Modeling gene expression sample variables with an application ontology


ABSTRACT:  Describing biological sample variables with ontologies is complex due to the cross-domain nature of experiments. Ontologies help in providing annotation solutions, however, for cross-domain investigations, multiple ontologies are needed to represent the experimental data. Such ontologies are subject to rapid change, are often very large and are not interoperable and present complexities that are a barrier to users of biological resources. In this talk I will present our approach to curating gene expression experimental data with an application ontology, the Experimental Factor Ontology (EFO), designed to meet cross-domain, application focused use cases. I will also describe the additional querying power such an ontology now gives us in our applications, such as the Gene Expression Atlas and ArrayExpress at EBI. Our approach demonstrates a successful methodology for consuming and applying domain ontologies in a real-world data repository application.