!"#!"#$%
#%
The wonderful world of
structure archiving
What’s happening and what’s next?
Gerard Kleywegt, PDBe, EMBL-EBI
Winter School on Structural Biology,
CEITEC, Brno, 13 February 2015
Summary…
More!
Bigger!
Better! Cooler!
Blobbier!
Outline
•! PDB in 2014/2015
•! More! (100,000+ entries)
•! Bigger! (Large structures & “Formageddon”)
•! Better! (D&A & validation)
•! PDBe in 2015
•! Cooler! (Nifty new things coming this year)
•! The future
•! Blobbier! (Cellular imaging and hybrid methods)
(images unrelated)
Who’s who again?
PDBe, wwPDB, EMDataBank
PDBe at a glance
•! Mission: Bringing Structure to Biology
•! Founding partner of Worldwide Protein Data Bank (wwPDB)
•! Birthplace of Electron Microscopy Data Bank (EMDB)
•! Founding partner of EMDataBank
•! Major activities:
•! Deposition and annotation site for structural data on
biomacromolecules & complexes (X-ray, NMR, EM)
•! Integrated resource to serve structural data and information
•! Liaise with structural biology community
•! Guided by advisory bodies made up of community experts
•! PDBe, wwPDB and EMDataBank advisory committees
•! (Validation) Task Forces (method-specific) & ad-hoc working groups
pdbe.org
!"#!"#$%
!%
wwPDB
wwpdb.org
wwPDB partnership
•! Collaborate on “data in”
•! Policy issues
•! Weekly releases
•! Validation standards
•! Format specifications
•! Chemical Component Dictionary
•! Deposition and annotation procedures
•! Archive quality and remediation
•! Journal interactions
•! Community interactions
•! Friendly competition on “data out”
•! Serving PDB data with added-value
•! PDB-based services
•! Other services, resources and activities
wwpdb.org
Electron Microscopy Data Bank (EMDB)
•! Founded at EMBL-EBI in 2002
•! Since 2007 - operated jointly by PDBe and RCSB
•! EMDataBank resource (NCMI, PDBe, RCSB) funded by NIH since 2007
•! Additional funding from EMBL-EBI, BBSRC, EU and MRC
•! Own Advisory Committee and Validation Task Force
emdatabank.org – pdbe.org/emdb
Roles of PDBe, wwPDB and EMDataBank
Gutmanas et al., Acta Cryst. D69, 710 (2013)
Roles PDBe wwPDB EMDataBank
Community interactions
CCP4, CCPN,
CCP-EM
(V)TFs, IUCr,
journals, …
EM-VTF,
workshops, portal
Challenges CAPRI, CASD-NMR (CASP) EM modelling
Formats
(3D cellular imaging
data?)
PDB, PDBx,
working groups
Maps, FSC,
segmentations
Data models & ontologies
Crystallisation
ontology, CCPN
PDBx
EMDB data model,
EMX
New methods
(SXT? 3DSEM?
CLEM?)
SAS? Hybrids? (?)
Deposition, annotation, validation,
archiving, distribution
(3D cellular imaging
archive?)
PDB, BMRB EMDB
Integration SIFTS and more PDB annotation EMDB annotation
Advanced services exposing
structural information
Many! - PDB
in 2014/2015
More! Bigger! Better!
PDB in 2014/2015
•! More!
•! Over 100,000 entries in the active archive
•! More than 10,000 of these are NMR entries
•! Bigger!
•! Large structures now released intact (not SPLIT)
•! Required move to mmCIF/PDBx (“Formageddon”)
•! Better!
•! New common Deposition & Annotation system
•! Validation reports
•! Archive remediation
!"#!"#$%
&%
14 May 2014: 100,000+ PDB entries! Over 10,000 NMR entries in the PDB
•! Including >100 membrane proteins
Large structures now released intact
•! Limitations of PDB format
necessitated SPLIT entries
•! mmCIF/PDBx format does
not have these limitations
•! Workshop at EMBL-EBI in
2011 – decision to support
PDBx in major refinement
packages and to switch to
PDBx as the distribution
format for the PDB archive
mmcif.wwpdb.org
Large structures to be released intact
•! 2013
•! Large structures can be deposited intact (ADIT,
AutoDep)
•! Distribution as PDBx and PDBML in separate ftp area,
and also as SPLIT entries in regular archive
•! 2014
•! New wwPDB Deposition & Annotation system designed to handle
large structures
•! July: previously SPLIT entries reunited and distributed in parallel in
separate ftp area
•! 10 December: “Formageddon”!
•! Switched to PDBx as archive distribution format
•! SPLIT entries replaced by reunited entries
mmCIF/PDBx workshop for programmers
EMBL-EBI, November 2013
New Deposition & Annotation system
•! Designed to handle large structures from any combination of techniques
(X-ray, NMR, EM)
•! Jointly developed by wwPDB partners
•! Replaces AutoDep, ADIT, EMDep
•! Validation integral part of deposition and annotation
•! X-ray module in production since January 2014
•! X-ray/NMR/EM/neutron coming in 2015
http://deposit.wwpdb.org/deposition
!"#!"#$%
'%
Validation of structural data and models is
crucial
Archive remediation
•! Never-ending process to improve the content, description
and consistency of data in the PDB archive
•! Incidental: affecting one or a few entries, often acting on
information from users
•! Archive-wide
•! In the past: literature citations, sequence references, taxonomy
assignments, peptide ligands (inhibitors, antibiotics)
•! Future: carbohydrates, protein modifications,
metal-containing ligands
PDBe in 2015
Cooler!
PDBe in 2015 – lots of cool stuff coming soon!
•! Cooler!
•! Redesigned PDB and EMDB entry pages, based on extensive
user-testing
•! Several unique features to enhance the content of the archives
•! New search/browse system that makes finding things easier
and allows for further analysis
•! Redesigned “corporate” web site (pdbe.org)
•! For developers: all our data will be available through an API
•! A bit later: validation portal to make analysing and validating Xray,
NMR and EM structures easier
•! Lots of 3DEM-related activity, too!
PDBe website needed a make-over
•! Many improvements:
•! Quality of underlying data (literature, assemblies, …)
•! Unique data (e.g., through mining full-text articles)
•! Data access through API (PDB, EMDB, SIFTS, PISA,
validation)
•! Data discovery and analysis (search/browse)
•! User-driven redesign of entry pages
•! User interfaces (e.g., linked 1ry/2ry/3ry structure viewer)
•! Javascript-based 3D viewers (instead of Java applets)
•! Improved “corporate” website
Expression data
Genomic context
Bio. Process
Bio. Function
Pathway
Disease
Taxonomy
Cellular Location
Protein
Enzymes
Active site
Small mol. binding
Small mol.
Representative lig. conformer
Related str.
Publication
Macromolecular Structure
Sample Info.
Exp. info
Representative Binding site
Quality of Structure
Representative structure
Seq. Similar
Struct. Similar
Flexible/disorder
Sec. structure
Seq. motif
Biol. Complex
Interfaces
Struct. Domain
PDBe “tubemap” relationships
Close relation (>7 users)
Relation (>4 users)
Relation within a block
Relation between blocks
UniProt card-sorting names:
Function
Subcellular location
Names and taxonomy
Family and domains
!"#!"#$%
$%
Citation information
•! Primary citation:
figures with
captions if full-text
publication
•! Reviews/articles
that reference
primary citation or
mention PDB entry
in full text (datamining
by Europe
PMC)
Entry images
•! Display “preferred” assembly ! smallest assembly that contains all
entities (i.e., distinct molecules)
•! If no such assembly, use the one with the most entities
•! … and find out if annotation is correct or not
Colour by chain Colour by entity
PDB entry 3MIN
Entry/assembly images – virus structures
1NQX 3ZFF
!"#!"#$%
(%
PDBe – new search/browse system
•! Provides suggestions based on reference data (auto-
complete)
•! E.g., enzyme names, GO terms, protein names, CATH
classification
•! Supports categorisation of result set (“faceting” – like
Amazon)
•! By method of structure determination, release date, resolution,
ligands, author, CATH domain, …
•! Supports multiple views of the result set
•! Entry view, molecule view (more to come)
Auto-complete interface Search results – entry view
Search results – entry view Search results – macromolecule view
!"#!"#$%
)%
Search results – compound view Search results – protein-family view
PDBe – new “corporate” web site The PDBe API
•! Application
Programming Interface
•! PDB, EMDB, CCD,
SIFTS, PISA, SSM,
validation, topology,
search system
•! Used to populate new
PDBe entry pages
•! Available to external
developers
Info: sameer@ebi.ac.uk
Delivering, analysing and validating
experimental data at PDBe
EMDB volume viewer
pdbe.org/emd-1788
!"#!"#$%
*%
EMDB visual map analysis
pdbe.org/emd-1788
Slice viewer for tomograms
Collaboration with OME (Dundee)
pdbe.org/emd-1053
OLDERADO
•! Helps you analyse
NMR ensembles
•! How many
clusters?
•! Representative
models?
•! Which rigid
domains?
•! pdbe.org/olderado
Vivaldi
pdbe.org/vivaldi
ChartsInteractive3Dviewer
ModelselectionSummaryMoregraphs
Vivaldi++ (design study) Prototype/Mock-up - “EDS+”
+,--./0123%45/6%
7898:;80<280%=->1%
7/?5<,/.-/4/-%
@,8-513%
7/?5<,/.:/0125:%
45/6%651;%/-/:12>0%
</0?513%98=?%
7/?5<,/%A2>,=%
?/-/:1>2%
7/?5<,/%50%B>:,?%
C4/28--%@,8-513%
!"#!"#$%
D%
Ligand-validation prototype PDBe - validation portal
•! Unified delivery of experimental data and validation
information for X-ray, NMR and EM
•! Visualisation (1D, 2D, 3D, linked)
•! Selections
•! Terminology
•! Look-and-feel
(Mock-up!)
PDBe - validation portal
•! If you can use one, you can
use them all!
•! Lowers barrier for non-
experts
•! Interactive pages will include
•! Interactive 3D viewer
•! Graphs, plots and tables,
tightly coupled to 3D viewer
•! Static entry pages will
provide intermediate level of
detail
(Bob Hanson of Jmol/JSmol fame)
The future
Blobbier!
The world is changing
•! Biology
•! Structural biology
•! Bioinformatics
•! ICT
•! Funding landscape
•! Emerging nations
•! How about structural biology
archives?
Challenges facing archives
•! Increasing size and complexity of structures and data
•! More heterogeneous information at a range of scales
•! Need to coordinate across disciplines
•! Need to integrate structural data on scales from atoms to
cells
•! Need to integrate structure with other biological and
chemical data
•! Need to deliver appropriate structure data to non-experts (in
context of their work)
•! (Funding)
!"#!"#$%
#E%
Structural biology archives today
Technique Models Data
X-ray PDB PDB
NMR PDB BMRB + PDB
3DEM PDB EMDB
•! Simple world
•! 3 techniques
•! 3 archives
•! atomistic models
•! Trends
•! Many techniques, ranging in scale from atoms to cells
•! Many types of model: atomistic, “residue-istic”, map
segmentations (with contrast), envelopes (no contrast),
geometric shapes
•! Hybrid methods (integrative modelling): mixed models
(atomistic/blobby; experimental/theoretical) and heterogeneous
data with variable information content
Structural biology archives tomorrow?
Technique Models Data
X-ray PDB PDB
NMR PDB BMRB + PDB
3DEM PDB + “ModelDB” + “BlobDB” EMDB
SAXS/SANS PDB + “SASDB” “SASDB” + PDB
Theoretical models “ModelDB” n/a
Hybrid methods PDB + “ModelDB” + “BlobDB” “BlobDB”? PDB?
SXT, 3DSEM, … “BlobDB” “3DCellDB”? EMDB?
CLEM PDB + “ModelDB” + “BlobDB” “3DCellDB”? EMDB?
•! “ModelDB” = theoretical atomistic models?
•! “BlobDB” = non-atomistic models and hybrid methods data?
•! “SASDB” = SAXS/SANS data and models not archived elsewhere?
•! “3DCellDB” = 3D cellular imaging data and segmentations?
The key will be “integration”
Image: Zeev-Ben-Mordehai et al., 2014
Integrating imaging and 3D structural data
Instruct & Euro-BioImaging
•! Enriching biology with structural information
Scales / Methods
Atoms Molecules Machines SamplesCells
NMR X-ray EM CLEMSAXS ET 3DSEMSXT
Integrating imaging and 3D structural data
Elixir
•! Annotating and linking structure through biological information
ChEBI ENA/UniProt ChEMBL/IntAct Reactome
Information / Resources
Chemistry Sequences Interactions PathwaysVariation
Ensembl
Integrating imaging and 3D structural data
• Archive 3D cellular imaging data (EM, ET; later: SXT, 3DSEM, CLEM)
• Annotate using bioinformatics resources, classifications and ontologies (UniProt, GO)
• Link to 3D molecular structure data (X-ray, NMR, EM) in PDB and EMDB
• Provide tools to make the information easily accessible and to facilitate discovery
Sample
Atomic details
!"#!"#$%
##%
Archiving cellular imaging data
•! EMDB: mainly single-particle, tomogram-averaging and
tomographic data
•! EMPIAR:
•! Archive raw data associated with EMDB entries (for validation,
methods development, teaching, …)
•! Develop tools for segmentation annotation of 3D imaging data
•! Develop integrated viewer for structural data on scales from
molecules to cells (or atoms to small samples)
•! Future (if funded): develop archive for SXT and 3DSEM data
and accommodate CLEM data in EMDB/EMPIAR
EMPIAR
pdbe.org/empiar
Prototype browser (EMD-2179)
PDBe team
Funding & acknowledgments
•! Many thanks to many colleagues at PDBe, EMBL-EBI,
wwPDB, EMDataBank and elsewhere!
Questions & discussion…
http://pdbe.org/
http://www.facebook.com/proteindatabank
http://twitter.com/PDBeurope
http://youtube.com/user/ProteinDataBank
http://wwwdev.ebi.ac.uk/pdbe/
http://wwwdev.ebi.ac.uk/pdbe/entry/search/index
http://wwwdev.ebi.ac.uk/pdbe/api/doc/api-index.html
http://wwwdev.ebi.ac.uk/pdbe/entry/pdb/4i91
!"#!"#$%
#!%
If you see this slide, I’ve gone too far