[buster-discuss] Help needed : constrained occupancy refinement

[Clemens Vonrhein]

Hi Jose,

On Mon, Mar 06, 2017 at 03:36:18PM +0000, Jose Artur Brito wrote:
>    Dear All,
> 
>    I'm currently refining a structure with BUSTER where in the active site
>    I have a thiosulfate molecule (D1) and a sulfide molecule (E1). I want
>    to refine the occupancies of both molecules but constrain the sum of
>    them to 1 (i.e., I either have a sulfide or a thiosulfate in the active
>    site). I'm running the -Gelly command with the following cards in a
>    .occ file:
> 
>    NOTE BUSTER_RESET_CONSTANT_COMBINE
>    NOTE ============= alternate conformations ===============================
>    NOTE ------------- AltOccAll contains all atoms in alternate conformations
>    NOTE BUSTER_SET AltOccAll = Empty
> 
>    NOTE BUSTER_SET AltOcc1    = { D|1:*.A}
>    NOTE BUSTER_SET AltOccAll = AltOccAll + AltOcc1
>    NOTE BUSTER_SET AltOcc2    = { E|1:*.B}
>    NOTE BUSTER_OCCSUM 1.0 0.005 D|1:*.A E|1:*.B
>    NOTE BUSTER_SET AltOccAll = AltOccAll + AltOcc2
>    NOTE ------------- fix all non-refined occupancies
>    NOTE BUSTER_SET FixOcc = All
>    NOTE BUSTER_SET FixOcc = FixOcc \ AltOccAll
>    NOTE BUSTER_CONSTANT OCC FixOcc
> 
>    BUSTER complains with:
>    Have found    2 'NOTE BUSTER_OCCSUM' cards
>     =====NOTE BUSTER_OCCSUM 1.0 0.005 S|243:O.A S|243:O.B
>     =====NOTE BUSTER_OCCSUM 1.0 0.005 D|1:S2.A E|1:*.B
>     *** ERROR Have NOTE BUSTER_OCCSUM card:
>     NOTE BUSTER_OCCSUM 1.0 0.005 D|1:*.A E|1:*.B
>     *** ERROR atom descriptor 'D|1:*.A'
>     *** ERROR cannot find atom??

Is this the error message when running with the above Gelly file? It
seems different: there is no BUSTER_OCCSUM card involving atom
"S|243:O.A" in the above Gelly file although it is reported here ...

>    What am I doing wrong?? Even if I change the "*" with an atom
>    identifier (e.g., S2 in thiosulfate), it complains the same ("cannot
>    find atom").

Your BUSTER_OCCSUM needs to explicitly reference atoms, not groups of
atoms. So something like

  NOTE BUSTER_OCCSUM 1.0 0.005 D|1:S2.A E|1:S1.B

if you have an atom " S2 " with altConf "A" in residue D1 and an atom
" S1 " with altConf B in residue E1. Remember that you use

  <Chain>|<ResNum>:<Atom>.<altConf>

for atom specification. When you use

  <Chain>|<ResNum>:*.<altConf>

it will refer to all atoms that match the remainder of the pattern
(which can be multiple atoms in your case).

If your thiosulfate is THJ (https://www3.rcsb.org/ligand/THJ) you
would have something like

  HETATM    1  S1 ATHJ D   1      -0.407   0.000  -0.000  0.50 10.00           S
  HETATM    2  O1 ATHJ D   1      -0.826  -1.393   0.012  0.50 10.00           O
  HETATM    3  O2 ATHJ D   1      -0.826   0.686  -1.212  0.50 10.00           O
  HETATM    4  O3 ATHJ D   1      -0.826   0.707   1.200  0.50 10.00           O
  HETATM    5  S2 ATHJ D   1       1.643  -0.000   0.000  0.50 10.00           S

and H2S (http://ligand-expo.rcsb.org/reports/H/H2S/) with

  ATOM      1  S  BH2S E   1      -0.051   0.000   0.000  0.50 10.00           S

you would use

 NOTE BUSTER_RESET_CONSTANT_COMBINE

 NOTE BUSTER_SET AltOccAll = Empty

 NOTE BUSTER_SET AltOcc1    = { D|1:*.A}
 NOTE BUSTER_SET AltOccAll = AltOccAll + AltOcc1

 NOTE BUSTER_SET AltOcc2    = { E|1:*.B}
 NOTE BUSTER_SET AltOccAll = AltOccAll + AltOcc2

 NOTE BUSTER_OCCSUM 1.0 0.005 D|1:S1.A E|1:S.B

 NOTE BUSTER_SET FixOcc = All
 NOTE BUSTER_SET FixOcc = FixOcc \ AltOccAll
 NOTE BUSTER_CONSTANT OCC FixOcc

This /should/ work ... if not, please let us know.

Cheers

Clemens & Claus