[phenixbb] question in geometry refinement

Pavel Afonine pafonine at lbl.gov
Thu Jan 26 06:44:47 PST 2017


Hi Noelia,

it sounds like that molecule (the 5th) may be disordered and the fact 
that its occupancy refines to 0.8 is a good indication of this.. I 
suggest you do your best fixing outliers manually, then run refinement with
   - Ramachandran plot restraints enabled (you can apply them to all 
model or to just that 5th molecule only);
   - weights optimization (it runs much slower but if you can use 
multiple CPUs that will speed it up);
   - may be secondary structure restraints (that will rely on SS 
annotation provided in HELIX/SHEET records, so make sure it is correct 
if you choose to use it).

If this does not help please feel free to send me files and I will have 
a look to see what else can be done!

Good luck (and I guess I'll see you at Crystallography School in Madrid 
in May?),
Pavel

> Dear colleagues,
>
> I am dealing with the refinement of a protein with 5 monomers in the 
> asymmetric unit at 2.2 Å resolution. Four of them have correct 
> geometry and good density map, however, the fifth molecule has poor 
> density and bad geometry. Ramachandran plot is presents 91.05% 
> (favored), 3.23 % (outliers).
> I have modified the occupancy of the molecule, the refinement results 
> in 0.8 occupancy for this chain. I also tried to refined the structure 
> without this chain, but the R values increased and the density map 
> indicated that it should be there. I would like to refine this chain 
> fixing the geometry of the others which are correct. Do you have any 
> ideas about how to proceed?
>
>
> Some info:
> Protein residues 305
> 5 monomers in the a.u
> Space group C121 a=151.390, b= 67.81, c=155.420Å, 94.53º
> R value 0.23 / Rfree 0.28
>
>
> Many thanks,
>
> Noelia
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20170126/5b17b1b9/attachment.htm>


More information about the phenixbb mailing list