*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
There is definitely a role for online supplementary material, and I think
Savvas' suggestions are one good use of them.
Think of it not as a way of moving important material from prominance
to obscurity, but a way of moving less important material from oblivion
to obscurity. The types of statistics he is refering to would mean nothing
to the average reader, and there is no point in wasting space in the
journal on them. However they may be vital for letting the reviewer
decide the validity of the paper, or for future scientists trying to
sort out apparently irreconcilable conflicting results. If you accept
my paper at face value, fine. If not, the proof is in the supplementary
materials.
Another good use is for electronic illustrations that could not
be typeset- the movies or vrmls and such.
ed
Yoder, Marilyn wrote:
Dear colleagues,
I read Savvas comments (below) with interest. As I've been trying to write a
couple of manuscripts with colleagues, the 'online supplementary material' has
been a focus of attention on my plate. I must say I don't like the concept. I
guess my concern is that important information to judge the quality of the data
and the validity of a model are relegated to such unimportant status that they
don't merit ink. Frankly, I don't think a lot of readers will take additional
effort to track down supplementary material, and online access to many journals
is not a trival matter.
Am I simply old-fashioned, or do others have similar concerns?
Best regards,
Marilyn
Marilyn D. Yoder
Division of Cell Biology & Biophysics
School of Biological Sciences
University of Missouri-Kansas City
5007 Rockhill Rd.
Kansas City, MO 64110-2499
phone 816-235-1503
fax 816-235-1503
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Savvas Savvides
Sent: Monday, April 17, 2006 2:01 PM
To: Bulletin Board CCP4
Subject: Re: [ccp4bb]: Acceptable Rmerge values in last resolution shell!
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Dear colleagues,
the discussions on data quality indicators and what is acceptable for reviewes
appears to be a recurring theme on the ccp4bb on a yearly basis.
The era of 'online supplementary material' may be the golden opportunity we
have been waiting for to communicate data quality issues that the standard
'Table 1'
or 'materials and methods' do not readily address. One can imagine a number of
informative optional or even compulsory 'supplementary tables/figures'. For
example, a plot of I/sigI vs resolution for cylinders of reflections along the
three reciprocal space axes, or other relevant directions in reciprocal space
could provide useful information about the effective resolution of the data.
Other possibilities could be a couple of representative diffraction patterns,
or relevant tables from the programs we have processed our data with. Or maybe
a plot of B-values vs sequence for the well refined structures at high
resolution.
Or perhaps the electron density correlation plots of Kleywegt et al. for
structures with NCS.
Have ideas along these lines ever been discussed on the ccp4bb? I do not recall
that they have been (so please do not unleash the dobbermans if they have!!.
Best wishes
Savvas
Quoting Mischa Machius <[EMAIL PROTECTED]>:
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
To me, it almost looks like there is currently no good way to
describe, or even assess, the information content of diffraction data.
Something else seems to be needed. If we, for example, accept the
notion that there is information in data with I/sigma(I) of 1 and
below, we will have to learn to live with Rmerges of 100% and greater.
Also, some phenomena, such as radiation damage, aren't properly
modeled yet, making things look worse than they probably are. As Bart
rightfully says, people are throwing away perfectly useful data, just
because the numbers don't look too good.
I think the various cutoffs that are in use are primarily there to
satisfy reviewers. Nobody-knows-who came up with these thresholds,
when that was and - most importantly - what the rationale behind them
is. They seemed to have been passed down through the generations
without too much contemplation. I, for one, am often surprised at what
reviewers come up with, and I am getting more and more frustrated with
arguing with them. Many reviewers are not even structural biologists,
so who knows where they got their wisdom from.
It sure looks it's our own fault. We got to correct that.
One could just as well be pragmatic and use all data one can get, even
if the completeness in the highest resolution shell is only 20%, the
Rmerges approach 100%, and the I/sigma(I) is "only" 1.5. Current
maximum-likelihood methods seem to be very good at making proper use
of the information in the data. As long as the electron density
improves, use the data.
Of course, this would lead to the problem of how to assess the quality
of the electron density and, finally, the resolution of a structure. I
think this is where the real issue is. We need tools to determine the
effective resolution of an electron density in various regions. There
should not be just a single resolution value, because one portion of
the electron density may have a "resolution" of 1.8Å, whereas another
one may have 3.0Å, etc. Such a statistic would be much more useful for
scientists who are using our models than some esoteric way of
calculating Rmerges. Of course, the quality of the electron density
depends on the correctness of the model, which in turn is quite
subjective and brings a whole set of other problems with it. So, it
seems we are still having a way to go to come up with reasonable
guidelines that make the best out of our data.
On Apr 17, 2006, at 9:08 AM, Bart Hazes wrote:
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Hi Ashima,
With these statistics you shouldn't have to worry about reviewers,
it looks perfectly sensible. Actually I'm much more concerned about
the recent epidemic of overly pessimistic resolution cutoffs. In our
journal club at least half the papers have I/SigI in the highest
resolution bin in the 3-6 range which means they could have gotten
significantly higher resolution. There are situations where data
quality is more important than resolution, for instance
(anomalous) phasing, but I see the same with many native data sets.
It is not clear to me if people are placing the detector too far
from the crystal and thus not even measure the highest resolution
data or that they just elect not to process those data. Why??? To
get nicer looking statistics???? That would be VERY bad practice!!!
A kinder view is that the detector distance is set based on the
apparent resolution of the first image(s) which underestimates the
true resolution of a high redundancy data set. If you don't need a
long detector distance to resolve spots I prefer to select a
distance where my visible diffraction uses the central 80-90% of the
detector allowing mosflm to try to extract some sensible information
from beyond what the eye can see.
This looks like something James Holton may have looked at. If so I'd
be interested to hear if he or the elves have come up with a magic
rule.
Bart
Ashima Bagaria wrote:
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
HI all,
In regards to my CCP4 question about the acceptable Rmerge values
in last resolution shell..various other parameters pertaining to
the protein data at 3.5 A are I/sigmaI = 13.1 (2.3) %completeness =
95.7(96.8) multiplicity = 3.8 All suggestions are welcome Regards
ashima
----------------------------------------------------------------------
--
--------
Mischa Machius, PhD
Associate Professor
UT Southwestern Medical Center at Dallas
5323 Harry Hines Blvd.; ND10.214A
Dallas, TX 75390-8816; U.S.A.
Tel: +1 214 645 6381
Fax: +1 214 645 6353
--
