scientists looked for better ways to prevent and treat the disease. Researchers in Italy identified twelve antibodies from people who had been infected or vaccinated. Because it was not clear which viral proteins those antibodies targeted, a team at The University of Texas at Austin used the AlphaFold 3 artificial intelligence model to predict the likely targets among roughly thirty-five surface proteins.

AlphaFold 3 indicated with high confidence that several antibodies bind to a protein called OPG153. Follow-up laboratory work verified the prediction, and an engineered form of OPG153 given to mice triggered production of neutralizing antibodies. The study was published in Science Translational Medicine.

Researchers say the approach — starting from antibodies, finding the viral target and then engineering an antigen — could produce vaccine antigens or antibody therapies that are cheaper and easier to make than whole-virus vaccines. Patent applications have been filed and the work received partial funding from the Welch Foundation.