Generative Modeling of Full-Atom Protein Conformations using Latent Diffusion on Graph Embeddings

2025-06-20

AI Summary

Addressing the challenge of simultaneously achieving conformational diversity and atomic-level accuracy in generating all-atom structures of dynamic proteins (e.g., GPCRs), this work introduces the first graph-embedding latent diffusion model tailored for all-atom protein structure generation. We propose three novel graph pooling strategies—blind, sequence-aware, and residue-aware—to hierarchically encode structural context, and incorporate a dihedral-angle loss-regularized decoder to ensure system-specific, high-fidelity conformational sampling. The model employs Chebyshev graph neural networks (ChebNet) to explicitly represent all side-chain heavy atoms and learns conformational distributions directly from molecular dynamics (MD) trajectories. Evaluated on the D2R-MD dataset, it achieves an all-atom lDDT of 0.70 and Cα-lDDT of 0.80, with Jensen–Shannon divergence <0.03 for backbone and side-chain dihedral angle distributions—substantially outperforming existing methods. Code and data are publicly released.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Natural Language Processing: Generation

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

Generating diverse, all-atom conformational ensembles of dynamic proteins such as G-protein-coupled receptors (GPCRs) is critical for understanding their function, yet most generative models simplify atomic detail or ignore conformational diversity altogether. We present latent diffusion for full protein generation (LD-FPG), a framework that constructs complete all-atom protein structures, including every side-chain heavy atom, directly from molecular dynamics (MD) trajectories. LD-FPG employs a Chebyshev graph neural network (ChebNet) to obtain low-dimensional latent embeddings of protein conformations, which are processed using three pooling strategies: blind, sequential and residue-based. A diffusion model trained on these latent representations generates new samples that a decoder, optionally regularized by dihedral-angle losses, maps back to Cartesian coordinates. Using D2R-MD, a 2-microsecond MD trajectory (12 000 frames) of the human dopamine D2 receptor in a membrane environment, the sequential and residue-based pooling strategy reproduces the reference ensemble with high structural fidelity (all-atom lDDT of approximately 0.7; C-alpha-lDDT of approximately 0.8) and recovers backbone and side-chain dihedral-angle distributions with a Jensen-Shannon divergence of less than 0.03 compared to the MD data. LD-FPG thereby offers a practical route to system-specific, all-atom ensemble generation for large proteins, providing a promising tool for structure-based therapeutic design on complex, dynamic targets. The D2R-MD dataset and our implementation are freely available to facilitate further research.

Problem

• Generates full-atom protein conformations from molecular dynamics data
• Addresses lack of atomic detail in protein structure modeling
• Improves conformational diversity modeling for dynamic proteins like GPCRs

Innovation

• Latent diffusion for full protein generation
• Chebyshev graph neural network embeddings
• Dihedral-angle regularized decoder mapping