Post-classification refinement was used to improve the accuracy of the classification.
Models of Kir6.2 CTD were refined against maps after focused classification and refinement of Kir6.2 CTD in the ATP + GBM state.
After several cycles of 3D classification and refinement, a 2D classification step using the alignment parameters from the 3D refinement was performed to check the alignment accuracy and to further exclude particles liable to misalignment.
The classes with high resolution and similar features were combined for further classification or refinement.
Three dimension classification and refinement was performed with Relion1.2 (Scheres, 2012).
The map quality of specific regions was dramatically improved by focused classification and refinement (Fig. S5 and Table S3).
Therefore, after region-based 3D classification and refinement, a final data set of 42,000 nontop-view particles resulted in a 3.8 Å map (gold-standard FSC 0.143).
Micrographs inspection, automatic particle picking, 2D, 3D classification and refinement were performed within RELION 1.4 (Scheres, 2012) and the contrast transfer function parameters were estimated using CTFFIND3 (Mindell and Grigorieff, 2003).
Here we review some of the latest ribosome structures, where cryo-EM allowed gaining unprecedented insights based on 3D structure sorting with focused classification and refinement methods helping to reach local resolution levels better than 3 Å.
With the improved density map after "focus" classification and refinement, we also compared our reconstructed F-subcomplex with the three crystal structures of clathrin AP1 (the F-subcomplex homologue) in different conformations as well as the structure of membrane-bound F-subcomplex in the "hyper-open" state (Dodonova et al., 2015) to fit into the focus-refined density map.
4) 3D classification & refinement using Relion: $28.89 (10.32 hrs at $2.80/hr) This cost reflects of price of $0.35/hr/instance on a 128 CPU STARcluster, which comprises eight r3.8xlarge instances.
