Dear Sergii,

>refinement I checked a structure in Coot and all atoms of these 3 molecules
>were separated from each other and on much bigger distances then in DFIX
>instruction

In the header, you have set the connectivity to zero for residues 4001 to
4288. 

CONN 0 O_4001 > O_4288

This is also applied to your small molecules in the channel since in the
.ins file these residues are located in between water 4286 and 4287. The
CONN 0 instruction will activate the antibumping restraints for your LIM
atoms. Therefore they are pushed away from each other.

Either you can change the CONN (and ISOR) instruction and change the order
in the .ins file (move the LIM residues to the end). Or (as I would do it):
insert the LIM residues just after the protein chains, renumber the
waters/LIM and change CONN (and ISOR) accordingly.

>** MERG code changed to 0 for compatibility with HKLF and BASF parameters
>**
>I know, that if I use BASF-instruction I should have MERG 0, but I do not
>have it at all.

For a regular data set, the MERG 4 instruction should be included in the
header. Friedel opposites will be merged (and all f'' are set to zero).

Hope this helps. 

Best wishes, Tobias.


___________________________________________

Tobias Beck
Dept. Structural Chemistry
Georg-August University Goettingen
Tammannstr. 4
37077 Goettingen, Germany
phone:  +49 551 39-3068
fax:    +49 551 39-22582
net:    http://shelx.uni-ac.gwdg.de/tbeck/
___________________________________________


-----Ursprüngliche Nachricht-----
Von: CCP4 bulletin board [mailto:[email protected]] Im Auftrag von
Sergii Buth
Gesendet: Mittwoch, 9. September 2009 21:23
An: [email protected]
Betreff: [ccp4bb] Ignoring DFIX instructions in the refinement (SHELX)

Hello All,

I am working on a structure, which is triple helix with a channel inside
where are present some inclusion compounds.
I started to refine the structure of my protein in REFMAC, then I went to
ShelX to refine occupancy of double conformations and finish my structure.
But in ShelX I faced a few problems, and some of them I can not solve by
myself:

1) I have 3 unknown identical inclusion compounds in the channel. I built
the probable structure of the molecule from electron density map (crystal
diffracts up to 1.34 A), then generated for it an idealized PDB file and a
topology file. Then I fit this probable molecule in each of three positions,
found proper atom coordinates for each position, and updated with this
information main PDB file.
Then I refined (isotropically)... And now the problem starts... After
refinement I checked a structure in Coot and all atoms of these 3 molecules
were separated from each other and on much bigger distances then in DFIX
instruction. After Real space refinement in Coot (for this procedure I
imported REFMAC topology to Coot) the atoms stuck together for entire
molecule.
Then ->ShelX refinement-> Coot - and the same - I see separate atoms...

2) And I am wondering about one more warning:

** MERG code changed to 0 for compatibility with HKLF and BASF parameters **
I know, that if I use BASF-instruction I should have MERG 0, but I do not
have it at all.
But I also know, that "the reflections are always merged, and Friedel
opposites combined, before performing Fourier calculations in SHELXL so that
the (difference) electron density is real and correctly scaled."
Does it mean, that I should have MERG instruction anyway?

Please, find attached .ins file for more clarity. (Change extension to .ins,
'coze my proxy blocks it)

Thank you very much for your help,

Best regards, Sergii Buth

Reply via email to