Hi John,
Just to play devils advocate (is that the right terminology?), a
random list of potential issues that could cause that .. seen that,
been there, suffered ...
* Rfactor tells you something about your X-ray term, so if you have
the weight on the X-ray term too low (or the weight on geometry too
high) you will get a model of perfect geometry that doesn't fit your
data anymore (i.e. the R-factors go up). This can happen especially
if you have a poor(ish) model where a (basically) geometry-only
refinement will move those poor residues/loops etc around a lot
Have you compared the structures before and after this R-factor
increase?
* maybe you have really bad clashes - even if your model is near
perfect. That can happen e.g if you pull over the set of test-set
flags (FreeR_flag in MTZ terminology) from a previous
structure/refinement. This is obviously a good thing and the right
thing to do - but be sure to do this in a way that the cell
parameters in your MTZ file are the ones from the actual data (and
not the ones from the FreeR_flag of a crystal with 6A shorter A
axis).
It's surprising how much difference a hklin1/hklin2 swap in CAD can
make!
* you don't mention if the R and Rfree move up identically - or if you
have a faster increase in R than in Rfree, which would mean that
your R-factors are increasing (bad I guess) but your Rfree-R gap is
closing down (good).
So moving from R/Rfree=0.20/0.35 to R/Rfree=0.32/37 is different
than moving from R/Rfree=0.20/0.25 to R/Rfree=0.23/0.28.
* resolution limits: are you suddenly including all those poorly
measured or non existent reflections at the low resolution end (10A
and lower) that are only present because the beamstop masking wasn't
don properly during data processing/integration?
These bogus reflections can mess up your bulk-solvent correction and
scaling with weird effects. Better check the scale factors you're
getting during refinement and if they make sense.
* mixing up reflection files (very, VERY easy to do). Maybe you
re-processed the data and the indexing is now chosing a different
alternative. This depends on the spacegroup - see
http://www.ccp4.ac.uk/dist/html/reindexing.html
* Some graphics programs are very good in re-setting B-factors of your
PDB file to something else (usually 20.0 or 30.0 ... but I've seen
0.00 as well) - mainly in parts you've rebuilt. These values could
be very wrong from the correct value.
So this is just a few possible mess-ups I could think of (if you give
some more info I'm sure we all could come up with a dozen more). I
would first look into reasons similar to those - the refinement
programs are usually pretty good.
Cheers
Clemens
On Fri, Feb 13, 2009 at 08:45:18AM -0800, Pavel Afonine wrote:
> Dear John,
>
> If by unstable refinement you mean a significant increase of Rfactors
> during refinement, then there is a problem in the software you use and the
> best thing you can do is to notify the developers of that software so they
> can fix the problem.
>
> I can't tell for other software, but I'm not aware of such thing (unstable
> refinement) in phenix.refine. If you have a problem case I would appreciate
> if you send it me so I fix the problem.
>
> Pavel.
>
> On 2/13/2009 6:36 AM, John Bruning wrote:
>> Hi,
>> I would like to survey the group for causes and potential solutions to
>> unstable refinement (the steady and upward drift of R and Rfree in each
>> cycle of refinement). I am currently aware of the following potential
>> causes: twinning (meohedral, pseduomerohedral, or epitaxial), anisotropy,
>> pseudosymmetry, the lack of appropriately tight geometric constraints, and
>> extraordinarily high b-factors. If anyone has run into any other causes
>> of unstable refinement, I would like to hear about them.
>> Thanks,
>> John Bruning
--
***************************************************************
* Clemens Vonrhein, Ph.D. vonrhein AT GlobalPhasing DOT com
*
* Global Phasing Ltd.
* Sheraton House, Castle Park
* Cambridge CB3 0AX, UK
*--------------------------------------------------------------
* BUSTER Development Group (http://www.globalphasing.com)
***************************************************************
