[buster-discuss] How to do occupancy refinement of alternative conformations with associated waters in buster

[Clemens Vonrhein]

Dear Stefan,

On Wed, Aug 05, 2015 at 05:14:58PM +0200, Stefan Gruener wrote:
>    I am currently building a structure at 1.5A resolution and have
>    modelled an Arginine residue with 2 alternative conformations, giving
>    occupancies of 0.52 vs. 0.48. However, there is some positive
>    difference density remaining (see attached file, contour level is 1.0
>    sigma for 2Fo-Fc and 3.0 sigma for Fo-Fc) which to me may suggest the
>    presence of waters with occupancies matching the respective alternative
>    Arg conformations.

Yes.

>    Is there any way to group these waters together with their respective
>    arginine conformer in BUSTER refinement?

Yes.

Let's assume that in your picture we see the A-conformation (occ=0.52)
at the left and the B-conformation (occ=0.48) to the right. You would
then

 * add a HOH in the left positive Fo-Fc density and assign this to
   conformation B with occ=0.48 (you can do this easily in Coot with
   "Edit -> Residue Info...")

 * add a second HOH into the right Fo-Fc peak with conformation A and
   occ=0.52.

The resulting PDB file can be run through pdb2occ with

  pdb2occ -p refine-coot-0.pdb -o refine-coot-0.occ

In that output gelly file the two Arg conformations are handled
correctly (sum of occupancies restraint to one). But the two waters
are treated independently - which is what you will need to edit. First
find the two AltOccNN groups for the Arg, e.g.

 NOTE BUSTER_SET AltOcc27   = { A|442:*.A}
 NOTE BUSTER_SET AltOccAll  = AltOccAll    + AltOcc27  
 NOTE BUSTER_COMBINE OCC AltOcc27

 NOTE BUSTER_SET AltOcc28   = { A|442:*.B}
 NOTE BUSTER_SET AltOccAll  = AltOccAll    + AltOcc28  
 NOTE BUSTER_OCCSUM 1.0 0.005 A|442:N.A A|442:N.B
 NOTE BUSTER_COMBINE OCC AltOcc28

Then find the two waters, e.g.

 NOTE BUSTER_SET AltOcc37   = { W|101:*.B}
 NOTE BUSTER_SET AltOccAll  = AltOccAll    + AltOcc37
 NOTE BUSTER_COMBINE OCC AltOcc37

 NOTE BUSTER_SET AltOcc38   = { W|102:*.A}
 NOTE BUSTER_SET AltOccAll  = AltOccAll    + AltOcc38
 NOTE BUSTER_COMBINE OCC AltOcc38

You want to remove hose two last groups (here AltOcc37 and AltOcc38)
and instead add the two waters to the groups that already contain
your Arg:

 NOTE BUSTER_SET AltOcc27   = { A|442:*.A}
 NOTE BUSTER_SET AltOcc27   = { W|102:*.A} + AltOcc27
 NOTE BUSTER_SET AltOccAll  = AltOccAll    + AltOcc27
 NOTE BUSTER_COMBINE OCC AltOcc27

 NOTE BUSTER_SET AltOcc28   = { A|442:*.B}
 NOTE BUSTER_SET AltOcc28   = { W|101:*.B} + AltOcc28
 NOTE BUSTER_SET AltOccAll  = AltOccAll    + AltOcc28
 NOTE BUSTER_OCCSUM 1.0 0.005 A|442:N.A A|442:N.B
 NOTE BUSTER_COMBINE OCC AltOcc28

So now you have a group AltOcc27 for your A conformation of (Arg+HOH)
and a second group AltOcc28 for your B conformation of (Arg+HOH). The
occupancy of atoms within each group will be refined together while
the sum is restrained to add up to one. Please note that all atoms in
A- and B-conformation should start with the same occupancy (see
above).

Hopefully there are no typos in the above examples. You can then run
it with

  refine -p refine-coot-0.pdb -Gelly refine-coot-0.occ ...


See also

  https://www.globalphasing.com/buster/manual/gelly/manual/gelly3.html
  https://www.globalphasing.com/buster/manual/gelly/manual/gelly5.html

>    I tried some naive modifications in the gelly file but that just gave a
>    refine-0057 error when trying to run BUSTER.

Maybe those edits were somehow wrong? Can you send the diff between
the output from pdb2occ before and after your edits?

Does the above 'recipe' work for you?

Cheers

Clemens Vonrhein & Claus Flensburg

-- 

***************************************************************
* Clemens Vonrhein, Ph.D.     vonrhein AT GlobalPhasing DOT com
*
*  Global Phasing Ltd.
*  Sheraton House, Castle Park 
*  Cambridge CB3 0AX, UK
*--------------------------------------------------------------
* BUSTER Development Group      (http://www.globalphasing.com)
***************************************************************