Dear Colleagues,
Warren makes a good point, but let me recast it slightly. We are
all part of the structural biology community, and we, as a community,
need to get together and explore our options for interchange of
information. Some of us already have a proposal pending with the
NSF for a collaboration on Management of Experimental Data in Structural
Biology (MEDSBIO) to institutionalize that process for our raw
experimental data. The wwPDB, as a member of the community, is
cooperating in that effort. Perhaps it is time for another formal
collaboration, say, a collaboration of Management of Derived
Information in Structural Biology (MDISBIO) that would focus on
coordinates and other derived information. I hope and believe
that the wwPDB would cooperate in that effort, and I am certain the
MEDSBIO would cooperate as well.
I am not proposing a standardization effort. There are too many
valid, but diverse, needs and the pace of change is too rapid for one
representation to satisfy everybody, but we could work to reduce
the current chaos to more manageable levels and work to achieve
clean, well-documented interoperability among the necessary minimum
number of standards.
In the computer science community, we have learned to live with
multiple programming languages, and to treat the "standards" for
those languages as living documents to be studied and considered
in an open and collaborative manner and to revise those standards
every few years. This practice might serve the structural biology
community as well.
Regards,
Herbert
At 6:19 PM -0700 9/19/07, Warren DeLano wrote:
Nevertheless, there is a huge problem with interoperability between
software packages, and this core issue greatly frustrates everyone, not
only Joe.
Frankly, it is a disgrace that even in 2007, software users in
structural biology and computational chemistry still cannot reliably
exchange basic molecular structure information between programs.
But we can only blame ourselves for this: it is *WE SOFTWARE DEVELOPERS*
who must pick up where the PDB mandate lets off in order to create
robust and workable interchange systems between our various packages.
We cannot expect to the PDB to do something only we, collectively, know
how to do, and likewise, the PDB should not continue to be scapegoated
for our own failings (as Joe may be doing, at some level).
Hijacking the PDB format may not be the best way to go, but without
question, something must be done. The status quo IS A MESS!
Whether it is a virtual gathering via Wiki, or an old-fashioned
round-up, there are several dozen key molecular software developers who
need to be "locked in a cell together" until some minimum reasonable
solution shakes out for representing and sharing molecular structure
information: something that we can all live with, and something that we
will all commit to adopting.
But how? when? & where? And what would actually motivate participation?
I don't yet have those answers, and perhaps the wwPDB could indeed play
a key role in this process. But ultimately, this isn't their job or
their burden.
It is ours, collectively, as developers of molecular software.
Sincerely,
Warren L. DeLano, Ph.D.
DeLano Scientific LLC
-----Original Message-----
From: CCP4 bulletin board [mailto:[EMAIL PROTECTED] On Behalf Of
Paul D. Adams
Sent: Wednesday, September 19, 2007 1:17 PM
To: CCP4BB@JISCMAIL.AC.UK
Subject: Re: [ccp4bb] Why wwPDB and members are doing a poor job.
Joe,
I think this is neither a constructive or accurate statement. I
have collaborated with the PDB (both the RCSB and the EBI) for many
years and I believe that they are doing a very good job under very
difficult circumstances. They have been tasked with curating an ever
increasing volume of data from a community that, as you point out, is
constantly developing new methods. As a community we need to
communicate and collaborate with the PDB, not try to sideline them. I
know that they are willing to listen so we should take that opportunity.
Cheers,
Paul
On Aug 29, 2007, at 8:08 AM, Joe Krahn wrote:
This is a reply to the below message posted under "[ccp4bb] The
importance of USING our validation tool", which is a rather long
thread now.
This is part of why I claim that wwPDB and its members are doing a bad
job. They have worked to systematically remove "general purpose"
information that does not fit their pre-defined schemes, which are
developed with out much interaction with the user community. the
problem
is that we are doing RESEARCH, which means that we will continue to
develop new methods over time. The sensible thing to do is to allow
unformatted user-defined information, and eventually work it in to a
properly formatted, standard item if that information is seen as
generally useful by the user community.
I think that the lack of community involvement by the database
administrations should be a clear indication of why we should NOT
switch
from PDB to mmCIF format for coordinate files. Instead, we should take
this opportunity of wwPDB members abandoning the PDB format to take
over
management of the format ourselves. I was quite irate with them for
going against our wishes on several features of the PDB format, like
supporting the SegID. Instead, I think we should realize that "modern
database" management goals are different from experimentalist
goals, and
that we should not rely on them to decide how our own data should be
represented.
I think that we should intentionally avoid mmCIF for coordinate files,
and stick to the PDB format. The wwPDB has absolutely no policy for
user
involvement, and RCSB has clearly dropped the previously establish
PDB-format change policy. Their task was never to manage a public file
format standard. This is an opportunity to turn the PDB file format
into
a public standard.
I have started a PDB Format Wiki, running on my home computer, at
http://pdb.homeunix.org. If it gains interest, I will see about moving
it to a proper Internet host.
Joe Krahn
Miller, Mitchell D. wrote:
Hi Boaz,
We were informed by an RCSB annotator in April 2006 that the
RCSB had suspended including REMARK 42 records in PDB files
pending the review of the process by the wwPDB.
In looking at the new annotation guidelines, it looks
like the result of that review was to reject the REMARK 42
record and the listing of additional validation items.
See page 23 of the July 2007 "wwPDB Processing Procedures
and Policies Document"
http://www.wwpdb.org/documentation/wwPDB-A-20070712.pdf
"REMARK 42 and use of other programs for validation Use of REMARK
42 is
discontinued.
If authors wish to indicate presubmission validation and other
programs used before
deposition, the programs may be listed in a new remark, REMARK 40.
This remark will
list the software name, authors and function of the program.
Results of the software will
not be listed. Use of this remark is voluntary."
It seems that the wwPDB only allows the inclusion of validation
statistics output by the refinement program but not from additional
validation programs. So for additional statistics to be included
in the PDB header, they will either need to be implemented by the
refinement package or the wwPDB annotators.
Regards,
Mitch
--
Paul Adams
Senior Scientist, Physical Biosciences Division, Lawrence Berkeley Lab
Adjunct Professor, Department of Bioengineering, U.C. Berkeley
Head, Berkeley Center for Structural Biology
Deputy Principal Investigator, Berkeley Structural Genomics Center
Building 64, Room 248
Tel: 510-486-4225, Fax: 510-486-5909
http://cci.lbl.gov/paul
Lawrence Berkeley Laboratory
1 Cyclotron Road
BLDG 64R0121
Berkeley, CA 94720, USA.
--
--
=====================================================
Herbert J. Bernstein, Professor of Computer Science
Dowling College, Kramer Science Center, KSC 121
Idle Hour Blvd, Oakdale, NY, 11769
+1-631-244-3035
[EMAIL PROTECTED]
=====================================================
