structural refinement using low temperature data

Aaron Oakley s_aarono at EDUSERV.ITS.UNIMELB.EDU.AU
Wed Apr 10 19:54:30 EST 1996


> Hi,
> 	Does anyone has experience of structural refinement using low
> temperature data? I try to refine my protein structure with low temperature
> data, but the r value is about 0.3, which can't go down further without B-value
> refinement.

Low temperature structures often have lower B-factors than the equivalent
structures refined against room-temp. data.
If you are at R=30%, then you could try doing some B factor refinement.
(Individual restrained, or Group, depending on the resolution of your data)
If you don't want to do group or individual B factor refinement yet, you
could try refining a single B factor for the entire structure. (using
the group protocol with group=(not hydrogen) )
Also, if you suspect that the data is anisotropic, try compensating
for this by refining an overall anisotropic temperature factor
(which will be applied to the HKL's, not your structure), using the
baoverall.inp script.

I have found that there are no special problems associated with refining
against low temperature data. Of course, you should use Rfree in all
cases to see if you are over-fitting.

Hope this helps,

Aaron





More information about the X-plor mailing list