On Wednesday, 01 June 2016 08:48:41 PM Jessica Bruhn wrote:
> Hi all,
>
> I’m refining a structure and ran into an odd situation where the solvent and ligand molecules have lower b-factors than the protein:
>
> Average B-factor 60.73
> Macromolecules 60.90
> Ligands 28.25 (BME and Na)
> Solvent 38.14
I do not see anything odd about that.
A ligand binding site is often, though not always, the best ordered part of the
molecule. So the B factors for both the ligand and the protein in that region
are expected to be lower than the overall mean B.
By the same token you probably only added water molecules in places where
they were clearly visible. I.e. the better-ordered part of the structure.
In fact, my own rule of thumb for water molecules is that any water molecule
that refines to a B factor >60 gets removed from the model. So in my case
for any structure with mean B > 60 inevitably all water molecules in the final model
will have lower than average B. YMMV.
Ethan
> I was wondering if anyone had any explanations or tips for fixing things. This dataset was highly anisotropic, so I ellipsoidally truncated my data (2.6A, 3.5A and 3.7A in each direction). This dataset is also twinned with a 49% twin fraction. Despite all of that, I was able to build a relatively decent structure (28%/31%) that agrees well with the same protein in a different space group. My b factors vary widely across this structure and clearly one end of the protein is better ordered than the other end. The water molecules that I have built are all near the more ordered (lower b-factor) portion of the protein. I also built one BME molecule (b factor 27) and two sodium ions (b factors 30). I’ve tried re-setting all b factors and then re-refining b factors, which helped a little. Are these statistics reasonable given the issues with my data? Is there any way to fix this? I’d love any advice you can give.
>
> Thanks,
> Jessica
>
--
mail: Biomolecular Structure Center, K-428 Health Sciences Bldg
MS 357742, University of Washington, Seattle 98195-7742
|