[phenixbb] Clearing old NCS group definitions
Nathaniel Echols
NEchols at lbl.gov
Wed Dec 9 03:54:50 PST 2009
On Dec 8, 2009, at 11:00 PM, John Pascal wrote:
> I am using NCS restraints in phenix.refine, and I am defining the restraints in
> a parameter file that I read into the GUI interface. After modifying the ncs
> parameters file, I see in the log file that phenix.refine has used the updated
> NCS groups, but also retains the old restraints settings. Therefore, the
> number of NCS groups doubles each time I make an edit. Is there a way to erase
> the previous NCS restraints?
There are some quirks in the way the GUI handles parameter files that are input as separate files - specifically, they will not overwrite existing parameters. (There is a good reason for this, but I can't remember it at the moment.) In general, I don't recommend setting parameters this way, because the parameters are used internally to draw controls and there are some built-in limitations to how flexibly it can handle different sources. Editing files externally that are used by the GUI is also problematic. I would recommend using the built-in graphical NCS restraint editor, especially since this will allow you to save default restraints so you don't have to re-load them in the future. If you want it to start with the restraints in your file, you have two options:
1. Drag-and-drop the parameter file onto the phenix.refine button in the main GUI - this will open a new window with the NCS restraints already loaded. Alternately, if you start phenix.refine_gui from the command line, add the file to the argument list. (In general, these are the safest methods for supplying parameters from a separate file.)
2. Copy the file to <project_dir>/phenix/defaults/ncs.params, and it will be loaded whenever the GUI starts.
> I see that the restraints are accumulating in the project_data/*.eff files. Are
> those read in by default on consecutive runs?
No, the only time these are used after the job is complete is when you restore the job in the main GUI.
-Nat
--------------------
Nathaniel Echols
Lawrence Berkeley Lab
510-486-5136
NEchols at lbl.gov
More information about the phenixbb
mailing list