[phenixbb] Missing map for model

[Pavel Afonine]

Hi Folmer,

I think I dealt with something similar a while ago. Unfortunately I 
don't remember the details but your description of the problem 
definitely rings the bells.... If I remember, the problem was some 
miscommunication between Autobuild and phenix.refine in terms of using 
incorrect data columns from input mtz file.... Using the right data was 
a solution.

I don't know the answer, but I would try the following which shouldn't 
take more than 5 minutes or even less.

Run phenix.refine to just compute the maps:

phenix.refine model.pdb data.hkl strategy=none main.number_of_macro_cycles=1

where data.hkl is your original file with Fobs.

This will do not do any refinement but will compute the maps. Now, the 
map coefficients file (xxx__map_coeffs.mtz) contains two kinds of maps: 
computed using original unmodified Fobs (labels ending with "_no_fill") 
and computed using modified Fobs, where missing reflections are filled 
with DFmodel. To see it, just run this command:

phenix.mtz.dump xxx_map_coeffs.mtz

So I would suggest to compare these two kinds of maps. If this doesn't 
give you any clue, the next question is: does your input data file (the 
one that is used in Autobuild) contain HL coefficients? If it does, then 
try why I suggested above but run phenix.refine slightly different:

phenix.refine model.pdb data.hkl strategy=none 
main.number_of_macro_cycles=1 use_experimental_phases=false

and have a look at the maps once again.

If this does not help, and no-one else suggests anything else to try 
out, you may want to send me  your input files and I will try to find 
what's going on. All files will be treated confidentially.

Pavel.


On 3/28/09 4:29 AM, Folmer Fredslund wrote:
> Dear all,
>
> In phenix.refine (1.4-3) my electron density maps look weird. I get
> beautiful density for a helix that is missing from the current model,
> but absolutely horrible density for the rest of the molecule. Any
> ideas what is going on? Maps from automr look fine and the density
> modified map from autobuild also look fine, whereas the autobuild map
> from refinement has the same features.
>
> Any ideas what is going on? I am running all programs with default parameters.
>
> Best regards,
> Folmer Fredslund
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://www.phenix-online.org/mailman/listinfo/phenixbb
>