Database Connection and Annotating Models

Database connection

Import model from BioModels.net

Database connection

Akira Funahashi Keio University, Japan

Database connection Database connection Import model from BioModels.net Import model from JWS Online Database connection Database connection

Import model from PANTHER Import model from PANTHER

Database connection Comparison Import model from PANTHER •BioModels.net •Curated •Can simulate •No CellDesigner Layout •Panther DB •Curated •Can not simulate •CellDesigner Layout SABIO-RK CellDesigner ⇔ SABIO RK •Web-accessible database •Users can import additional information •http://sabio.h-its.org/ to each object (reaction) on-the-ﬂy •Contains information about biochemical •SBML (Systems Biology Markup Language) reactions, related kinetic equations and is used to exchange information parameters CellDesigner

E Name, EC number

S P Vmax[S] Km +[S] kinetic law, parameters, function / unit deﬁnitions

Integration Example •Import kinetic law, parameters to the model from SABIO-RK Example Example

Example Example Example

Annotating a model

Akira Funahashi Keio University, Japan

V IEWPOINT Peripheral insulin resistance contributes to Jnk1 in obese mice, or inhibition of serine mice increases insulin sensitivity, revealing type 2 diabetes, but ␤-cell failure is the es- kinases by salicylates, reduces Ser phosphor- targets for inhibitor design. However, inhibi- sential feature of all types of diabetes. ␤ cells ylation of IRS proteins and reverses hyper- tion of SHIP2 or pTEN might be risky be- frequently fail to compensate for insulin re- glycemia, hyperinsulinemia, and dyslipide- cause they stabilize PI(3,4,5)P3, which can sistance, apparently because the IRS2-branch mia in obese rodents by sensitizing insulin stimulate cell growth. By contrast, PTP1B of the insulin and IGF signaling cascade is signaling pathways (18). resides in the endoplasmic reticulum, where it also essential for ␤-cell growth, function, and Ubiquitin-mediated degradation of IRS1 dephosphorylates the insulin receptor during inter- survival (14). Increased expression of IRS2 and IRS2 also promotes insulin resistance nalization and recycling to the plasma membrane in ␤ cells promotes compensatory insulin secre- (Fig. 1). IL-6 secreted from leukocytes and (21). This specialized mechanism appears to limit tion in obese mice and prevents ␤-cell destruction adipocytes increases expression of SOCS dangerous side effects of PTP1B inhibitors. induced by streptozotocin—adrugthatinduces proteins, which are best known for their abil- The close association between obesity and type 1 diabetes; IRS2 also improves the survivalIGFity to inhibit cytokine signaling receptor signaling. insulin resistance, and their progression topathway type IGF signaling pathway

and function of islet transplants in mice. Drugs However, SOCS1 and SOCS3 also recruit an 2diabetes,isaserioushealthproblem.Whether

that increase IRS2 synthesis might be useful treat- elongin BC-based ubiquitin ligase into the better management of chronicReview inflammation can Trends in Pharmacological Sciences March 2012, Vol. 33, No. 3 ments for diabetes. For improve insulin action,

example, expression of promote ␤I -cell func- Key:

Induces the IRS2 gene in ␤ tion, andInsuli restoren IGF- centralI/II Hsp90 client

RTK

cells is increased by nervous system appe-transph osphorylation Inhibits Attenuated path agonists that stimulate tite control is an impor- I I adenosine 3Ј,5Ј-mono- tant area ofα investiga-α α α α α Plasma membrane β β β β β β PIP PIP phosphate (cAMP) tion. Finding drugs that 2 3 PIP3 PDK1

Shc P P production—including stimulate IRS2 synthe- P P P PI3K PDK1 P P P P Grb2 IRS-1/2

glucagon-like peptide– sis or promote its sig- SoS PTEN IR IGF- IR 1orglucoseitself— naling might be a good p53 through pathways that starting point. Howev- Ras Ras PKC P Raf-1 P GTP GDP activate the transcrip- er, too much insulin ac- MDM2 P PI3K pathway PI3K tion factor CREB (15) tion might be detrimen- P TSC2 on January 17, 2007 Akt P MEK1 P

(Fig. 1). tal, so future work must TSC1

Understanding dys- better resolve the net- P ERK-1/2 P mTOR 4EBP1 FoxO3a P BAD P GSK-3β regulated insulin sig- work of insulin re- Raptor eIF4E (inactive)

aling is an impor- sponses that are gener-pathway MAPK/ERK

P Cytochrome C c-Myc P P Fas tant goal because it ated in various tissues, p70S6K 4EBP1 release + /Raptor pathway /Raptor causes a cohort of and attempt toc-Fos distin-P eIF4E systemic disorders— guish the ones that mTOR

dyslipidemia, hyper- prolong health from Membrane docking/integration Membrane

tension, cardiovascular the ones that mightCell proliferation Cell cycle arrest,

pro-survival transcription www.sciencemag.org and translation autophagy and apoptosis disease, stroke, blind- diminish it. ness, kidney disease, Fig. 1. Regulation of insulin and IGF signaling. Insulin and IGF1 receptors form hybrids that BAX P

References modulate the selectivity and afﬁnity for insulin and insulin-like growth factors (IGF1 and IGF2). female infertility, and Met GSK-3β 1. M. Bliss,Met-tRNTheAi Discovery+ eIF2 P eIF2B P neurodegeneration. Insulin or IGF binding stimulates tyrosine autophosphorylation in the receptor ␤ subunits, which (active) activates the kinase and recruits cellular substrates—IRS1 and IRS2—for tyrosine phosphorylation. of Insulin (Univ. of ChicagoHSF1 Press, Chica- P

Subtle genetic poly- Recruitment is regulated by serine phosphorylation of the IRS proteins, which inhibits the Hsp27 Tau n CREB P

go, IL, 1982). morphism influenc- ERK-1 P (aggregate) interaction between its PTB domain and the phosphorylated receptor. Proinﬂammatory cytokines 2. M. Kasuga et al., Nature P es lifelong insulin increase the synthesis of SOCS1 or SOCS3, which promote ubiquitination and degradation of IRS1 298,667(1982).HSF1 PKC P I GLUT4 GSK-3β or I sensitivity, whereas and IRS2. Production of cAMP enhances expression of IRS2 through the activity of phosphorylated 3. P. Zimmet et al., Na- Downloaded from I vesicle (active) CREB. Tyrosine phosphorylation of IRS1 or IRS2 recruits and activates various SH2 domain– ture 414HSF1, 782P (2001). Akt P Neurotransmission

distinct monogenic P

4. M. White, Insulin Signal- • Synaptic plasticity

containing proteins, including the PI 3-kinase, which activates the PKB cascade. Abbreviations: pY, Pancreatic β-islet cells

Hsp27 • Transmitter release Adipose, muscle and liver

disorders are diffi-

ing Pathway, Sci. STKE cerebellum, sensorimotor

Insulin vesicles HSR

phosphotyrosine; pS, phosphoserine; PKC␭/␨, protein kinase C ␭ or ␨; E2, ubiquitin conjugating (functional) cult to identify and (Connections Map, as and reuptake cortices, hippocampus, enzymes; TNF␣R, tumor necrosis factor–␣ receptor; GLP-1R, glucagon-like peptide–1 receptor; pituitary, hypothalamus seen November 2003), are usually associat- TRENDS in Pharmacological Sciences

IL6R, interleukin-6 receptor; for other abbreviations, see the text. http://stke.sciencemag.

ed with rare metabol- org/cgi/cm/cmp_12069.Figure 2. Insulin, IGF-I and IGF-II binding to IR and IGF-IR. Ligand binding triggers receptor autophosphorylation and RTK activation, permitting IRS-1, IRS-2 or Shc

phosphorylation. Mitogenic signaling pursues via Grb2 activation of the GEF (guanine nucleotide exchange factor) SoS, enabling Ras/Rheb activation. Activated Ras/Rheb

ic diseases. Acute and 5. F. Frasca et al., Mol. Cell. Biol.initiates19,3278 the MAPK/ERK (1999).pathway through Raf-1 phosphorylation and enhances mTOR/raptor signaling, collectively increasing cell growth, proliferation, protein expression

and viability. Induction of the PI3K pathway enhances metabolic insulin effects and mitigates pro-apoptotic and anti-mitogenic signaling. Hsp90 client proteins are indicated

chronic inflammation cause insulin resistance, IRS-protein complexes to mediate ubiquiti- 6. R. S. Savkur et. al., Nature Genet.with a29pink,bar 40and (2001).serve crucial roles in regulating insulin/IGF signaling. Potential modulation effects of insulin/IGF signaling components upon Hsp expression and

7. M. F. White et al., Nature 318function, 183 are (1985).shown in the black box. Abbreviations: BAD, Bcl-2-associated death promoter; BAX, Bcl-2-associated X; 4EBP1, eIF4E binding protein-1; CREB, cAMP-response

which provides a frameworkInsulin to understand Signaling how nylation (19). in Ubiquitin-mediated Health degrada-and Disease Heatelement binding; eIF2, shockeukaryotic initiation factor response2; eIF4E, eukaryotic translation initiation factor-4E;andERK-1/2, extracellular-signal-regulatedinsulin-kinase-1/2; FoxO3a, forkhead

8. X. J. Sun et al., Nature 352, 73box (1991). ‘other’ 3a; GLUT4, glucose transporter 4; Grb2, growth factor receptor-bound protein 2; GSK-3b, glycogen synthase kinase-3b; HSF1, heat shock factor-1; Hsp, heat http://www.sbgn.org/

diet, physiological stress, and obesity promote tion of IRS1 and IRS2 might be a general shock protein; HSR, heat shock response; IGF-IR, insulin-like growth factor-I receptor; IR, insulin receptor; IRS-1/2, insulin receptor substrate-1/2; MAPK, mitogen-activated

Met

9. M. Schubert et al., J. Neurosci.protein23, kinase; 7084MEK1, (2003).MAPK/ERK kinase 1; Met-tRNAi , initiator methionyl-tRNA; mTOR, mammalian target of rapamycin; raptor, regulatory-associated protein of mTOR;

Science 302 (5651), 2003, 1710. associated neurodegeneration

insulin resistance. Proinflammatory cytokines, mechanism of cytokine-induced insulin resis- 10. S. Ogg et al., Nature 389, 994MDM2, (1997).murine double minute-2; PDK1, phosphoinositide-dependent kinase-1; PI3K, phosphatidylinositol 3-kinase; PIP2, phosphatidylinositol 4,5-bisphosphate; PIP3,

phosphatidylinositol 3,4,5-triphosphate; PKC, protein kinase C; PTEN, phosphatase and tensin homolog; RTK, receptor tyrosine kinase; Shc, Src homology 2 domain- Documents/PD_L1_Examples

including IL-6 and tumor necrosis factor–␣ tance that contributes to diabetes and ␤-cell 11. J. Nakae et al., Dev. Cell 4, 119containing; (2003). SoS, son of sevenless; TSC, tuberous sclerosis complex. (TNF-␣)thataresecretedfromleukocytesdur- failure. Modern genomic approaches have re- 12. T. Kitamura et al., J. Clin. InvestTrends110, 1839 (2002). in Pharmacological Sciences, ing inflammation, are also produced in adipose vealed new cytokines secreted from adipo- 13. P. Puigserver et al., Nature 423, 550 (2003). 14. A. M. Hennige et al., J. Clin. Invest.33112(3),, 1521 (2003). 2012, 129–137 tissue. TNF- promotes serine phosphorylation cytes that directly influence nutrient ho- ␣ 15. U. S. Jhala et al., Genes Dev. 17, 1575 (2003). 133 of IRS1 and IRS2, which correlates closely meostasis and insulin sensitivity, including 16. G. S. Hotamisligil, B. M. Spiegelman, Science 259,87 with insulin resistance (16). Although TNF-␣ leptin, resistin, and adiponectin. How adi- (1999). regulates various kinases, the c-Jun N-terminal ponectin signaling stimulates insulin signal- 17. V. Aguirre et al., J. Biol. Chem. 277, 1531 (2002). kinase (Jnk) is a prominent effector because it ing deserves attention (20). 18. J. Hirosumi et al., Nature 420, 333 (2002). 19. D. L. Krebs, D. J. Hilton, Sci. STKE. 2003, E6 (2003). binds to IRS1 and IRS2 and phosphorylates Protein or lipid phosphatases, including 20. T. Yamauchi et al., Nature 423, 762 (2003). serine residues that inhibit insulin-stimulated PTP1B, SHIP2, or PTEN, inhibit insulin sig- 21. F. G. Haj, P. J. Verveer, A. Squire, B. G. Neel, P. I. tyrosine phosphorylation (17). The knockout of naling (Fig. 1). Disruption of each gene in Bastiaens, Science 295, 1708 (2002).

www.sciencemag.org SCIENCE VOL 302 5 DECEMBER 2003 1711 Exercise Exercise Create a following model on CellDesigner Search Database from CellDesigner

http://www.sbgn.org/Documents/PD_L1_Examples http://www.sbgn.org/Documents/PD_L1_Examples

Database Connection Exercise Search Database by Notes: Add UniProt ID for Proteins, PubMed ID for reactions and call “Connect to UniProt” PubMed: PMID: 123456

Entrez Gene: GeneID: 4015 UniProt: P05019

UniProt: P08069

PMID: 14657487 PMID: 22172248 MIRIAM annotation MIRIAM annotation

http://www.ebi.ac.uk/miriam/main/qualiﬁers/

PERSPECTIVE is IGF_IGFR hasPart: IGF IGF_IGFR hasPart: IGFR Minimum information requested in the annotation of biochemical models (MIRIAM)

Nicolas Le Novère1,15, Andrew Finney2,15, Michael Hucka3, Upinder S Bhalla4, Fabien Campagne5, hasPart Julio Collado-Vides6, Edmund J Crampin7, Matt Halstead7, Edda Klipp8, Pedro Mendes9, Poul Nielsen7, Herbert Sauro10, Bruce Shapiro11, Jacky L Snoep12, Hugh D Spence13 & Barry L Wanner14

Most of the published quantitative models in biology are During the genomic era we have witnessed a vast increase in availabil- lost for the community because they are either not made ity of large amounts of quantitative data. This is motivating a shift in available or they are insufficiently characterized to allow the focus of molecular and cellular research from qualitative descrip- them to be reused. The lack of a standard description format, tions of biochemical interactions towards the quantification of such

http://www.nature.com/naturebiotechnology lack of stringent reviewing and authors’ carelessness are interactions and their dynamics. One of the tenets of systems biology isPartOf the main causes for incomplete model descriptions. With is the use of quantitative models (see Box 1 for definitions) as a mech- today’s increased interest in detailed biochemical models, anism for capturing precise hypotheses and making predictions1,2. it is necessary to define a minimum quality standard for Many specialized models exist that attempt to explain aspects of the the encoding of those models. We propose a set of rules for cellular machinery. However, as has happened with other types of bio- curating quantitative models of biological systems. These logical information, such as sequences, macromolecular structures or rules define procedures for encoding and annotating models represented in machine-readable form. We believe their application will enable users to (i) have confidence that Box 1 Glossary curated models are an accurate reflection of their associated reference descriptions, (ii) search collections of curated Some terms are used in a very specific way throughout the article. hasVersion (isoform) models with precision, (iii) quickly identify the biological We provide here a precise definition of each one. phenomena that a given curated model or model constituent 2005 Nature Publishing Group Quantitative biochemical model. A formal model of a biological

© represents and (iv) facilitate model reuse and composition system, based on the mathematical description of its molecular into large subcellular models. and cellular components, and the interactions between those components. Encoded model. A mathematical model written in a formal 1 European Bioinformatics Institute, Hinxton, CB10 1SD, UK. machine-readable language, such that it can be systematically 2Physiomics PLC, Magdalen Centre, Oxford Science Park, Oxford, OX4 4GA,K. 3Control and Dynamical Systems, California Institute of parsed and employed by simulation and analysis software without Technology,Pasadena, California 91125, USA. 4National Centre for Biological further human translation. Sciences, TIFR, UAS-GKVK Campus, Bangalore 560065, India. 5Institute isVersionOf (superclass, parent) MIRIAM-compliant model. A model that passes all the tests and for Computational Biomedicine, Weill Medical College of Cornell University, New York, New York 10021, USA. 6Center for Genomic Sciences, Universidad fulfills all the conditions listed in MIRIAM. Nacional Autónoma de México, Av. Universidad s/n, Cuernavaca, Morelos, Reference description. A unique document that describes, or 62100, Mexico. 7Bioengineering Institute and Department of Engineering Science, The University of Auckland, Private Bag 92019, Auckland, New references the description of the model, the structure of the Zealand. 8Max-Planck Institute for Molecular Genetics, Berlin Center for model, the numerical values necessary to instantiate a simulation Genome based Bioinformatics (BCB), Ihnestr. 73, 14195 Berlin, Germany. from the model, or to perform a mathematical analysis of the 9 Virginia Bioinformatics Institute, Virginia Tech, Washington St., Blacksburg, model, and the results one expects from such a simulation or Virginia 24061-0477, USA. 10Keck Graduate Institute, 535 Watson Drive, Claremont, California 91711, USA. 11Jet Propulsion Laboratory, California analysis. 12 Institute of Technology, Pasadena, California 91109, USA. Triple-J Group Curation process. The process by which the compliance of an for Molecular Cell Physiology, Department of Biochemistry, Stellenbosch University, Private Bag X1, Matieland 7602, South Africa. 13Department of encoded model with MIRIAM is achieved and/or verified. The IGF isPartOf: IGF_IGFR Scientific Computing & Mathematical Modeling, GlaxoSmithKline Research curation process may encompass some or all of the following & Development Limited, Medicines Research Centre, Gummels Wood tasks: encoding of the model, verification of the reference 14 Road, Stevenage, Herts, SG1 2NY, UK. Purdue University, Department of correspondence and annotation of the model. Biological Sciences, Lilly Hall of Life Sciences, 915 W. State Street, West Lafayette, Indiana 47907-2054, USA. 15These authors have contributed Reference correspondence. The fact that the structure of a equally to the work. Correspondence should be addressed to N.L.N. model and the results of a simulation or an analysis match the (e-mail: [email protected]). information present in the reference description. Published online 6 December 2005; doi:10.1038/nbt1156

NATURE BIOTECHNOLOGY VOLUME 23 NUMBER 12 DECEMBER 2005 1509

Minimum information requested in the annotation of biochemical models (MIRIAM) IGFR isPartOf: IGF_IGFR Nature Biotechnology 23, 1509 - 1515 (2005)

Exercise Notes or MIRIAM? Add MIRIAM annotation

hasPart: UniProt: P05019 hasPart: UniProt: P08069 CellDesigner Notes Easy to add (text) MIRIAM Tool neutral (SBML) Precise annotation is: ChEBI: CHEBI:15422 is: ChEBI: CHEBI:16761 Summary Introduction of CellDesigner What kind of model you can build

Create a full version of Mathematical model IGF signaling pathway and annotate! Pathway map How to build a model with CellDesigner From scratch Import a model from BioModels.net Import kinetic law and parameters from SABIO-RK

Acknowledgement Acknowledgement SBML SABIO-RK (HITS gGmbH) * SBML community * Martin Golebiewski Grant-in-Aid for Young Scientists SBGN * Lenneke Jong (21700328) * SBGN community * Wolfgang Mueller * Isabel Rojas Grant-in-Aid for Scientiﬁc Research on SBW * Frank Bergmann CellDesigner Innovative Areas (23136513) * Herbert Sauro (SBI) * Yukiko Matsuoka Grant-in-Aid for Scientiﬁc Research SOSlib * Samik Ghosh (24300112) * Rainer Machne * Hiroaki Kitano * Christoph Flamm (Keio Univ.) * Noriko Hiroi ERATO Kawaoka Infection-Induced Simulation Core (Univ. of Tuebingen) * Kazushige Nakamura Host Response Project * Andreas Dräger * Kei Sumiyoshi * Kota Mashimo COPASI (Univ. of Heidelberg) * Kaito Ii * Ralph Gauges * Sven Sahle * Akiya Jouraku (NIAS) * Ursula Kummer * Norihiro Kikuchi (MKI)