Re: [ccp4bb] Could I improve an inconsecutive experimental map?

[Eleanor Dodson]

I presume you have done density modification after calculating the exptl 
phases?
If you use REFMAC to refine your partial model with the exptl/density 
modified phase restraints your FWT PHWT fourier coefficients already use 
the combined phase.

I am not sure how to choose a damping factor. If you think the exptl/dm 
phases are good and have realistic weights  I guess there shouldnt be one..
SHARP claims to get the weights more or less right; parrot is meant to 
do a better job of weighting than DM.
Eleanor




Fengyun Ni wrote:
Hi,

I have tried SHARP, but no positive results could be obtained.

Do you have some articles on how to combine the model phase and experimental
phases?
I could only seen the consecutive density in mir 8 to 6 A resolution map, so
I could only locate roughly the position of helix.
After that I use this helix to calculate the structure factor in SFALL.
In SIGMAA, what's the DAMP coefficient should I use?
After combination, do I need to do density modification?
Hope for your help!

Regards,
Alphar


2009/2/21 Charlie Bond <charles.b...@uwa.edu.au>

  
Hi Alphar,
A while ago I had success with a poor MIR map by building in a few
fragments I could identify and then combining phases from this model with
the MIR phases using SIGMAA (CCP4 program) and then calculating new maps. A
couple of cycles of this resulted in phases in which I had confidence and I
was able to build the complete structure. (There's probably a more modern
way to do this now ...)
Cheers,
Charlie


Quoting Fengyun Ni <alphar...@gmail.com>:

Thanks, Ezra!
    
I have a native dataset.

I'll try what you said, Hope it works in my case.

Thanks!

Alphar

2009/2/21 Ezra Peisach <epeis...@gmail.com>

Personally, I would try to build into what you see and then do
      
phase
      
recombination.  I had a SeMet crystal in which I could only see
      
part of a
      
sheet and helix after DM... Things were not connected - until I
      
improved
      
  the phases with my model.   Also - do you have a native dataset?
      
You do not
      
mention it.
      
Ezra



Fengyun Ni wrote:

Hi everyone!
        
I have a question on my poor MIR map.  Four datasets (two AU and

          
two HG)
        
 were used for phasing upto 3 A resolution in my case. I could  locate
      
several
heavy atom sites for each dataset with occupancy finally refined

          
to about
        
 0.3 to 0.4, and with B value refined to about 50 to 80. I did not
      
include
        
 the anomalous data because once I refine against anomalous data,
      
the
        
 anomalous occupancy was only about 0.02, and even smaller as
      
0.004 for some
        
 sites. I guess the anomalous data are not good enough, so I did  not use
      
them
right now (Could I do this? Or must I include them somehow?).
  The FOM I got is about 0.6, the Cullis-R factors were about 0.6

          
for all
        
 four data sets, and the phasing power was about 2.5 in each  dataset.
      
After I
did the density modification, the FOM could increas to about 0.8.
  My problem is that the protein should form a long helix

          
structure as
        
 indicated by other homology protein, but in the map after density
      
modification, the densities are not consecutive though the

          
overall shape
        
 seemed to be a long helix. The good thing is that some short
      
helix-like
        
 densities could be observed in the current map, but they are not
      
connected
        
 to each other.
      
  Right now, I am totally lost whether I should believe in the

          
mir-map I
        
 have. Could I improve this map with some other method? Or should
      
I
        
 re-do the
      
phasing?
BTW, the space group for my crystal is R3 or R32 with a=50, b=50,

          
c=380,
        
 alpha=90, beta=90 and gamma=120. The c-axis is much longer than
      
a- and
        
 b-axis, so the reflection data suffered from the anisotroy
      
effect.
        
   Any suggestion on my phasing problem is welcome.
      
--
Alphar Ni
Call me alphar~
:)


          
        
--
Fengyun Ni
Call me alphar~
:)


      
--
Charlie Bond
Professorial Fellow
University of Western Australia
School of Biomedical, Biomolecular and Chemical Sciences
M310,  35 Stirling Highway
Crawley WA 6009, Australia