[ccp4bb]: Missing Domain Search

[Martyn Symmons <martyn@cryst.bioc.cam.ac.uk>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***

Dear Tristan
     - one possibility is that the 'lost' domains do not conform to all
the symmetry operations of your spacegroup - so you made a 'mistake' (so
far as those domains were concerned) when you merged all the reflections.
Did you have a 'difficult' choice between a lower and higher symmetry
spacegroup during data reduction? The down side of exploring this is you
will need a model for an 'expanded' asu with multiple copies of the 'good'
domains which maybe dominate the data. If you decide to proceed this way
you must make sure your Rfree set contains all 'pseudo'-symmetry related
measurements otherwise the refinement program gets a 'sneak preview' of
the Rfree dataset from unflagged data that are highly correlated with it.
And you will have to set very strict NCS on the 'good' domains. So if your
maps are not noticeably better calculated from the lower symmetry group
then it is not worth following up.
     - I remember Prof. Lawrence Pearl describing this happening for a
DNA:protein complex where too high a symmetry was imposed that only one
component really satisfied.
     - I would be interested to hear if anyone has had this happen with
multidomain proteins.
         - good luck
                    Martyn
  On Thu, 20 Mar 2003, Tristan Fiedler wrote:

> Date: Thu, 20 Mar 2003 14:56:12 -0500
> From: Tristan Fiedler <tfiedler@newssun.med.miami.edu>
> Cc: ccp4bb@dl.ac.uk
> Subject: [ccp4bb]: Missing Domain Search
>
> ***  For details on how to be removed from this list visit the  ***
> ***          CCP4 home page http://www.ccp4.ac.uk         ***
>
> Dear CCP4 a/o CNS users,
>
> We are working with a MAD dataset with phase information extending to
> 2.0 Angstrom resolution which has allowed unambiguous modelling of 250
> out of 350 residues.   Current R/Rfree are high 20's/low 30's without
> modeling any waters.
>
> The problem is that the N-terminal domain of ~100 residues can not be
> located visually in the density using O. Homologs of the N-terminal
> domain exist in the PDB.  Testing CNS:rigid.inp using a homolog was
> unsuccessful.
>
> Any suggestions as to how to locate the missing N-terminal domain would
> be appreciated.  We realize it may (most likely is) disordered and no
> density may exist in the maps.
>
> Thank you in advance,
>
> Tristan Fiedler
> Mark Del Campo
> --
> Tristan J. Fiedler, Ph.D.
> Malhotra Laboratory
> Department of Biochemistry & Molecular Biology
> University of Miami Medical School
>
> t.fiedler@umiami.edu
> tfiedler@newssun.med.miami.edu
>

Martyn F. Symmons
Research Associate
Crystallography and Biocomputing Group
Department of Biochemistry
University of Cambridge
Phone: 01223 766020
-----------------------
Am fear a tha grad gu gealladh, 's tric leis mealladh
Seanfhacal Gaidhlig (Gaelic Proverb).