Solving Inverse Problems in Protein Space using Diffusion-Based Priors
ADP-3D (Atomic Denoising Prior for 3D reconstruction) is a framework that conditions a diffusion model in protein space with any observations for which the measurement process can be physically modeled. Inspired from plug-n-play, ADP-3D demonstrates versatility in solving inverse problems in protein space with a pretrained diffusion model as a learned prior. It outperforms existing posterior sampling methods at reconstructing full protein structures from partial structures. It shows that a protein diffusion model can be guided to perform atomic model refinement in simulated cryo-EM density maps and that it can be conditioned on a sparse distance matrix. For more details, please read the preprint (Levy et al., 2024) and visit the project page.
GitHub Repository
References
2024
Protein Folding
Solving Inverse Problems in Protein Space Using Diffusion-Based Priors
Axel Levy, Eric R Chan, Sara Fridovich-Keil, and 3 more authors
The interaction of a protein with its environment can be understood and controlled via its 3D structure. Experimental methods for protein structure determination, such as X-ray crystallography or cryogenic electron microscopy, shed light on biological processes but introduce challenging inverse problems. Learning-based approaches have emerged as accurate and efficient methods to solve these inverse problems for 3D structure determination, but are specialized for a predefined type of measurement. Here, we introduce a versatile framework to turn raw biophysical measurements of varying types into 3D atomic models. Our method combines a physics-based forward model of the measurement process with a pretrained generative model providing a task-agnostic, data-driven prior. Our method outperforms posterior sampling baselines on both linear and non-linear inverse problems. In particular, it is the first diffusion-based method for refining atomic models from cryo-EM density maps.
