Enregistré dans:
Détails bibliographiques
Auteurs principaux: Krook, Jonathan, Janson, Axel, Andén, Joakim, Weber, Melanie, Öktem, Ozan
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2602.21915
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866914374362333184
author Krook, Jonathan
Janson, Axel
Andén, Joakim
Weber, Melanie
Öktem, Ozan
author_facet Krook, Jonathan
Janson, Axel
Andén, Joakim
Weber, Melanie
Öktem, Ozan
contents We present a geometry-aware method for heterogeneous single-particle cryogenic electron microscopy (cryo-EM) reconstruction that predicts atomic backbone conformations. To incorporate protein-structure priors, we represent the backbone as a graph and use a graph neural network (GNN) autodecoder that maps per-image latent variables to 3D displacements of a template conformation. The objective combines a data-discrepancy term based on a differentiable cryo-EM forward model with geometric regularization, and it supports unknown orientations via ellipsoidal support lifting (ESL) pose estimation. On synthetic datasets derived from molecular dynamics trajectories, the proposed GNN achieves higher accuracy compared to a multilayer perceptron (MLP) of comparable size, highlighting the benefits of a geometry-informed inductive bias.
format Preprint
id arxiv_https___arxiv_org_abs_2602_21915
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Protein Graph Neural Networks for Heterogeneous Cryo-EM Reconstruction
Krook, Jonathan
Janson, Axel
Andén, Joakim
Weber, Melanie
Öktem, Ozan
Computer Vision and Pattern Recognition
We present a geometry-aware method for heterogeneous single-particle cryogenic electron microscopy (cryo-EM) reconstruction that predicts atomic backbone conformations. To incorporate protein-structure priors, we represent the backbone as a graph and use a graph neural network (GNN) autodecoder that maps per-image latent variables to 3D displacements of a template conformation. The objective combines a data-discrepancy term based on a differentiable cryo-EM forward model with geometric regularization, and it supports unknown orientations via ellipsoidal support lifting (ESL) pose estimation. On synthetic datasets derived from molecular dynamics trajectories, the proposed GNN achieves higher accuracy compared to a multilayer perceptron (MLP) of comparable size, highlighting the benefits of a geometry-informed inductive bias.
title Protein Graph Neural Networks for Heterogeneous Cryo-EM Reconstruction
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2602.21915