Protein interaction networks from yeast to human Lars Juhl Jensen EMBL Heidelberg
Jun 27, 2015
Protein interaction networksfrom yeast to human
Lars Juhl Jensen
EMBL Heidelberg
Current Opinion in Structural Biology, 2005
STRING
integrate diverse evidence
functional interactions
Bork et al., Current Opinion in Structural Biology, 2005
hundreds of proteomes
evolution
statistics
(the original sin)
prokaryotes
genomic context methods
gene fusion
gene neighborhood
phylogenetic profiles
Cell
Cellulosomes
Cellulose
eukaryotes
integrate diverse datasets
Jensen et al., Drug Discovery Today: Targets, 2004
curated knowledge
MIPSMunich Information center
for Protein Sequences
KEGGKyoto Encyclopedia of Genes and Genomes
STKESignal Transduction Knowledge Environment
Reactome
literature mining
co-mentioning
NLPNatural Language Processing
MEDLINE
SGDSaccharomyces Genome Database
The Interactive Fly
OMIMOnline Mendelian Inheritance in Man
primary experimental data
microarray expression data
GEOGene Expression Omnibus
physical protein interactions
BINDBiomolecular Interaction Network Database
MINTMolecular Interactions Database
GRIDGeneral Repository for Interaction Datasets
DIPDatabase of Interacting Proteins
HPRDHuman Protein Reference Database
problems
many sources
(different gene identifiers)
many types of evidence
questionable quality
not directly comparable
spread over many species
huge synonyms lists
calculate raw quality scores
calibrate vs. gold standard
KEGGKyoto Encyclopedia of Genes and Genomes
von Mering et al., Nucleic Acids Research, 2005
transfer based on orthology
combine all evidence
Bork et al., Current Opinion in Structural Biology, 2005
Networ-Kin™
Phospho-peptidedata (from MS)
Predict the kinase class(NetPhosK and Scansite)
the human interactome
yeast two-hybrid
1936
13
4
4
1385
65
18465
Stelzl et al. Rual et al.
Small-scale studies
32
0
3
4
18
4
23
Stelzl et al. Rual et al.
Small-scale studies
62 8 39
Small-scale studies
Stelzl et al. Rual et al.
852
17
473
432
69
260
3.5% and 21% sensitivity
in a couple of years
the human interactome
100% = 1/5?
the yeast interactome
five years ago
yeast two-hybrid
1150
117
117
72
4053
118
4469
Uetz et al. Ito et al.
Small-scale studies
162
53
34
72
180
29
338
Uetz et al. Ito et al.
Small-scale studies
511 189 616
Small-scale studies
Uetz et al. Ito et al.
439
178
759
897
190
1347
19% and 12% sensitivity
three years ago
complex purification and MS
14186
784
1178
287
19492
230
3475
Gavin et al. Ho et al.
Small-scale studies
2341
125
656
287
2725
42
149
Gavin et al. Ho et al.
Small-scale studies
4431 1465 5041
Small-scale studies
Gavin et al. Ho et al.
14913
1071
632
8047
517
746
63% and 41% sensitivity
today
combining all available data
scoring and filtering
de Lichtenberg et al., Science, 2005
de Lichtenberg et al., Science, 2005
de Lichtenberg et al., Science, 2005
de Lichtenberg et al., Science, 2005
de Lichtenberg et al., Science, 2005
the challenge
how to get from here …
1936
13
4
4
1385
65
18465
Stelzl et al. Rual et al.
Small-scale studies
… to there …
de Lichtenberg et al., Science, 2005
yeast two-hybrid
complex purification and MS
scoring and filtering
Acknowledgments
• The STRING team (EMBL)– Christian von Mering
– Berend Snel
– Martijn Huynen
– Sean Hooper
– Mathilde Foglierini
– Julien Lagarde
– Peer Bork
• Literature mining project(EML Research)– Jasmin Saric
– Rossitza Ouzounova
– Isabel Rojas
• Cell cycle studies (CBS)– Ulrik de Lichtenberg
– Thomas Skøt Jensen
– Søren Brunak
• Networ-Kin™ (Mt. Sinai)– Rune Linding
– Tony Pawson
• Inspiration for presentation– Lawrence Lessig
– Dick Clarence Hardt
– Anders Gorm Pedersen
Thank you!