Unlocking hidden biomolecular conformational landscapes in diffusion models at inference time
This work addresses the challenge of predicting conformational landscapes for biomolecules like proteins, which is crucial for understanding their function, and represents an incremental improvement by enhancing existing diffusion models at inference time.
The authors tackled the problem of predicting biomolecular conformational distributions by introducing ConforMix, an inference-time algorithm that enhances sampling in diffusion models, enabling efficient discovery of conformational variability without prior knowledge of major degrees of freedom, and demonstrated its scalability and accuracy in capturing structural changes like domain motion and cryptic pocket flexibility.
The function of biomolecules such as proteins depends on their ability to interconvert between a wide range of structures or "conformations." Researchers have endeavored for decades to develop computational methods to predict the distribution of conformations, which is far harder to determine experimentally than a static folded structure. We present ConforMix, an inference-time algorithm that enhances sampling of conformational distributions using a combination of classifier guidance, filtering, and free energy estimation. Our approach upgrades diffusion models -- whether trained for static structure prediction or conformational generation -- to enable more efficient discovery of conformational variability without requiring prior knowledge of major degrees of freedom. ConforMix is orthogonal to improvements in model pretraining and would benefit even a hypothetical model that perfectly reproduced the Boltzmann distribution. Remarkably, when applied to a diffusion model trained for static structure prediction, ConforMix captures structural changes including domain motion, cryptic pocket flexibility, and transporter cycling, while avoiding unphysical states. Case studies of biologically critical proteins demonstrate the scalability, accuracy, and utility of this method.