Low/High B's

[Jiang Jiansheng]

> We have been having trouble with our B's.  In some structures we have an average B of
> ~30, other structures, same protein and data collection/reduction routine, have an average
> B of ~10.  The electron density does not appear to be that much different between these
> structures.  Our data are collected to 2.5A resolution.  We are, of course, using XPLOR 

> for refinement.  Has anyone experienced this before?  If so how did you get the situation
> resolved?  Thanks.
>
> brad


It is probably that the overall B scale was set too high/low.

1. Check with the Wilson plot by using the same resolution range
that was used in the B refinement.  The determined B scale from
Wilson plot would be comparable with the overall average
B value.  If there is big differences between two Bs,
you can reset all B values to the B_Wilson scale and
re-carry out the B refinement again in X-PLOR.

2. If there is also a big difference on Wilson B values
between different structures, and same data correction/
reduction routine are used, It may be a meaningful
difference! It is normal that average B values are around 10-30 

for macromolecular structures. Usually, the lower resolution
data the higher average B value.

3. Check with data completeness.  Missing data in the
high resolution range could also lead to an artificial average
B value.

Jiansheng Jiang.