[sharp-discuss] Re: problem with sharp

Clemens Vonrhein vonrhein at globalphasing.com
Tue Oct 26 13:44:26 BST 2004


Dear Cholien,

there is some documentation about the different phasing scenarios:
please see e.g.

  http://www.globalphasing.com/sharp/manual/chapter2.html#scenarios

Now, one important thing you have to keep in mind is the basic
difference between an autoSHARP run and a SHARP run (both can be
started from the main SHARP Control Panel - which is what you see when
you connect to the httpd server):

  - SHARP is doing purely heavy atom refinement and phasing (can be
    followed by density modification etc). You can accomodate any kind
    of phasing scenario (so e.g. also MAD+native).

  - autoSHARP tries to do everything from the merged data onwards
    (including scaling, heavy atom detection, refinement, phasing,
    density modification etc). Due to this automation, however, you
    can only do the more basic phasing scenarios: SAD, MAD, SIR(AS),
    MIR(AS).

Your case seems to be a two-step process:

  1. get to a complete and refined heavy-atom model for the Se-Met
     dataset(s)

  2. get best map possible

The first step (since it is mainly concerned with the Se atoms) can
easily be done with your three MAD wavelengths alone. This means you
can use autoSHARP (MAD) with some very simple input. One note though:
the heavy-atom detection in autoSHARP is not as powerful as other
available algorithms. Therefore, if autoSHARP doesn't find the sites
(or it finds 'funny' ones): just create a file
sharpfiles/datafiles/Something.hatom (format is explained in the
manual) with the fractional coordinates of your heavy atoms (e.g. from
SHELXD, SnB, Solve, ... whatever). Then select this file as known
heavy atom positions in the autoSHARP control panel.

For the second step you might want to add the native (high-resolution)
dataset: after your (hopefully successful) autoSHARP run you can use
the final SIN file (which should have the complete heavy-atom model
already in the correct hand) and use your MAD+native mtz file. But be
careful about the Se-S issue (see documentation)!

I would highly recommend doing an autoSHARP MAD run with a set of
known heavy atoms (e.g. from Solve). And make sure to have a look at
any message/warning it presents you!

Cheers

Clemens

On Tue, Oct 26, 2004 at 08:26:49PM +0800, ?????? wrote:
> Dear Sir,
> 
> 
> 
> Thank you for replying. I input the MTZ file which is merged native, peak, 
> edge and remote data. I have no idea would I select "MAD" or "MAD+native" 
> model? I will rerun SHARP follow your suggestion about spacegroup H3/H32.
> 
> 
> 
> I have a result (PDB) from solve/resolve but the PDB file lost many 
> residues so now I use SHARP to solve and improve  MAD phasing. The resolve 
> result show the structure contains 12 Se sites. I have been used SHARP to 
> input 2-3 Se sites each SHARP experiment because I am not sure these Se 
> sites are all correct. Besides, I want to obtain more probably Se sites by  
> judging FOM value and residual map of SHARP. And I have been used autoSHARP 
> in MAD model that result show the structure contains 2 Se sites. The result 
> of autoSHARP is difference from solve/resolve. Would you give me any 
> recommend?
> 
> 
> 
> Thanks in advanced.
> 
> 
> 
> 
> 
> cholien.
> 
> 

-- 

***************************************************************
* Clemens Vonrhein, Ph.D.     vonrhein AT GlobalPhasing DOT com
*
*  Global Phasing Ltd.
*  Sheraton House, Castle Park 
*  Cambridge CB3 0AX, UK
*--------------------------------------------------------------
* BUSTER Development Group      (http://www.globalphasing.com)
***************************************************************
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://www.globalphasing.com/pipermail/sharp-discuss/attachments/20041026/c590bb21/attachment.pgp


More information about the sharp-discuss mailing list