New Research
Learning what cancer reveals:
Invoke Bio's NEMo Model
Every tumor evolves its own path. Our new preprint introduces NEMo, a model that learns which mutations a patient's immune system will actually recognize and attack, by watching what happens inside real tumors instead of just in a test tube.
Personalized cancer vaccines only work if they target the right piece of a tumor. NEMo (the Neoantigen Elimination Model) is Invoke Bio's answer to that targeting problem, and it's the subject of a new preprint we're sharing today.
Most target-selection tools only ask whether a mutated protein fragment, a neoantigen, can bind to the machinery that displays it on a cell's surface. Binding matters, but it isn't the whole story: a lot has to happen between binding and an actual immune response. So instead of relying only on binding affinity, we trained NEMo on immunoediting: longitudinal data showing which mutations disappear from a tumor under immunotherapy, and which ones persist. Disappearance is the clearest signal we have that a T cell really did recognize and kill the cells carrying that mutation.
Want the full interactive deep dive?
Explore the interactive piecePrimary Research
Predicting Neoantigen Immunogenicity from In Vivo Immune Editing
Sears, T.J., Lee, K., Munoz Perez, M., Rasmussen, R., Pagadala, M.S., Tanaka, K., Subramanian, A., Moding, E.J., Zanetti, M., Carter, H.
T.J. Sears, M. Munoz Perez, and R. Rasmussen are employees and shareholders of Invoke Bio. E.J. Moding, M. Zanetti, and H. Carter serve as scientific advisors to Invoke Bio. Full competing-interest statement available on bioRxiv.
View on bioRxiv