[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[ccp4bb]: Comments on ligands
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
-------- Original Message --------
Date: Fri, 15 Nov 2002 10:43:39 -0000
From: "Jim Naismith" <jhn@st-andrews.ac.uk>
Reply-To: <jhn@st-andrews.ac.uk>
Organization: University of St. Andrews
To: <ccp4@ysbl.york.ac.uk>
Dear Eleanor,
My email account is playing up. Incoming only as a
result I cannot send out on naismith (my name of CCP4BB). Can you
forward this for me please.
Best
Jim
James H. Naismith | Research mailto:naismith@st-and.ac.uk
Professor of Chemical Biology | Teaching mailto:jhn@st-and.ac.uk
BBSRC Career Development Fellow |
Centre for Biomolecular Sciences | Office: 1334-463792 (24 hr)
The North Haugh | Fax : 1334-467229
The University | Lab : 1334-467245
St. Andrews | In UK add 0 to start of number
Fife Scotland, U.K., KY16 9ST | www http://speedy.st-and.ac.uk/
I agree with Ludovic. Most of us take on trust what is in dictionaries,
when was the last time you checked every bond in ATP with every new
release of CCP4? Let me say that I don't. Improved dictionaries have
made real progress.
Further in complex drug molecules, the chemical state you crystallised
may be different than you started with. It can sometimes be difficult to
be sure you have the right compound formulation. When is disorder a
chemical change? So your data might be moving atoms to 'unrealistic'
geometries because you have the wrong atom types. If your maps are poor,
it is often an informed best guess. There is thin line between informed
best guess and seeing what you want to see. For instance we see a very
buckled NADH ring in one of our structures, I can restrain this to be
flat but the maps tell me it is not. However, at lower resolution (even
1.8A) the maps could be considered noisy. Where is the balance of error
here? Stick to the 'safe' flat ring or model the distortion. Having
thought and worked deeply on this, I can tell you the geometry of NADH,
planarity etc is a can of worms. There is no right answer. I suspect
this is true for some drugs. These are 'honest' mistakes and perhaps we
have to clearer about the level of uncertainty.
Saying that I suspect one of the underlying tension here is a vague fear
of dishonesty. This is much harder to police. We all know well there
have been some incidences where there is a feeling that deliberate
deception has indeed occurred. Stopping this happening has got be one of
our goals. Probably the best way to stop it is to highlight when it has
happened. One way to do this is to ensure we all deposit data as well as
model.
(A sure way to discourage this is for people to make careers out of
highlighting trivial errors or honest mistakes.)
Thus I feel there is a delicate balance. Honest mistakes are best dealt
with by the collegial nature followed by Joel. I would say deposition of
F's is a sign of an honest mistake. I guess some the stuff I have done
has small errors (I don't know this is true, I believe that all the
stuff I have submitted is correct at the time), although I do not think
they materially affect the publications. I would try to work with anyone
who identified these and if required inform the Journal and PDB. But in
my and most cases, the data are there, my cooperation is not required
for you to expose me by the normal peer review route.
If no F's are deposited and no satisfactory response obtained, then
perhaps stronger action is needed. One solution would be for the
databank to use peer review to assess the coordinates and the alleged
error. If peer review feels the coordinates cannot be relied on, the
entry would be removed and the Journal contacted to inform them.
Best
Jim
James H. Naismith | Research mailto:naismith@st-and.ac.uk
Professor of Chemical Biology | Teaching mailto:jhn@st-and.ac.uk
BBSRC Career Development Fellow |
Centre for Biomolecular Sciences | Office: 1334-463792 (24 hr)
The North Haugh | Fax : 1334-467229
The University | Lab : 1334-467245
St. Andrews | In UK add 0 to start of number
Fife Scotland, U.K., KY16 9ST | www http://speedy.st-and.ac.uk/