We are excited to add the first tool for automated model building into cryo-EM maps: Model Angelo.
In short, this tool uses advances in machine learning representation of protein structure with cryo-EM maps to leverage the combined information of chain tracing, amino acid geometry, and cryo-EM density simultaneously.
It works really well – in 10 minutes on COSMIC2 we were able to build a ~40 kDa model for KIFBP (a project in the Cianfrocco laboratory) for the published map EMD-24677 (!):
While subtle differences exist between the manually built model and Model Angelo, Model Angelo got the sequence registry and topology of the protein exactly the same as what we found after the painstaking work of model building.
Try it out! Model Angelo tool page
Preprint: ModelAngelo: Automated Model Building in Cryo-EM Maps. Jimali, Kimanius, Scheres. arXiv:2210.00006