[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ccp4bb]: FALLOFF warnings from truncate
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Martyn Symmons wrote:
>
> *** For details on how to be removed from this list visit the ***
> *** CCP4 home page http://www.ccp4.ac.uk ***
>
> Dear Pedro
> once you have a model then refinement programs can apply an
> anisotropic correction to the data to get a good fit between Fcalc and
> Fobs - they usually do this on the fly so your data stays unchanged. If
> you do want to make it stick then you can use SFALL to calculate Fcalcs
> and then use SCALEIT to scale the FPs to these Fcalcs (by telling it that
> your native is a heavy atom derivative...). Or if you have some other
> estimate of the Bfactor matrix (say whatever the refinement program tells
> you _it_ applied) then you can apply it directly in SCALEIT. I guess that
> this is looked on as dubious practice since you are cooking up new FPs -
> but of course SCALEIT keeps track of the sigmaFPs and so the larger
> relative errors in the weak data keep everything in proportion....
> I have to admit to having done this before SIGMAA and DM which
> don't do anisotropic corrections for themselves. But that is just between
> me and you, ok? (oops - should have edited that Cc line....)
> good luck with everything
> Martyn
>
Yes you should have done that! But Steve Gamblin swears by this..
However no amount of scaling up produces useful data if you didnt
measure it..
Most software is sort of self correcting..
Exptl phasing usually cant handle the weak unreliable observations
anyway, so you effectively fdont get any information for the highly
anisotropic reflns.
Refinement corrects for it, and extracts what it can..
Molecular replacement may benefit - the model is usually isotropic, and
any procedure which uses normalised data - shelxds/shelxd / acorn/ etc
cant handle aniso data very well..
Eleanor