[phenixbb] refinement of a ligand with alternate conformations in one region but not the other.

[Ralf W. Grosse-Kunstleve]

Hi Pascal,

> Can I refine these two alternate conformations in Phenix knowing that a part
> of the ligand will be common (have an occupancy of one) while the two other
> ones will be separate?
>
> Do I need to enter two ligand molecules or do I have to do more elaborate
> things?

All you need to do is copy the atoms in multiple conformations and set
the "altloc" column 17. For example:

HETATM15878  C1  BHO D1256      27.570  38.520  40.214  1.00 27.53           C
HETATM15879  C2  BHO D1256      28.853  38.200  39.517  1.00 26.33           C
HETATM15880  C3  BHO D1256      28.892  38.373  38.135  1.00 25.37           C
HETATM15881  C4  BHO D1256      27.764  38.820  37.463  1.00 27.30           C
HETATM15882  C5  BHO D1256      26.572  39.118  38.117  1.00 25.47           C
HETATM15883  C6  BHO D1256      26.457  38.972  39.492  1.00 26.90           C
HETATM15884  C  ABHO D1256      27.381  38.412  41.698  0.50 25.89           C
HETATM15885  O1 ABHO D1256      27.250  37.337  42.252  0.50 27.11           O
HETATM15886  N  ABHO D1256      27.301  39.545  42.391  0.50 27.93           N
HETATM15887  O2 ABHO D1256      26.266  39.863  42.994  0.50 25.79           O
HETATM15888  C  BBHO D1256      27.585  38.324  41.685  0.50 29.66           C
HETATM15889  O1 BBHO D1256      28.651  37.997  42.159  0.50 32.42           O
HETATM15890  N  BBHO D1256      26.491  38.496  42.420  0.50 32.55           N
HETATM15891  O2 BBHO D1256      25.358  38.331  41.899  0.50 39.06           O

I think coot has helpful tools for creating multiple conformers.

> Last question is it possible that for the variable regions I may end up with
> a sum of occupancies that is not 1 while for the fixed region it might end
> up being one?

Reasonably recent versions of phenix.refine (e.g. 1.4-3) automatically
constrain the sum of multiple-conformer occupancies to one.

Ralf