![](https://secure.gravatar.com/avatar/9a025121d42a94a7cfb1d3f1d74b6ffb.jpg?s=120&d=mm&r=g)
Dear All, I first run a default phenix.real_space_refine with the map and initial model, with the model got, I run phenix.real_space_refine with run=adp, I find for the second phenix.real_space_refine step with run=adp, it can gives a reasonable value of b-factors for each residues in the model, and it cannot modify the coordinates of any atoms in the model (in comparison with the model got by the default phenix.real_space_refine in the first step). I just read a Nature paper from David Baker group "Atomic-accuracy models from 4.5-Å cryo-electron microscopy data with density-guided iterative local refinement", it says "Because cryo-EM maps are frequently better in resolution in some regions than others, atomic B factors are fit against cryo-EM density data to maximize the real-space correlation between model and map". For Real-space B-factor refinement, the math functions used by that Nature paper include, ( f is a scattering factor fit to each element) My question is, does b-factor refine only give a reasonable b-factor to each atom or residue in the model (so that we know the displacement of each atom or residue), or besides giving reasonable b-factor, it can also change and optimize the x y z coordinates of each atom in the model, and the phenomenon I observed in phenix.real_space_refine on that it only gives reasonable b-factor without modification of the x y z coordinates was only limited to the case I see in phenix.real_space_refine? Best regards. Smith