parameter sets

John Kuszewski johnk at spasm.niddk.nih.gov
Wed Nov 29 03:13:03 EST 1995


In article <4954kd$91i at mserv1.dl.ac.uk>, GERARD at XRAY.BMC.UU.SE (Gerard 'CD' Kleywegt) writes:

|> NOE violations, how accurate are those ?  People who don't
|> fit buildup curves used to classify NOEs as short, medium
|> or long (fuzzy science ? ;-) when I was young (not sure
|> if they still do that, having joined "the other side").

The terms are "strong," "medium," and "weak."  :-)  

|> I would guess that errors don't build up in one direction,
|> but more or less compensate each other (you could find out

...

|> and see which one gives the best model (don't use procheck,
|> since that is heavily biased towards anything refined
|> against engh & & huber; also don't use noe violations

Huh?  First you agree with me that the precise values of the
angles probably don't introduce significant errors in a 
structures, and then you say that you shouldn't look at 
PROCHECK because it's biased toward structures refined
with a particular (as well as accurate, and similar to 
those values usually used in parallhdg) set of angles.
Which is it?  

|> xtallography) - instead, use the ramachandran plot and

Again, that's the first page of PROCHECK's output.

|> vriend & sander's directional atomic contact analysis;
|> these check things not restrained in refinement and
|> are therefore good validation criteria; you could also

It's true that you need cross-validation terms, and that
individual NOEs are not as good as individual reflections,
but I don't think that the WHAT-IF "quality control" is 
the best choice either.  The best choice is to use your
own data--the proton chemical shifts.  They're already 
measured, they're very sensitive to the local environment, 
and they're (fairly) well predicted from theory.  You
could throw out a fraction of your 1H shifts (you are
refining against them directly, aren't you? :-) and
use that for cross-validation.  Hmm--perhaps someone
should write a paper on that?

-- 
                                   _____________
                                   |        ___/_
                                   |        |/  /
                                   --  /\  //  /--
                                   ||  ||  /  /||
                                   ||  || /  / ||
                                   ||  ||/  /  ||
John Kuszewski                     ||  |/  /|  ||      
johnk at spasm.niddk.nih.gov          ||  /  /||  ||
                                   \/ /  / ||  \/
that's MISTER protein G to you!     |/__/|      |
                                      /_________|

"Biophysics has driven me to an attitude of apocalyptic doom"
   --Frank Delaglio



More information about the X-plor mailing list