Search papers, labs, and topics across Lattice.
This survey reviews recent AI methods for modeling protein dynamics, focusing on learning from structural ensembles/trajectories, energy signals, and accelerating molecular simulations. It covers methods for conformation and trajectory generation, Boltzmann generators, physics-aware adaptation, ML potentials, coarse-grained modeling, and collective variable discovery. The authors highlight open challenges including scalability, thermodynamic consistency, kinetic fidelity, and integration with experimental data, providing a valuable resource for researchers in this area.
AI is poised to revolutionize protein dynamics research, but key challenges remain in ensuring scalability, thermodynamic consistency, and kinetic fidelity.
Protein dynamics underlie many biological functions, yet remain difficult to characterize due to the high computational cost of molecular dynamics simulations and the scarcity of dynamic structural data. This survey reviews recent advances in artificial intelligence for protein dynamics from three perspectives: learning from structural ensembles and trajectories, learning from physical energy signals, and learning to accelerate molecular simulations. We summarize representative methods for conformation ensemble generation, trajectory generation, Boltzmann generators, physics-aware adaptation, machine learning potentials, coarse-grained modeling, and collective variable discovery. We further discuss available datasets and key open challenges, such as scalability, thermodynamic consistency, kinetic fidelity, and integration with experimental constraints.
Mila