Ontologies in Bioinformatics: Growing-up challenges
Janna Hastings1,2,3
1 Cheminformatics and Metabolism, European Bioinformatics Institute2 Swiss Center for Affective Sciences, University of Geneva
3 Swiss Institute of Bioinformatics
Robinson-Rechavi 2-day Group MeetingLausanne, June 4-5 2012
STANDARDS:RESEARCH
REPRODUCABILITY,BIOCURATION,
DATA INTEGRATION,EXCHANGE,
SEMANTIC WEB…
ANALYSIS: ENRICHMENT,
SEMANTIC SIMILARITY,…
DATABASE MANAGEMENT:
BROWSING,VISUALIZATION,CLASSIFICATION,
AUTOMATED REASONING,
…
Applications of ontologies in bioinformatics
Growing-up challenges for ontologies in bioinformatics
1. Visualization2. Moving from OBO to OWL –
tools, software3. Moving from single ontologies to multiple
ontologies (using cross-products)4. Training aimed at biologists and
bioinformaticians5. Community building
Selected ontology-related publications (JH @EBI)
• Janna Hastings et al., Structure-based classification and ontology in chemistry. Journal of Cheminformatics. 2012 Apr 5; 4:8.
• Barry Hardy, [...] Janna Hastings et al. Toxicology ontology perspectives. ALTEX 2012, 29 (2), 139.
• Janna Hastings et al. The Chemical Information Ontology: provenance and disambiguation for chemical data on the biological semantic web. PLoS ONE, 6(10).
• Melanie Courtot, […] Janna Hastings et al. Controlled vocabularies and semantics in Systems Biology. Molecular Systems Biology, 7:543.
• Janna Hastings et al., Modularization requirements in bio-ontologies: A case study of ChEBI. Frontiers in Artificial Intelligence and Applications Volume 230, p. 63--70.
• Kirill Degtyarenko, […] Janna Hastings et al., ChEBI: a database and ontology for chemical entities of biological interest. Nucleic Acids Res. 36, D344-D350.
ChEBI: Chemicals of Biological Interest
1. Ontology visualization
There is a need to hide irrelevant information and concisely display relevant information
ChEBI has introduced a new heuristic for hiding redundant paths in the visualiation
See also http://ols.wordvis.com/
2. Moving from OBO to OWL
DEPRECATED
OBO-Edit is the most commonlyused bio-ontology editor, but it isno longer actively being developed
UNUSABLE
by biologist
s?
Protégé is the tool that will replace OBO-Edit, but it urgently needsinvestment in usability improvements for use by biologists and curators
APIs for programmaticaly working with OWL ontologies
• OWLTools (Java) … wrapper for, and much easier to use than, the OWL API (Java)
• OntoCat (R)• POSH (Prolog, Unix shell)• InfixOWL (Python)
Normalization of ontologies+ automated reasoning for hierarchies
FBbt
Normalized: More relationshipsMore inferred classificationsEasier to maintain but harder to navigate
Non-Normalized: Only is-a, maybe part-of relationsAll relevant relations are not asserted, onlythose found salient by the human annotators
3. Working with multiple ontologies
• Cross-products, e.g. chemicals in GO (for terms such as ‘carbohydrate metabolism’)
• No online visualization tools yet are able to display cross-products across ontologies
• Most databases don’t support annotation with composite terms, necessitating pre-composition
4. Training
5. Community building• ICBO conference• ISMB Bio-ontologies SIG
Not well attended by biologists/curators?
• Bio-curation conferenceNot well attended by ontologists?
• OBO Foundry Not very active lately?
• OntoSIB
Ontology @Pub nights Blog: www.bioontology.ch
COMPUTER SCIENCE
BIOLOGY
Thank you, AcknowledgementsChristoph SteinbeckColin BatchelorMarcus EnnisVenkatesh Muthukrishnan
Jane LomaxChris Mungall
BBSRCEU-OPENSCREEN
Barry SmithKevin MulliganWerner CeustersNicolas le NovereDavid Osumi-Sutherland
FNS (SNF)EMBL