Ontological design patterns: Metadata of molecular biological ontologies, information and knowledge

Abstract

Within the post-genomic era, scientists want to exploit the information of genome sequencing results, protein interactions, and gene deleted phenotypes. Molecular biologists build ontologies to model their knowledge in nonambiguous ways. Ontologies are metadata describing the meaning of data. They are useful for query answering in databases, hypotheses testing, and for the representation of context or explanation models. The provision of ontologies of complex information is challenging. Ontological Design Patterns (ODPs) make the ontological structure and content explicit. ODPs can be traced to reuse and adapt ontologies to objectives of information search, analysis and comparison. The three ODPs shown in this paper provide methods to hide unnecessary information (InteractionHider ODP), to notify interdependent concepts if one concept is changed (UpdateDependency ODP), and to handle multiple concepts, that can fulfil a specific context (ChainOfConcept ODP). All ODPs are available from the author with code examples.