3j9d: Capsid protein VP1 from Bluetongue virus 1
3.3Å cryo-EM model, 746 protein residues (Zhang et al., 2016)
(Thank you very much to Hong Zhou and Polly Roy for agreeing to my use of their data for this case study)
Ever since the so-called "resolution revolution" in 2013-2014, the number of structures derived from cryo-EM data deposited into the PDB has increased exponentially over time. While a good cryo-EM map has some clear advantages over a crystallographic map of similar resolution (the biggest, of course, being that there is no phase problem to solve - that is, you already have the map before you build your model), the fact remains that most EM maps are still what a crystallographer would define as "low resolution". As of Jan 2019, of the 2,728 cryo-EM models in the wwPDB just 64 have resolutions higher than 3Å.
3j9d is a moderately large model of 746 residues hand-built into 3.3Å density, with a MolProbity score of 3.48. The video to the left captures the full process of rebuilding it over the course of just over 1 hour: loading model and map from the PDB/EMDB, some simple edits (adding missing atoms) to prepare the model for simulation, correcting cis peptide bonds flagged in the validation panel, then working over the entire model from N to C terminus to reach a final MolProbity score of 0.97.