*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
oi mario !
I am on the final stage of refinement and I am making sure, on coot, that
the residues are in the allowed regions of the Ramachandran Plot. The
problem is that, after Refmac, those residues are thrown back on the
disallowed regions.
the main question, which i'm surprised to see nobody has asked yet, is *how*
are you "making sure, on coot, that the residues are in the allowed regions of
the Ramachandran Plot" ?
do you do a careful analysis of each outlier, trying to assess if it is an
error in the model or a genuine feature of the structure ("outlier for a
reason"), then looking for explanations in case you think it is an error, and
doing some sensible rebuilding (probably also of some neighbouring residues)
to try and fix the underlying *problem*?
or do you twiddle phi and psi to fix the *symptom*?
if you do the former: stout fellow!
if you do the latter: naughty boy! read some of the literature about model
rebuilding and validation, e.g.:
- http://xray.bmc.uu.se/gerard/gmrp/gmrp.html (esp. the sections "Quality
control" and "Rebuilding")
- http://xray.bmc.uu.se/cgi-bin/gerard/reprint_mailer.pl?pref=34
- http://xray.bmc.uu.se/cgi-bin/gerard/reprint_mailer.pl?pref=56
also, at lower resolution you expect your model to have more outliers than at
higher resolution, so having zero ramachandran outliers is not a goal you
should be pursuing in the first place. see for instance:
- http://xray.bmc.uu.se/cgi-bin/gerard/reprint_mailer.pl?pref=65
- http://xray.bmc.uu.se/cgi-bin/gerard/reprint_mailer.pl?pref=39
- http://xray.bmc.uu.se/cgi-bin/gerard/reprint_mailer.pl?pref=29
as for the discussion on "desirable" rmsds - i thought we (the community) had
agreed many years ago that the lower the resolution, the tighter your
restraints *must* be. (in the limit of zero reflections: use constraints)
sure, you can tune the restraints to get an "atomic resolution rmsd" of 0.02 A
or whatever, but you are unable to determine the individual bond lengths with
an accuracy that warrants it. so you get the "right" distribution, but for the
wrong reasons. i seem to remember (but my memory is write-only nowadays, so
don't take my word for it!) that ian tickle wrote the definitive posting about
this on ccp4bb a couple of years back.
it looks as if every five years or so, the community forgets what it knew to
be sensible previously and dusts off old misconceptions and fallacies (about
such things as rfree, rmsds, ncs, sigma cut-offs, ramachandran, etc.). sure,
some "truths" change over time as our methodology improves (that is the way of
science, and how it should be) - but we still cannot extract information from
reflections we haven't observed! so, with respect to restraints at least,
every five years someone will have to tell us to remember WWW (Wayne's Wise
Words): "where freedom is given, liberties are taken".
--dvd
******************************************************************
Gerard J. Kleywegt
[Research Fellow of the Royal Swedish Academy of Sciences]
Dept. of Cell & Molecular Biology University of Uppsala
Biomedical Centre Box 596
SE-751 24 Uppsala SWEDEN
http://xray.bmc.uu.se/gerard/ mailto:[EMAIL PROTECTED]
******************************************************************
The opinions in this message are fictional. Any similarity
to actual opinions, living or dead, is purely coincidental.
******************************************************************
