CryoAtom improves model building for cryo-EM

Constructing atomic models from cryogenic electron microscopy (cryo-EM) density maps is essential for interpreting molecular mechanisms. Here we present CryoAtom, an approach for de novo model building for cryo-EM maps, leveraging recent advancements in AlphaFold2 to improve the state-of-the-art method ModelAngelo. To accommodate the cryo-EM map information, CryoAtom replaces the global attention mechanism in AlphaFold2 with local attention, which is further enhanced by a novel three-dimensional rotary position embedding. CryoAtom produces more complete models, reduces the resolution requirement and accelerates the modeling. The application of CryoAtom to three large maps demonstrates its ability to detect previously uncharacterized proteins with unknown functions, improve the modeling of conformational changes and compartmentalize the map to isolate nonprotein components. A particular case includes a 104-protein complex that was modeled within a few hours and a minor conformational change of a single protein domain was detected at the periphery when models from two different maps were compared. CryoAtom stands as an accurate method currently available for model building of proteins in cryo-EM structure determination. The source code and model parameters are available from GitHub ( https://github.com/YangLab-SDU/CryoAtom ).