DiffPIE: Guiding Deep Generative Models to Explore Protein Conformations under External Interactions

Abstract

In recent years, many foundation generative models have been developed to pre-dict structures of molecules and materials. Although these foundation models have achieved great success, it is challenging to collect enough data to train foundation generative models. One such example is to predict protein conformations with protein-environmental interactions (PEI), such as interactions introduced by organic linkers or material surfaces. We propose a physics-guided route to extrapolate foundation mod-els beyond their training domain. Our method couples a pretrained deep generative model with explicit, physics-based interaction potentials for PEI, steering sampling to-ward conformations consistent with external constraints without any retraining or fine-tuning. We demonstrate accurate and efficient conformation prediction of (i) cyclic peptide with organic linkers and (ii) peptide adsorbed on the gold surface. The gen-erated structures serve as high-quality initial conditions for downstream simulations, providing a general, systematic approach to extend foundation models to proteins under system-specific environmental interactions.