Accurate protein structure determination from cryo-EM maps using deep learning and structure prediction

Abstract

Cryo-electron microscopy (cryo-EM) has become the main-stream technique for macromolecular structure determination. However, due to intrinsic resolution heterogeneity, accurate modeling of all-atom structure from cryo-EM maps remains challenging even for maps at near-atomic resolutions. Addressing the challenge, we present EMProt, a fully automated method for accurate protein structure determination from cryo-EM maps by efficiently integrating map information and structure prediction with a three-track attention network. EMProt is extensively evaluated on a diverse test set of 177 experimental cryo-EM maps with up to 54 chains in a case at <4.0 [A] resolutions, and compared with state-of-the-art methods including DeepMainmast, ModelAngelo, phenix.dock_and_rebuild, and AlphaFold3. It is shown that EMProt significantly outperforms the existing methods in recovering the protein structure and building the complete structure. In addition, the built models by EMProt also exhibit a high accuracy in model-to-map fit and structure validations. EMProt is freely available at http://huanglab.phys.hust.edu.cn/EMProt/.