Session outline
1.
Standards and the problem of data integration
•
2.
Example: PSICQUIC and the PSICQUIC game
Introduction to ontologies.
•
Exploring the Gene Ontology
-Coffee break-
3.
Introduction to protein-protein interactions (PPIs)
•
IntAct: the molecular interactions database at the EBI
-Lunch break-
4.
Introduction to pathways
•
Reactome: a database of human biological pathways
-Coffee break-
5.
Network representation and analysis: strategies and limitations
•
Network generation and analysis through Cytoscape and PSICQUIC
Introduction to protein-protein
interactions (PPIs)
A couple of definitions…
Protein-protein interactions (PPIs): physical and selective contacts
that happen between pairs of proteins, in certain molecular regions and
in a defined biological context.
Interactome: the totality of PPIs that happen in a cell / in an organism /
in a specific biological context...
Protein-protein interaction network: Graphical representation of a
group of PPIs in which proteins are represented as nodes and
interactions as edges.
EMBL-EBI
Why protein-protein interactions?
Gene level
DNA
Protein level
RNA
1 protein
=
1 protein
n functions
=
=
function
n1networks!
WRONG!
1. To predict a protein biological function
• “guilt by association”
• proteins with similar functions should cluster together
2. To improve characterization of protein complexes and pathways
• interaction networks work as a draft map that brings detail to
biological processes and pathways
EMBL-EBI
Guilt by association
Cyclin-dependent kinase
Role in transcription regulation
Proto-oncogene
Transcription factor
Histone methyltransferase
Role in early development
and hematopoiesis
Cyclin
Regulation of cyclin kinase,
transcription regulation and
cell cycle control
Mediator of RNA
polymerase
transcription activity
Role in transcription
regulation
Transcriptional
modulator
Transcription
regulation dependent
on kinases
Something to do
with transcription
and cell cycle control
Putative oncoprotein
Involved in transcription
regulation
Putative oncoprotein
Transcription activation
Receptor-activated
transcription modulator
Transcription modulation,
signal transduction,
activated by kinases.
Role in inhibition of wound
healing
EMBL-EBI
Characterization of protein complexes
and pathways
Hook, B. and Schagat, T. [Internet] 2011.
Available from:
www.promega.com/resources/articles/pu
bhub/functional-proteomics-techniquesto-isolate-and-characterize-the-humanproteasome/
Glaab et al., BMC Bioinformatics, PMID: 21144022.
EMBL-EBI
Protein-protein interaction detection methods
Lowthroughput
High-throughput
Yeast-two hybrid (Y2H)
Tandem affinity purification+
mass spectrometry (TAP-MS)
No single method can accurately reproduce a true
binary interaction observed under physiological
conditions – every interaction detected
experimentally is fundamentally artefactual.
X-ray diffraction studies
EMBL-EBI
Protein-protein interaction detection methods
Braun et al., Nat Methods, 2008, PMID: 19060903.
EMBL-EBI
Types of protein-protein interactions
• Binary interactions – Two participants
• Co-localization – Two or more participants closely
located
• N-ary ints. (associations) – Complex purification
• Functional / direct ints. – E. g. enzymatic interactions
EMBL-EBI
Binary interactions
1
0
EMBL-EBI
Colocalization
1
1
EMBL-EBI
N-ary interactions (association)
Schleiff et al., Nat Rev Mol Cell Biol. 2011.
PMID: 211396380
EMBL-EBI
Direct / functional interactions
• Usually in vitro assays
• Participants IDs are known in advance (predetermined)
• Examples – enzymatic assays, SPR, crystallography,
methods using purified protein…
Images from Wikipedia:
http://en.wikipedia.org/wiki/X-ray_crystallography
http://en.wikipedia.org/wiki/Surface_plasmon_resonance
http://en.wikipedia.org/wiki/Protein_adsorption_in_the_food_industry
EMBL-EBI
Representing PPIs: interaction domains
interaction domains
Overlap in sequence ranges:
EMBL-EBI
Representing PPIs:
The problem with complexes
•
Some experimental methods generate complex data: E. g.
Tandem affinity purification (TAP)
•
There are two algorithms to transform this information into
binary data:
EMBL-EBI
Interactions databases: types
De Las Rivas & Fontanillo, PLoS Computational biology, PMID: 20589078.
EMBL-EBI
Databases: curation levels
SHALLOW
CURATION
DEEP CURATION
EMBL-EBI
Databases: curation levels
Shallow curation
BioGRID – active curation, limited number of model organisms
HPRD – active curation, human focus, predicted interactions
MPIDB – curation stopped, data re-located to IntAct, interactions in microorganisms
InnateDB – active curation, interactions related with innate immunity
Deep curation
IntAct – active curation, wide species coverage, all types of molecules
MINT – active curation, wide species coverage, PPIs only
DIP – active curation, wide species coverage, PPIs only
MPACT – curation currently stopped, limited species coverage, PPIs only
MatrixDB – active curation, extracellular matrix molecules only
BIND – curation stopped in 2006/7, wide species coverage, all types of molecules –
information getting outdated
I2D – active curation, PPIs involved in cancer
EMBL-EBI
Primary databases: coverage
Human
PPIs
coverage in
the main
public
primary
databases
(Dec 2009)
De Las Rivas & Fontanillo, PLoS Computational biology, PMID: 20589078.
EMBL-EBI
A standard for PPIs representation:
the IMEx consortium
www.imexconsortium.org
Orchard et al., Nature Methods, PMID: 22453911.
EMBL-EBI
IntAct:
The molecular interactions database
at the EBI
IntAct goals & achievements
1. Publicly available repository of molecular interactions (mainly
PPIs) - >430K binary interactions taken from >12,000 publications
(October 2013)
2. Data is standards-compliant and available via our website, for
download at our ftp site or via PSICQUIC
www.ebi.ac.uk/intact
ftp://ftp.ebi.ac.uk/pub/databases/intact
www.ebi.ac.uk/Tools/webservices/psicquic/view/main.xhtml
3. Provide open-access versions of the software to allow installation
of local IntAct nodes.
EMBL-EBI
IntAct:
Data storage schema
Entry
[A] Publication level (entry)
Publication
Experiment 1
Experiment 2
[B] Experiment level
[C] Interaction level
Interaction 1
Interaction 2
…
[D] Participant level
[E] Feature level
Participant 1
Participant 2
Features
Features
Interaction 3
Interaction 4
…
…
EMBL-EBI
UniProt Knowledge Base
Interactions can
be mapped to
the canonical
sequence…
... to splice variants...
... or to postprocessed chains
www.uniprot.org
EMBL-EBI
IntAct: PSI-MI ontology
EMBL-EBI
IntAct Curation pipeline
“Lifecycle of an
Interaction”
Sanity Checks
(nightly)
reject
Public web site
Publication
(full text)
.
accept
p2
I
p1
Curation
manual
report
exp
FTP site
check
CVs
annotate
IMEx
report
MatrixDB
curator
Mint
DIP
Super curator
EMBL-EBI
IntAct: the role of the curator
DIRECT
SUBMISSIONS
LARGE DATASETS FROM
HIGH-THROUGHPUT PROJECTS
PUBLISHED
MOLECULAR
INTERACTIONS DATA
CURATION
EMBL-EBI
DIRECT SUBMISSION
LARGE DATASETS FROM
HIGH-THROUGHPUT PROJECTS
Gene Ontology
FUNCTION
UniProtKB
PROTEIN
SEQUENCES
PUBLISHED MOLECULAR
INTERACTIONS DATA
CROSS-REFERENCES
CURATION
ChEBI
SMALL MOLECULES
Ensembl
GENOME
SEQUENCES
InterPro
FAMILIES AND
DOMAINS
Others
STRUCTURES,
ORGANISM,
TISSUE...
EMBL-EBI
IntAct as a common curation platform
General
curation,
domain int.
General
curation,
large scale
UniProt
entry
related
Extracellular
matrix
Model
organisms
Immune
system
Commercial
curation
Cellular
mechanics
Regulatory
interactions
Specific curation focus/expertise
Host – pathogen
interactions
Cardiovascular
proteins
Other DBs
Common curation platform
Specific Data Dissemination Platforms
EMBL-EBI
www.ebi.ac.uk/intact
IntAct – Home Page
EMBL-EBI
IntAct webpage-based search
EMBL-EBI
IntAct webpage-based search
Details of
interaction
Choice of UniProtKB
or Dasty View
EMBL-EBI
IntAct: changing the layout
EMBL-EBI
IntAct: download formats
EMBL-EBI
PSIMITAB Columns
MITAB 2.5 Standard columns
(15):
ID(s) interactor A & B
• Alt. ID(s) interactor A & B
• Alias(es) interactor A & B
• Interaction detection method(s)
• Publication 1st author(s)
• Publication Identifier(s)
• Taxid interactor A & B
• Interaction type(s)
• Source database(s)
• Interaction identifier(s)
• Confidence value(s)
•
MITAB 2.7 specific columns (+27):
• Expansion method(s)
• Biological role(s) of interactors
• Experimental role(s) of interactors
• Type(s) of interactors
• Properties (CrossReference) of interactors /
interaction
• Annotation(s) of interactors / interaction
• HostOrganism(s)
• Parameters of interaction
• Creation and update dates
• Checksum(s) of interactors / interaction
• Negative
• Feature(s) interactors
• Stoichiometry(s) interactors
• Participant(s) identification method(s)
EMBL-EBI
Interaction detail in IntAct
EMBL-EBI
Detailed participant information:
Dasty view
EMBL-EBI
IntAct: filtering results
EMBL-EBI
IntAct: visualizing results
as a network
EMBL-EBI
IntAct: using lists
EMBL-EBI
IntAct: browse menu
EMBL-EBI
IntAct: advanced search
...
EMBL-EBI
IntAct: MIQL syntax search
EMBL-EBI
Searching and visualizing
EMBL-EBI
Using lists
EMBL-EBI
Checking details
EMBL-EBI
Advanced search
EMBL-EBI
Advanced search: using MIQL
EMBL-EBI
Ontology search
EMBL-EBI
Link to other PSICQUIC services
EMBL-EBI
More about IntAct: “on-line” EBI courses
www.ebi.ac.uk/training/online/course/intact-molecular-interactions-ebi
EMBL-EBI
The PSICQUIC client:
A unified gate to interactomics data
Unified query client: PSICQUIC
www.ebi.ac.uk/Tools/webservices/psicquic/view/main.xhtml
PSI-MI
MIQL
Interactions
Query
input
output
PSICQUIC
PSICQUIC
Service A
PSICQUIC
Service B
PSICQUIC
Service C
PSICQUIC
Registry
EMBL-EBI
Unified query client interface: PSICQUIC view
www.ebi.ac.uk/Tools/webservices/psicquic/view/main.xhtml
EMBL-EBI
Acknowledgements
Group leader
Henning Hermjakob
Coordinator
Sandra
Orchard
Curation team
Margaret
Duesbury
Birgit
Meldal
Developing team
Rafael
Jiménez
Marine
Dumousseau
Noemí
del Toro
EMBL-EBI