[phenixbb] Gold atoms in cryo-EM

[Pavel Afonine]

Hi Ali,

> I am trying to refine a large protein structure against a cryo-EM map using Phenix real-space refinement. The protein complex is unusual in that the oligomeric state is held together via coordinated gold atoms. However, there seems to be a problem with refinement, in that the the gold atoms show extremely high B-factors (>200) after refinement, whereas the B-factors of the protein residues, including those surrounding the gold atoms, adopt more reasonable values.
>
> My question is: could the unusually high B-factors of the gold atoms be somehow due to differences in the relative scattering of Au atoms when obtained from X-ray vs. electron beam irradiation? And if so, is there any way to correct for this?

- have you tried re-setting B-factors of gold to small(er) values (say 
similar to surrounding atoms) and then re-running refinement?

- did you make sure to use 'electron' scattering table (default is for 
X-rays)? Use scattering_table=electron for this.

- have you tried to convert the map into structure factors (generally 
not recommended! but in this case worth trying as part of investigation) 
and then use B-factor refinement (only) (so it runs faster) in 
phenix.refine?

If none of the above gives a clue then I may need to investigate deeper 
and that will likely require me having all relevant files (map, model etc).

Pavel