*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
".. purely anecdotal, but in general .."
In general (!), not so anecdotal afterall....
".. with luck .."
See Kuszewski et al., 1996: "it is invariably the case that
high-resolution X-ray structures show better agreement with solution
observables than the corresponding NMR structures, including the very
best ones".
It is therefore not surprising that one or more of the single NMR models
can be used (often leaving out the most flexible parts!) and this option
should be tested first. After all, the crystal will trap one or more
conformers found in solution, will it not? Therefore, in general you do
better with single models!
".. the average .."
Of course, sometimes an energy-minimized average structure will do the
job were the single models do not work.
However, this is somehow intriguing in view of the above-discussed
issue, is it not? Often, if the average structure performs better, the
single NMR models are more poorly refined/defined. Have a look at the
"constraints per amino acid" beforehand, the fewer constraints the
larger the range of possible structures that satisfy the data and, the
higher the possibilities of making mistakes, which in turn can get
"cured" through averaging. Therefore, sometimes you do better with an
averaged model!
Jeroen.
Eleanor Dodson wrote:
This is purely anecdotal, but in general I have done better with
testing single NMR models one by one rather than using the ensemble. I
have rationalised this by saying that with luck one of the predicted
structures will fit reasonably closely to the one in the crystal,
whereas the average must have horrible distortions from protein gemetry..
You can compare results - obv. if the SAME solution appears several
times that makes you feel very confident,
Eleanor
Jeroen Mesters wrote:
The reason(s) for obtaining (calculating) an ensemble of structures
are 1) true flexibility and of course 2) lack of data! This holds
true for both NMR and Xtallography.
Each single model is in agreement with the data (i.e. was refined
against the data) and an average model is probably the worst one of
all to take.
Better to use one of the models and remove the parts that differ most
among all the models.
J.
[EMAIL PROTECTED] wrote:
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
An easy alternative is to calculate the average structure and choose
the individual model with lowest rms deviation from the average as
representative of the ensemble. Since this is a typical task in the
NMR field, this option is mostly found in NMR software, e.g. MOLMOL.
Of course, conclusions drawn using these two methods may differ,
particularly if the individual models deviate a lot in certain parts
of the structure. In this case, an energy-minimized average
structure would probably be more "representative". CNS provides such
a model_minimize input (for use without experimental data).
Oliver
---------------------
Dr. Oliver H. Weiergraeber
IBI-2 (Structural Biology)
Research Centre Juelich
D-52425 Juelich
Germany
Phone: +49-2461-612028
Fax: +49-2461-612020
---------------------
----- Original Message -----
From: huangwei <[EMAIL PROTECTED]>
Date: Thursday, June 8, 2006 5:43 pm
Subject: [ccp4bb]: how to convert an ensemble model into a single model
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Hi guys,
I have tried AmoRe using a NMR ensemble, thus get an solution in
the form of ensemble.I was informed that "a single macromolecular
model derived from an ensemble of NMR models by averaging atom
positions and minimizing the energy".Could you please tell me some
software about this? Or any better way to convert an ensemble model
into a single model?
Best regards!
????????huangwei
[EMAIL PROTECTED]
??????????2006-06-08
--
Jeroen Raymundus Mesters, Ph.D.
Institut fuer Biochemie, Universitaet zu Luebeck
Ratzeburger Allee 160, D-23538 Luebeck
Tel: +49-451-5004070, Fax: +49-451-5004068
E-mail: [EMAIL PROTECTED]
Http://www.biochem.uni-luebeck.de
--
If you can look into the seeds of time and say
which grain will grow and which will not - speak then to me (Macbeth)
--
