The Systems Biology Markup Language SBML

1
The Systems Biology Markup Language(SBML)

• Andrew Finney1,2 Mike Hucka3
• Principal Investigators John Doyle3, Roger
  Brent2 and
• Hiroaki Kitano4
• 1 University of Hertfordshire, UK
• 2 Molecular Sciences Institute, USA
• 3 California Institute of Technology, USA
• 4 Sony Computer Science Laboratories Inc, Japan
• http//sbml.org/

2
SBMLWhat? And Why?

• What?
• Open standard for encoding biological models in
  XML
• Developed in collaboration with community at an
  annual forum
• Over 60 software applications claim to support
  SBML
• Primarily aimed at encoding biochemical reaction
  network models
• Why?
• Facilitate the exchange, sharing and reuse of
  models
• Facilitate the establishment of public model
  libraries
• Facilitate software tool interoperability
• Standard for publishing models

3
Overview of SBML

• Primary Domain biochemical network models
• XML with components that reflect the natural
  conceptual constructs used by modelers in the
  domain
• Reaction networks described by list of components
• Compartments
• well-stirred volumes
• Species
• a pool of biochemical entities of the same type
  located in a specific compartment
• Reactions
• any processes that change the amount of a species
• includes transport
• All parts are optional

4
SBML Other Features

• Functions
• allows definition of arbitrary functions for use
  in MathML expressions
• Parameters
• associate symbol with value
• constant or variable
• Rules
• mathematical constraints
• x f(Y), x f(X) or 0 f(X) types
• x can be species concentration, compartment
  volume or parameter
• X is set of all variables and parameters
• Y is set of all variables and parameters
  excluding x
• Metadata
• attach external information to components of
  model
• includes biological entity data e.g. GO term
• Events
• support for discrete event simulation
• mixed mode
• Units
• compound SI units can be assigned to components

5
SBML is Flexible

• SBML can be qualitative
• All the mathematical expressions and quantity
  Attributes are optional
• SBML can represent mathematical models without
  any reaction network
• Models can consist of just parameters, variables,
  differential equations and algebraic equations

6
Future SBML Development

• Proposals exist for
• Controlled vocabulary support
• Model composition
• Multicomponent species
• Diagrams
• Arrays and connection patterns
• Proposals requested for
• Spatial features
• Conditional objects
• Alternative mathematical representations of
  reactions
• e.g. Stochastic

7
Model Composition Proposals

• Proposals from
• Martin Ginkel, MPI Magdeburg
• Jonathan Webb, BBN
• Andrew Finney
• Common idea compose larger models from smaller
  ones
• Model contains
• Submodel definitions (or at least references to
  them)
• Instances of submodels
• Arbitrary links between objects inside instances
  and objects in enclosing model
• Links are directional and define attribute
  overload
• Direct links allow objects to refer directly to
  each other
• Issues
• Do we need interfaces?
• Should language define legality of links?
• Obviously can only link objects of same type
• Are XML standards, such as XLink, appropriate?
• Do we need to support arbitrary depth links?
• What are the semantics of links?

8
Multicomponent Species Proposal

• One of handful of proposals
• Hierarchical SpeciesType
• Biochemical entity type
• Independent of compartment
• Reaction generalized across compartments
• SpeciesType is graph of SpeciesTypeInstance nodes
• Arcs are pairs of binding sites
• State Generalized Reactions
• Products and Reactants are also graphs
• Graphs contain wildcards copied from products
  to reactants

9
SBML Development Process

• Not formalized
• SBML has been, and continues to be, developed as
  a collaboration amongst modellers and software
  developers
• The community interacts via
• SBML Forum meeting
• SBML Hackathon meeting
• sbml-discuss mailing list
• Mike Hucka and Andrew Finney
• Co-Chair SBML process and Co-Edit SBML documents
• SBML Team
• Based at U. Hertfordshire and Caltech
• Organizes meetings
• Provides services to support SBML standardization
• Creates proposals for extensions
• Not exclusive

10
SBML Meetings

• SBML Forum
• 2 day meeting
• Each year at International Conference on Systems
  Biology (ICSB)
• last one in October 2004
• attended by 50 delegates
• Heidelberg, Germany
• SBML Hackathon
• 2 day meeting
• April/May each year
• SBML community has been meeting biannually since
  April 2000

11
Software Developed by SBML Team

• libSBML Parser library
• Builds high level SBML object models inside
  applications from SBML Level 1 or Level 2 files
  or data streams
• Optionally validates the model against XML schema
  and for consistency as described by the SBML
  specification. (Units support under
  development.)
• APIs for C, C, Java, Perl, Lisp and Python all
  with the same underlying C implementation
• Same object model for both Level 1 and Level 2
• enables applications to seamlessly support both
  Levels
• Object model supports software modification of
  model structure and attributes
• Can write the object model out as Level 1 or
  Level 2
• Available for a range of operating systems
  including Linux, Windows, Mac OS X on both 32
  and 64 bit processors
• Comes with example applications that
• Validate an SBML file
• Translate a file from Level 1 to Level 2

12
Software Developed by SBML Team (continued)

• mathSBML
• Simulation and analysis environment for SBML
  models within Mathematica
• Support includes
• Events
• Algebraic Rules
• Complete MathML subset
• SBML Toolbox for MATLAB
• Simulation and analysis environment for SBML
  models within MATLAB
• Both allow users to access and modify SBML model
  structures

13
SBML Models Sets collated by SBML Team

• SBML Model Repository
• Manually transcribed from published models
• KEGG Database Conversion
• 10,000 models translated into Level 1 and Level 2
• Syntactic Test Suite
• Set of valid models for testing parsers
• Semantic Test Suite
• Set of models for testing simulators
• Each model comes with valid times series output
• Automatic test scripts are included
• Documented in SBML Wiki and available as a tar
  file
• Beta Release in use

14
Recommendations

• To Software Developers
• Software that supports the representation of
  biochemical reaction networks or pathways should
  support the import and export of SBML
• libSBML - which has bindings for C, C, Java,
  Python, Perl and Lisp - makes this relatively
  straightforward
• Software output should be valid SBML
• Can be checked using the online validation
  facility
• The interpretation of SBML by simulation software
  should be checked using the SBML Semantic Test
  Suite
• To Modelers
• Models should be made available in SBML
• Models should be submitted to open access
  databases of SBML models
• To Publishers
• Modelers should be encouraged to provide
  validated SBML when submitting articles
  containing models
• Readers should be able to download these models
  in the SBML format
• To the community
• Everyone is encouraged to contribute to the SBML
  development process to ensure SBML meets the
  needs of Systems Biology research

15
Conclusion

• SBML addresses the interoperability and model
  requirements of the systems biology modeling
  community
• SBML is a de facto open standard for representing
  biochemical network models
• Over 60 software applications support SBML
• SBML is being developed to incorporate the
  requirements of a large part of the community