[phenixbb] coordinate refinement and maps

Nathaniel Echols nechols at lbl.gov
Tue Nov 20 12:47:36 PST 2012

On Tue, Nov 20, 2012 at 12:41 PM, Subhani Bandara <ramssb17 at gmail.com> wrote:
>  I am refining  a protein with metal-ligand complexes. Is it possible to
> refine only y coordinates for an atom (on two fold)  while keeping X and Z
> fixed using Phenix? Also is there a way to calculate something like
> anomalous difference-difference map (Anomalous Fobs - Anomalous Fcalc) to
> identify any additional atoms bound at metal sites.

Not sure about the first question, but the answer to the second is
yes, if you download the latest nightly build, phenix.refine is now
able to output an anomalous log-likelihood gradient (LLG) map from
Phaser, which will indicate where any unmodelled anomalous scattering
is present.  This is most useful if you use the group_anomalous
strategy with any heavy atom selected (I would put S in this list),
and I suspect you need to refine them separately for optimal results
(i.e. each metal is its own group, rather than one group with the
selection "element FE" or whatever it is).  You can also get an LLG
map from phenix.maps, but this simply guesses the anomalous scattering
based on the wavelength, and will probably be noisier than if you
refine f' and f''.


More information about the phenixbb mailing list