[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[ccp4bb]: Deposition of Fcalc
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Dear Dirk,
I think the desire to deposit your own calculated structure factors is
perhaps wishing to exert too much control over your data. While not
disputing that researcher X may have done a very good job of refining
their structure, the freedom that researcher Y has to use another
refinement program and/or combination of corrections and weights may
well lead to an alternative interpretation of ambiguous density. The
info in the PDB header should be enough to reconstruct the maps as you
made them, or am I being naive?
In the spirit of "you never know when it might come in handy", why not
encourage the deposition of data with anomalous information (even
unmerged?), even if it was not used in the original experiment, since as
Eleanor and others frequently point out, many atoms have some anomalous
signals at wavelengths far from their edges and someone some day might
be interested in this. Then, since we are talking orders of magnitude
larger storage space, we might just as well have Fcalc in there as
well ;-)
Perhaps the main question is whether anyone apart from "structure
police" (apologies to Gerard!) will bother to use other people's data,
given that we're so busy with our own stuff. I've been meaning to look
at an old deposited data set for weeks now butnever got round to it. In
any case the possibility should be there in the deposition.
Derek
--
Derek Logan tel: +46 46 222 1443
Molecular Biophysics fax: +46 46 222 4692
Lund University
Box 124, Lund, Sweden
On Monday, November 18, 2002, at 09:10 AM, Dirk Kostrewa wrote:
> *** For details on how to be removed from this list visit the ***
> *** CCP4 home page http://www.ccp4.ac.uk ***
>
> On Friday 15 November 2002 17:54, Gerard DVD Kleywegt wrote:
>> a model deposited without data (or not at all) one has to take on faith
>> (i.e., more religion than science). for a model deposited with data,
>> you
>> can go and check the electron density yourself (e.g., using the uppsala
>> electron density server, EDS, a pre-release version of which is
>> available
>> at
>> http://fsrv1.bmc.uu.se/eds).
>
> On Friday 15 November 2002 19:06, Phil Evans wrote:
>> *** For details on how to be removed from this list visit the ***
>> *** CCP4 home page http://www.ccp4.ac.uk ***
>>
>>
>> On the question of depositing structure factors, this should be
>> compulsory (why isn't it?). There is no excuse for not depositing
>> them, and the most likely person to want them is you yourself.
>
> Dear Gerard and Phil,
>
> I think that in addition to deposit observed structure factors there is
> also a
> need to deposit the calculated structure factors _exactly_ as they came
> out
> of the scaling/refinement for the reported R-factors of the model. The
> reason
> for this is that if somebody takes the deposited model and the deposited
> observed structure factors and does not apply the proper anisotropic
> scaling,
> bulk solvent correction and weighting, the difference map will show
> features
> that are not consistent with the model. For that reason, personally, I
> still
> hesitate to publish observed structure factors until I can easily
> deposit the
> calculated structure factors and their phases _along_ with them. In
> principle, agree with both of you that these are data that should be
> published, but I must admit that I shy away from this until I can easily
> publish all the data.
>
> Best regards,
>
> Dirk.