It is tricky, but maps can work.
My procedure was:
Extend map file to P1
Put a dummy "atom" at the position you feel is about in the centre of
the molecule density.
Make a mask around this atom with a radius of nA which you feel is big
enough to cover most of your molecule.
Use this mask to mask the P1 map, ie set all points outside mask to 0.
Shift the density into a big cell - although maybe the I222 cell
dimensions with only one molecule in it will be big enough already.
Turn map into SFs - sfall will do it, and i am sure there are other
ways too..
Use those SFs as input to MR.
get a big headache trying to work out how to shift your rough model to
the new cell - what fractional shift to apply etc etc!
Eleanor
Edward A. Berry wrote:
> Heidi Schubert wrote:
>>
>> Hi
>>
>> I would like to convert a map into SFs to use for molecular
>> replacement. Read on if you think you can help.
>>
>> I have a 3.7A MAD data set and a few 30% similar homology models for
>> crystal in i222 space group. The combination of the two pieces of
>> information generated solvent flattened phases which reflect unique
>> information and have a FOM of 0.65. Workable, but certainly any model
>> I build is inherently poor.
>>
>> I have new data for the same protein (2 mols in the ASU now) in C2
>> space group with data to 2.1A. The poor models (original models and
>> an attempt at a "better" built model from the 3.7A phases) do not
>> work as molecular replacement solutions in Phaser or MolRep. It would
>> be better if I could use the i222-3.7A map for the molecular
>> replacement into the new space group.
>>
>> PHASER manual says "This density has been cut out and converted to
>> structure factors in a large cell"
>> So, how do I do that?
>>
> Randy J. Read wrote:
>> The procedure of cutting out electron density, putting it into a
>> large unit cell, and backtransforming to get structure factors can be
>> tricky (as you've discovered), so we put some instructions on our
>> webpage:
>>
>> http://www-structmed.cimr.cam.ac.uk/phaser/density_as_model.html
>>
>> The last time I tried that it worked, but let me know if you run into
>> any problems with your case.
>>
>> Randy Read
>
> HTH
> Ed
>
>
|