Temporal AI model predicts drivers of cell state trajectories across human aging
Gomez Ortega, J.; Nadadur, R. D.; Kunitomi, A.; Kothen-Hill, S.; Wagner, J. U. G.; Kurtoglu, S. D.; Kim, B.; Reid, M. M.; Lu, T.; Washizu, K.; Zanders, L.; Chen, H.; Zhang, Y.; Ancheta, S.; Lichtarge, S.; Johnson, W. A.; Thompson, C.; Phan, D. M.; Combes, A. J.; Yang, A. C.; Tadimeti, N.; Dimmeler, S.; Yamanaka, S.; Alexanian, M.; Theodoris, C. V.
Show abstract
Foundational AI models have recently shown promise for predicting the impact of perturbations on cell states. However, current models typically consider only one cell state at a time, limiting their ability to learn how cellular responses unfold over time, particularly across long trajectories such as diseases of aging. Here, we develop a temporal AI model, MaxToki, trained on nearly 1 trillion gene tokens including cell state trajectories across the human lifespan to generate cell states across long timelapses of human aging. MaxToki generalized to unseen trajectories through in-context learning and predicted novel age-modulating targets that were experimentally verified to influence age-related gene programs and functional decline in vivo. MaxToki represents a promising strategy for temporal modeling to accelerate the discovery of interventions for programming therapeutic cellular trajectories.
Matching journals
The top 7 journals account for 50% of the predicted probability mass.