On Wed, Sep 7, 2022 at 11:57 AM Matt McLeod <mjmcleo...@gmail.com> wrote:
> Hi everyone,
>
> I have a series of datasets at 253K (~2.0A), 273K (2.0A), 293K (2.0A),
> 313K (2.2A) and I am curious as to the details in determining B-factors.
>
In the larger physical universe those temperatures are not very far apart.
I would not expect much change in the physical properties of the crystal
or its content. There is a reason that "cryo" crystallography generally
means "liquid nitrogen temperature", i.e. 77K but more like 100K at the
position of a crystal in a boil-off nitrogen stream.
The first thing you should check when comparing B factors is whether the
different refinements use the same, or closely comparable, overall
B correction. Typically this is an anisotropic correction; you would
want to add an isotropic approximation of this overall term to the
individual refined atomic B factors and then compare the sums.
That gives you a back-of-the-envelope comparison.
To do a more thorough job is complicated.
I am of course highly biased towards the power of TLS 
My suggestion would be to first use the refined B factors you have to
construct a segmented TLS model. I would then reset all the Bs to a
constant and refine the data sets in parallel using a "pure TLS" model,
i.e. no individual atomic B factors.
Afterward I would compare the magnitude of the principle components
of each TLS segment as a function of the data collection temperature.
The rationale is that the individual TLS segments are chunks of the
structure, domains, subdomains, loops, etc, that may be relatively
free to vibrate within the crystal lattice. You would expect the
magnitude of this vibration to increase with temperature.
Not so much the individual atomic vibrations described by atomic
B factors.
Ethan
> I have treated these datasets more-or-less identically for comparison's
> sake. I used DIALS to index, integrate, and scale the data. I scaled the
> data to a ~0.6 CC1/2 cutoff.
>
> After fully refining the datasets, there is an odd trend with respect to
> temperature (from what has been previously published) and I assume that
> this is because of "behind-the-scenes" computation rather than a
> biophysical observation. The B-factors slightly decrease from 252-293K,
> and then significantly drop at 313K. The maps look pretty well identical
> across the datasets.
>
> 253K - 53.8 A^2
> 273K - 48.4 A^2
> 293K - 45.5 A^2
> 313K - 18.6 A^2
>
> I compared the wilson intensity plots from DIALS scaling for 273K and 313K
> and they are very comparable.
>
> I am looking for suggestions as to where to look at how these b-factors
> are selected or how to validate that these B-factor are or are not
> accurate. Also, any relevant literature would be welcomed. From what I
> have read, there is a general trend that as T increase, the atoms have more
> thermal energy which raises the b-factors and this trend is universal when
> comparing datasets from different temperatures.
>
> Thank you and happy to supply more information if that is helpful,
> Matt
>
--
Ethan A Merritt
Biomolecular Structure Center, K-428 Health Sciences Bldg
MS 357742, University of Washington, Seattle 98195-7742
########################################################################
To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1
This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list
hosted by www.jiscmail.ac.uk, terms & conditions are available at
https://www.jiscmail.ac.uk/policyandsecurity/
