[buster-discuss] Refine occupancy on two different polypeptide chains

[Clemens Vonrhein]

Dear Anna,

On Thu, Feb 01, 2018 at 12:14:54PM +0100, Anna Suarez Larsson wrote:
> > What happens if you remove that line (NOTE BUSTER_CONSTANT OCC FixOcc)
> > from your Gelly file?
> 
> I got a long list of errors in the sanity check about ‘contact
> between two ‘Protein’ type atoms from distinct residues’ so I added
> 'StopOnGellySanityCheckError=no’ to the command line to test.

That's always a possibility - but most of the time those sanity check
failures point to some modeling problem.

> The refinement then pushed the chains apart and the R-values rised.

Ah: so these two chains 'see' each other. Did you maybe not use
different altConf (alternate conformation) identifiers for your A and
B chain? If your A and B chain atoms have no altConf identifier then
the contact term is active and that's why they are getting pushed
apart.

There are two possibilities:

 (1) change all atoms in chain A to have altConf "A" and all atoms
     in chain B to have altConf "B" (see e.g. 2XGC or 2XGE)

 (2) add an EXCLUDE card to tell BUSTER to ignore contacts betwene
     chain A atoms and chain B atoms (see below)

> > You should be able to do that, although using pdb2occ might not be the
> > most elegant way: this tool is mainly intended to handle occupancy
> > refinement situations with some small(ish) localised items (alternate
> > conformation side-chains, partially occupied ligands, alternate loop
> > conformations etc).
> 
> Do you have any suggestions where to find documentation of how to do it otherwise?

You should be able to use a file (for -Gelly flag) containing

  NOTE BUSTER_RESET_CONSTANT_COMBINE
  NOTE BUSTER_COMBINE OCC { A|* }
  NOTE BUSTER_COMBINE OCC { B|* }
  # we want to restrain OCC(A)+OCC(B)=1:
  NOTE BUSTER_OCCSUM 1.0 0.005 A|1:CA B|1:CA
  # exclude contacts between chains
  EXCLUDE A|* B|*

This assumes that you have no altConf identifier, that you have a
residue A1 and B1 (to define occupancy sum restraint) and that you
have only chain A and B atoms in your model (with all A and B atoms at
the same starting occupancy).

See also

  https://www.globalphasing.com/buster/manual/gelly/manual/gelly3.html#buster_combine
  https://www.globalphasing.com/buster/manual/gelly/manual/gelly5.html#occsum
  http://www.uoxray.uoregon.edu/tnt/manual/node29.html#432

Does that work?

Cheers

Clemens

> 
> > 
> > Please let us know if you hit another issue.
> > 
> > Kind regards
> > 
> > Clemens & Andrew (for buster-develop)
> 
> Thank you for your answer!
> 
> Best regards,
> Anna
> 
> *****************************************
> Anna Larsson
> Laboratory of Molecular Biophysics
> Department of Cell and Molecular Biology 
> Uppsala University
> Box 596 
> SE-75124 Uppsala, Sweden
> 
> Tel. +46-18-4714983
> E-mail: anna.larsson at icm.uu.se
> 

-- 

*--------------------------------------------------------------
* Clemens Vonrhein, Ph.D.     vonrhein AT GlobalPhasing DOT com
* Global Phasing Ltd., Sheraton House, Castle Park 
* Cambridge CB3 0AX, UK                   www.globalphasing.com
*--------------------------------------------------------------
_______________________________________________
buster-dev mailing list
buster-dev at globalphasing.com
https://wwi.globalphasing.com/mailman/listinfo/buster-dev
-- 

*--------------------------------------------------------------
* Clemens Vonrhein, Ph.D.     vonrhein AT GlobalPhasing DOT com
* Global Phasing Ltd., Sheraton House, Castle Park 
* Cambridge CB3 0AX, UK                   www.globalphasing.com
*--------------------------------------------------------------