Dear Garib,

Thank you very much. I will try the keyword.

> Your case: Do you see some sort of modulation of intensities in your images?
> Something like weak strong intensities along c axis? Or do you see elongated peaks in the images.
Sorry I'm not sure what "modulation of intensities in image" is.
Do you mean something like described in this paper?
http://scripts.iucr.org/cgi-bin/paper?S0021889808010716
direct link to figure:
http://journals.iucr.org/j/issues/2008/03/00/he5398/he5398fig4.html

My diffraction images don't have such reflections.
But in some frames, severely elongated peaks along c* were observed.
Moreover, I found split peaks at high resolution (about 2.5A) area in
frames without elongated peaks.

I think my diffraction image is similar to what was discussed recently
in ccp4bb:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1011&L=CCP4BB&F=&S=&P=312247
you can see the attached files of that archive.


Yours truly,

K. Yamashita

2010/12/9 Garib N Murshudov <[log in to unmask]>:
> Dear Yamashita
>
> Refmac indeed does not apply vdw repulsions (antibumping restraints) if the sum of occupancies is less than 1 and atoms do not belong to the same residue with the same alt code. It means that if you have two or more molecules each with occupancy less than 0.5 then there will be no intramolecular antibumping restraints also. It is not what you would like. In these case you may want to apply antibumping restraints within molecule. For this reason using keyword
>
> (if you have three copies and chain names are A B and C)
>
> vdwrepulsions exclude between chains A B C
>
> would be better. Then the program would apply antibumping restraints within molecule but not between them.
>
> Your case: Do you see some sort of modulation of intensities in your images? Something like weak strong intensities along c axis? Or do you see elongated peaks in the images.
>
> regards
> Garib
>
>
>
>
> On 9 Dec 2010, at 01:11, Keitaro Yamashita wrote:
>
>> Dear Garib,
>>
>> Sorry I'm a little confused.
>>
>> As Eleanor said, Is it true that Refmac doesn't apply vdw restraints
>> if the sum of the atoms occupancies is <= 1?
>>
>>> If you want to deal with translational disorder (e.g. you have a modulated crystal then there is another way.
>> In my data, there are crystallographic 4-fold axis on c-axis (P4 21 2)
>> and pseudo-translation vector (0.1, 0.1, 0.5), which is indicated by
>> native patterson peak with 22% height of the origin.
>>
>> I can see very clear density at certain position around the 4-fold axis.
>> And around the shifted position (+0.1, +0.1, +0.5), there's definitely
>> the same density but they are superposed with their symmetry mates --
>> I think  they are statically (packing) disordered.
>>
>> Is my case "translational disorder" as you said?
>> If so, how can I solve it?
>>
>>
>> Yours truly,
>>
>> K. Yamashita
>>
>> 2010/12/9 Garib N Murshudov <[log in to unmask]>:
>>> In refmac you can remove vdw interactions between chains using the following command
>>>
>>> It is an example:
>>>
>>> vdwr exclude between chains A B
>>>
>>> or between resdues:
>>>
>>> vdwr exclude between residues first residue 123 chain A second residue 155 chain B
>>>
>>>
>>> Regards
>>> Garib
>>>
>>>
>>> On 8 Dec 2010, at 16:09, Keitaro Yamashita wrote:
>>>
>>>> Dear all,
>>>>
>>>> I'm refining complex structure against X-ray diffraction data with
>>>> packing disorder.
>>>> (Some domains overlap with their symmetry mates (4-fold), so their
>>>> occupancies are set to 0.25)
>>>>
>>>> I'd like to know whether refinement programs can exclude any
>>>> interaction among symmetry mates from geometric term in target
>>>> function.
>>>> Can it be done only for specific chains?
>>>>
>>>> I was thankfully taught phenix.refine could do that with the option
>>>> pdb_interpretation.custom_nonbonded_symmetry_exclusion in phenixbb.
>>>>
>>>> I'd very much like to know whether Refmac5, CNS or other programs can do that.
>>>>
>>>> (When I tried to refine using Refmac5, the output told many vdw
>>>> repulsions with symmetry mates.)
>>>>
>>>>
>>>> Thank you very much in advance,
>>>>
>>>> K. Yamashita
>>>
>>>
>
>