Worldwide Protein Data Bank
www.wwpdb.org
www.wwpdb.org
September 7, 2007
Worldwide Protein Data Bank
www.wwpdb.org
Agenda Welcome and introductions Accomplishments Remediation rollout summary Toward the futureBreak
Matters arising– Incorrect structures
Executive session Feedback to wwPDB Set next meeting date
Worldwide Protein Data Bank
www.wwpdb.org
wwPDB AchievementsOctober 2006 - September 2007 Continued growth of archive Website updates Publications and presentations Time-stamped archive Remediation rollout Annotation document One stop shop: NMR, cryoEM
Worldwide Protein Data Bank
www.wwpdb.org
Depositions since wwPDB establishment
Worldwide Protein Data Bank
www.wwpdb.org
PDB entry processing 1-1-2000 10,997 entries in PDB Today 10-Jul-2007 44,578 entries in PDB
Size now is 4 times larger than when the 3 sites started
In 1999, 2361 entries were deposited In 2006, 7282 entries were deposited
We handle more than 3 times as many entries per year with less staff – and all wwPDB sites produce high quality annotated PDB entries
No current backlog of unprocessed entries
Worldwide Protein Data Bank
www.wwpdb.org
Time-stamped copies of the archive
57 Gbytes of data for 2006, released January 2, 2007
68 Gbytes of data for July 2007 snapshot Both include
– PDB format entries– mmCIF format entries– PDBML format entries – Experimental data– Dictionary, schema, and format documentation
Worldwide Protein Data Bank
www.wwpdb.org
Outreach
wwPDB website Discussion forums NMR Task Force Publications Professional society meetings
Worldwide Protein Data Bank
www.wwpdb.org
Worldwide Protein Data Bank
www.wwpdb.org
Joint publications
Nucleic Acids Research, 35: D301 (2007)– The worldwide Protein Data Bank (wwPDB): ensuring a single, uniform
archive of PDB data Nature Structure Molecular Biology, 14:354 (2007)– Reply to: Building meaningful models of glycoproteins
Nature Biotechnology, 25: 854 (2007)– Response to “Overhauling the PDB”
Methods in Molecular Biology, in press– Data deposition and annotation at the wwPDB
Structural Bioinformatics 2nd Edition, in press– The wwPDB
Worldwide Protein Data Bank
www.wwpdb.org
Interactions since October 2006 Exchange visits
– MSD/RCSB PDB (4) – PDBj/RCSB PDB (1)– PDBj/BMRB (2)– BMRB/RCSB PDB (1)
Phone conference with site directors-twice a year VTC’s among staff
– BMRB/RCSB PDB twice a month (ADIT-NMR)– MSD/RCSB PDB twice a week (annotation procedures, remediation)– RCSB PDB/PDBj and BMRB/PDBj on necessary occasions
Email among staff– MSD/RCSB PDB ~2 per day– PDBj/RCSB PDB ~2 per day
Worldwide Protein Data Bank
www.wwpdb.org
New initiatives
One stop shop for NMR data and models One stop shop for electron microscopy maps
and models (NIH-funded)
Worldwide Protein Data Bank
www.wwpdb.org
Recommendations from 2006 wwPDBAC report Implement the recommendations from November 19-
20 2005 modeling workshop (Berman et al. Structure 14, 1211-1217) – Models phased out October 16, 2006
Rollout remediated data to superusers by December 31, 2006; to all users by July 1st 2007; Provide access to PDB formatted files following the most current format.– Superusers had access to data November 2006, all users in April
2007
Worldwide Protein Data Bank
www.wwpdb.org
Recommendations from 2006 wwPDBAC report Work with SAXS community to create appropriate
representation of these data, and circulate progress reports to the Committee as appropriate– Not done
Expand the four character PDB ID codes before the number of depositions reaches 400,000– Number of available PDB ID codes has been increased by allowing
IDs to start with a character
Develop and present a formal recommendation to the wwPDBAC regarding the purview of the PDB at our September 2007 meeting in Princeton, NJ – In process
Worldwide Protein Data Bank
www.wwpdb.org
Recommendations from 2006 wwPDBAC report Coordinate with the wwPDBAC to obtain formal
letters of support when seeking funding; establish a coordinated plan to both educate and lobby funding agency representatives; establish a charitable organization to serve as a conduit for receipt of both grant funding and gifts from pharmaceutical and biotechnology companies, involving individual Committee members as needed.– Funding Representatives Round Table Discussion
Worldwide Protein Data Bank
www.wwpdb.org
Remediation
Worldwide Protein Data Bank
www.wwpdb.org
Key drivers
Chemistry and nomenclature Sequence and taxonomy Citations Viruses
Worldwide Protein Data Bank
www.wwpdb.org
IUPAC, NMR, and the PDB
Atom nomenclature and NMR restraints
John L. Markley
Worldwide Protein Data Bank
www.wwpdb.org
History of the NMR-led requested remediation of hydrogen atom nomenclature When BMRB was established in the late 1980’s, it adopted the IUPAC atom
nomenclature recommendations from Biochemistry 9, 3471-3479, 1970
At that time, we noted that NMR structures being deposited in the PDB did not adhere to these recommendations (particularly for H-atoms; e.g. HB1/HB2 instead of HB2/HB3), and I brought this to the attention of the director of the PDB at Brookhaven with the request that it be remedied
A group of NMR spectroscopists led by Kurt Wüthrich worked with the NMR community to develop recommendations for the deposition of NMR structures; all agreed that the prior IUPAC recommendations be maintained (Pure & Appl. Chem., 70, 117-142, 1998)
Over the years, wwPDB Task Force on NMR has pushed strongly for remediation of atom nomenclature
Worldwide Protein Data Bank
www.wwpdb.org
Accomplished: atom nomenclature remediation Nomenclature in PDB now matches that in BMRB The single format will avoid confusion and errors All discrepancies have been resolved in the remediated
files, with the minor exception of atoms at the C-terminus
IUPAC-IUBMB-IUPAB wwPDB
H'' HXT
O' O
O'' OXT
– Since these atoms are not observed by NMR spectroscopists, we do not consider this to be a problem
– We plan to write an addendum to the IUPAC-IUBMB-IUPAB “Recommendations” for submission to Pure & Appl. Chem. to formalize these as “accepted atom designators”
Worldwide Protein Data Bank
www.wwpdb.org
Remediation of NMR structure files
Required the linking of structure files and restraint files
Atom names, residue numbers and chain identifiers needed to be updated
Remediation of restraint files required the unpacking, parsing, and regularization of legacy information contained in PDB “MR” files into the “NMR Restraints Grid”
Worldwide Protein Data Bank
www.wwpdb.org
NMR Restraints Grid development
BMRB, University of Wisconsin-Madison, USA
MSD, European Bioinformatics Institute, Hinxton, UK
Department of Computer Sciences/Condor Project, University of Wisconsin, USA
Department of NMR Spectroscopy, Utrecht University, The Netherlands
Centre for Molecular and Biomolecular Informatics, Radboud University, The Netherlands
Worldwide Protein Data Bank
www.wwpdb.org
NMR Restraints Grid development
PDB MR files are converted into NMR-STAR
NMR-STAR file and the corresponding PDB coordinate file are parsed;
the information is connected inside the CCPN framework; and the results
are written out as NMR-STAR files; converted restraint files are filtered to
remove redundant restraints
Files made available in the NMR Restraints Grid with access from links in
each corresponding PDB entry
NMR restraint data files with atom nomenclature corresponding to
remediated PDB data files will be available by the end of 2007
Worldwide Protein Data Bank
www.wwpdb.org
Current state of the NMR Restraints Grid
Grid contains 3583 entries with a total of 3,882,595 parsed restraints
3583 entries out of 6508 in PDB have restraints Database is updated continuously as new PDB entries
are released that have associated NMR restraints
Worldwide Protein Data Bank
www.wwpdb.org
Recent agenda items considered by the wwPDB NMR Task Force Strongly recommend that restraints be mandatory for
all NMR depositions to the PDB Commissioned the development of procedures for
representing uncertainty in NMR structures and for specifying the single model meant to be most representative of the structure
Task Force should write an article for J. Biomol. NMR on its recommendations for data representation and submission of experimental data
It was suggested that the Task Force begin to discuss validation issues
Worldwide Protein Data Bank
www.wwpdb.org
Most X-ray structures are supported by structure factorsDeposited Crystal Structures and Structures Factor Files
0
1000
2000
3000
4000
5000
6000
7000
1999 2000 2001 2002 2003 2004 2005 2006Year
Count
Crystal Structures Structure Factors
Worldwide Protein Data Bank
www.wwpdb.org
Less than half of NMR structures are supported by restraint dataDeposited NMR Structures and Restraint Files
0
200
400
600
800
1000
1200
1999 2000 2001 2002 2003 2004 2005 2006 2007Year
Count
NMR Structures Restraint Files
Worldwide Protein Data Bank
www.wwpdb.org
Most structural genomics centers regularly provide restraints, but the overall average is low
0%
20%
40%
60%
80%
100%
RIKEN OTHER SG TOTAL
Number of NMR structures deposited
1127
880
247
Per
cen
t o
f d
epo
site
d
stru
ctu
res
wit
h r
estr
ain
ts
Structural genomics center
Worldwide Protein Data Bank
www.wwpdb.org
Remediation rollout
Helen M. Berman
Worldwide Protein Data Bank
www.wwpdb.org
Remediation: scope and statistics
All primary citations verified (45K)
Sequences & taxonomy updated for 61K sequences
Ligand stereochemistry and nomenclature for 13M monomers and 170K non-polymer molecules
Symmetry and coordinate transformations for 280 virus entries
10814 diffraction source & beamline updates
~1000 miscellaneous uniformity issues
Worldwide Protein Data Bank
www.wwpdb.org
Remediation process Corrections contributed and reviewed by all wwPDB
members Corrections on the archival mmCIF data files tracked
in a version tracking system (CVS) New PDBx/mmCIF, PDBML-XML, and PDB format
data files produced Validated by each wwPDB group Staged public testing began January 2007 Iterative corrections based on external comments
made through July 2007 Remediated archive released August 1, 2007
Worldwide Protein Data Bank
www.wwpdb.org
Remediation-supporting infrastructure Internal (wwPDB) CVS archive remediation data files Internal (wwPDB) rsync distribution site for
remediated data files Early tests of web, rsync, & ftp distribution sites for
dictionaries, PDB, mmCIF, and XML data files Complete wwPDB ftp site for remediated data and
dictionaries updated with remediation corrections and weekly PDB updates
200K CVS remediated data file updates 1M+ remediated file updates to support testing and
distribute from January 2007 - present
Worldwide Protein Data Bank
www.wwpdb.org
Checking the remediated files
Haruki Nakamura
Worldwide Protein Data Bank
www.wwpdb.org
Different checks
References to external databases Data processing consistency checks PDBML/XML validation Database loads User-contributed diagnostics
Worldwide Protein Data Bank
www.wwpdb.org
References to external databases
Sequence and taxonomy (UniProt) Primary Citations (PubMed)
Worldwide Protein Data Bank
www.wwpdb.org
Data processing consistency checks
Covalent geometry and stereochemistry Compliance with wwPDB Chemical
Component Dictionary– Molecular and stereochemical assignment– Atom and residue nomenclature
Compliance with PDB Exchange Dictionary– Data types, controlled vocabularies, parent-child relations
External tools such as WhatIF
Worldwide Protein Data Bank
www.wwpdb.org
PDBML/XML schema validation
Version control Data type consistency Data ranges Controlled vocabularies Referential integrity XPath traversal of PDBML data hierarchy
Worldwide Protein Data Bank
www.wwpdb.org
Database loads
Diagnostics obtained from loading remediated data into existing database systems– Relational databases used by MSD-EBI and RCSB PDB– XML database used by PDBj
Worldwide Protein Data Bank
www.wwpdb.org
User-contributed diagnostics Batch checking of remediated files by Phenix revealed
consistency issues with alternate conformations - Ralf Grosse-Kunstleve
Batch checking for inconsistent linkages and missing residues by docking software - Tommy Carstensen
Nomenclature - Tom Goddard & Chimera Group Sequence and assembly diagnostics - Roland Dunbrack Relational data integrity diagnostics - Dan Bosler Nomenclature and experimental details - Clemens
Vonrhein Many specific issues related to chemical assignments,
disorder, and nomenclature
Worldwide Protein Data Bank
www.wwpdb.org
Looking toward the future
Kim Henrick
Worldwide Protein Data Bank
www.wwpdb.org
Annotation project
Standardize annotation rules and policies among wwPDB sites
Document annotation rules and policies Create venue to update annotation rules and
policies as necessary
Worldwide Protein Data Bank
www.wwpdb.org
Annotation projectHow did we get there? Review and discussion of each PDB field by
email and VTC Document written and reviewed by all staff Final review by site directors Software compliant to new annotation
procedures implemented Tested software and trained annotators Published document on web (January 2007)
Worldwide Protein Data Bank
www.wwpdb.org
Annotation document
Specification of ALL fields in PDB file Clarification of policies
– Assignment of PDB IDs– Release of files and information– Changes to entries
Clarification of data representation– Chain ID for all atoms in the file– Multi-model representation for alternate conformation or
disorder– Chimeras – Microheterogenity
Worldwide Protein Data Bank
www.wwpdb.org
PDB IDs and DOIs
Credit for a PDB entry in CVs
Used as a reference in publications– http://dx.doi.org/10.2210/
pdb4hhb/pdb
See also DOIs for Biological Databases Philip E. Bourne, CrossRef 7th Annual Meeting, 1 November 2006 Cambridge, MA
Worldwide Protein Data Bank
www.wwpdb.org
Outstanding issues
Microheterogeniety Disorder Large structures
Worldwide Protein Data Bank
www.wwpdb.org
wwPDB and software developers
ACA 24th July 2007 meeting in Salt Lake City
“Future Challenges for the PDB: What should the PDB be doing in 2015?”
Attended by software developers and wwPDB staff
Worldwide Protein Data Bank
www.wwpdb.org
July 24 meeting Technical discussions
TLS Multiple models Large structure
demand for one file per structure Microheterogeneity Twinning
George Sheldrick, Paul Adams and Garib Murshudov produce a draft of the PDB format to describe twinning and to represent the data in HKLF
Procedural outcomes Yearly developer meeting Editorial board to assist in difficult annotation problems Ongoing electronic forum
Worldwide Protein Data Bank
www.wwpdb.org
Toward a single processing tool
This weekend – wwPDB retreat with contributors from RCSB PDB Rutgers and UCSD, BMRB, PDBj, and EBI-EMBL
Task – come to agreement to pool resources to produce a single deposition tool and design of new processing pipeline