210 likes | 266 Views
libAnnotationSBML provides a library for linking annotated SBML models to MIRIAM and web services. It helps with writing and reading annotations in Java, aiming to enable automation of tedious tasks and enhance model annotations. An AnnotationWizard tool prompts users to confirm annotations from ChEBI, UniProt, and others. This tool supports ontology mapping, ontology term retrieval, and model corrections, contributing to improved model quality. The wishlist includes C++ support, dynamic database linking, and better web service consistency for enhanced usability.
E N D
libAnnotationSBML Neil Swainston Manchester Centre for Integrative Systems Biology 29 March 2009
Introduction • Library (not a tool) for linking annotated SBML models to MIRIAM and related web services • Writing and reading of annotations • Java – could be extended
Purpose of annotations • Provide definitive, unique identifier for “stuff” • One motivation: mapping experimental data to models • Need annotated models AND annotated data
AnnotationWizard • Shameless copy of Falko’s semanticSBML • See also SAINT (Allyson) • Principle: • Iterate through model • Submit each species to an appropriate web service search facility (ChEBI or UniProt) • Prompt user for confirmation • Update model with annotation • Has also been wrapped as a web service itself
Annotating models: yeast • Genome-scale SBML model of yeast metabolism • All 2153 molecules have MIRIAM annotations
Project overview Enzyme kinetics Quantitative proteomics Quantitative metabolomics PRIDE MeMo MeMo-RK SABIO-RK Web service Web service Web service Variables (metabolite, protein concentrations) Parameters (KM, Kcat) SBML Model
Project overview Enzyme kinetics Quantitative proteomics Quantitative metabolomics PRIDE MeMo MeMo-RK SABIO-RK Web service Web service Web service Variables (metabolite, protein concentrations) Parameters (KM, Kcat) SBML Model
Purpose of annotations • Provide definitive, unique identifier for “stuff” • One motivation: mapping experimental data to models • Need annotated models AND annotated data • “Tagging” • More interesting purpose… • Use the annotations to reason over the model
Purpose of annotations • Easy question • Is my model annotated? • Simple answer: yes • Can be answered solely by libSBML • More interesting question • Is my model annotated well? • Do my reactions elementally balance? • Simple answer may be yes or no, but harder to determine • Can’t be determined solely by libSBML
Isn’t ChEBI grrrreat… <species id=”glc" name="D-Glucose"> <annotation> <rdf:li rdf:resource="urn:miriam:obo.chebi:CHEBI:17634"/> </annotation> </species>
libAnnotationSBML: what it does • Provides unified interface to following web services: • ChEBI • KEGG • UniProt • SBO • others… • Wraps these calls so that it appears that you are accessing these locally
MIRIAM KEGG ChEBI UniProt Determine ontology Map to web service libAnnotationSBML OntologyTerm(s) SBase annotation
OntologyTerm • Model components can be parsed to return OntologyTerms • OntologyTerms have the following interface: • Development of tools to exploit annotations becomes greatly simplified: • OntologyTerm[] ots = sbmlUtils.getOntologyTerms( sbase ); • OntologyTerm ot = ots[0]; • ot.getName(); ot.getSynonyms(); ot.getFormula();
Purpose of annotations: to allow automation of tedious jobs D-Glyceraldehyde + NAD+ <=> D-Glycerate + NADH + H+
Purpose of annotations: to allow automation of tedious jobs • Suggest corrections to existing models • Incorrect annotations • Missing reactants / products • Stoichiometry • Better still: intelligent generation of models? • AutoComplete?
Wish list • Support more web services • Write it in C++ • WSDLs in MIRIAM??
Wish list • BUT dynamic linking of models to databases is still a little way off • Web services are inconsistent • getName()? getDescription()? getTitle()? • Semantic web services needed
Finally… • Hopefully makes doing things with annotations easier • Have a look if you’re interested: • http://mcisb.sourceforge.net
libAnnotationSBML Neil Swainston Manchester Centre for Integrative Systems Biology 29 March 2009