Garib Murshudov schrieb:
> What is the version of refmac you arre using. There was a bug and I 
> think we have fixed it. If you take the version from 
> www.ysbl.york.ac.uk/refmac/latest_refmac.html it may give consistent 
> results. Only mac and linux versions are available. For windows 
> version you will need to take from the latest ccp4 version (6.1.2 
> should have the newer version of refmac also)
>
>
> regards
> Garib
>
> On 20 Oct 2009, at 15:34, Gesa Volkers wrote:
>
>> Dear Ccp4 Users,
>>
>> I have refined a 2.3 A structure with Refmac in space group I4(1)22. 
>> Lately I had to switch from refmac version 5.2.0019 running on a 
>> Linux computer to refmac version 6.0.2 on a Windows Vista system. 
>> Refmac works fine but I’m a little concerned about the fact that 
>> R/Rfree/FOM became really worse from 19.5/25.7/80.0 (older version) 
>> to 21.5/27.3/77.1 (newer version) just by changing the refmac 
>> version. Also the correlation coefficients became worse. The overall 
>> temperature factor was significantly lower. I observed this also with 
>> another newer version on a Linux system. I strictly used the same 
>> protocol with TLS & restrained refinement using Babinet scaling 
>> without changing anything in the structure.
>> Are there other restraints for bonds etc. in the different versions 
>> leading to such differences? I would like to know what is going on.
>> I would appreciate your help or any hints!
>>
>> Thank you in advance,
>>
>> Gesa
>> -- 
Thank you, Garib.

The old version of refmac which gave good results, was 5.2.0019. The 
vista version with worst results of the same data was 5.5.0088. Another 
linux version, which gave worse results than the older linux version was 
5.5.0047.
Then I will install ccp4 6.1.2 on the vista system and see what happens.

Regards,
Gesa


-- 
Gesa Volkers
Institut für Biochemie, Molekulare Strukturbiologie
Felix-Hausdorff-Straße 4
17489 Greifswald
03834/864392