CryoACE: An Atom-centric Framework for Accurate and Automated Model Building in Cryo-EM

Protein automodeling from cryo-EM density maps faces unique challenges in enforcing physicochemical validity and managing conformational heterogeneity. Current solvers are often limited to static predictions or require computationally intensive heuristic searches. We present CryoACE, an end-to-end framework that reconstructs precise atomic graphs for both homogeneous and heterogeneous structures. Our method features two key innovations: an atom-centric reconstruction paradigm, where density features are sampled directly at atomic coordinates and iteratively recycled to refine structures—replacing expensive voxel convolutions for efficient multimodal fusion—and a training-free guidance mechanism that leverages predicted local resolution priors to resolve dynamic ambiguity. Validated on a newly constructed high-quality dataset, CryoACE significantly outperforms existing baselines on static benchmarks and, for the first time, unveils atomic-level dynamic conformations on complex real-world datasets like EMPIAR-10345 without relying on pre-built static structures. We release our code, model weights, and dataset to facilitate future research.

Automated macromolecular model building from cryo-EM density maps frequently struggles to maintain strict stereochemical validity and map-model fitness, especially when resolving noisy, low-resolution regions and complex conformational heterogeneity. To tackle these challenges, we introduce CryoACE, an end-to-end multi-modal framework that reformulates structure reconstruction into an atom-centric coordinate generation task. Instead of relying on expensive voxel-wise convolutions, CryoACE directly samples density features at predicted atomic coordinates and iteratively recycles them for precise structural refinement, further utilizing an auxiliary local resolution prior to guide dynamic conformation sampling. Validated on a newly curated dataset of 10,915 high-quality triplets, CryoACE achieves state-of-the-art accuracy with nearly 100% structural completeness on static benchmarks and, for the first time, successfully unveils atomic-level dynamic conformations on complex real-world heterogeneous datasets like EMPIAR-10345.