A novel web portal for large biological ontology

Abstract

As a formal, explicit specification of a domain of interest, ontologies consist of precisely defined terms and the relationships between them, thus imparting a hierarchical organization. Ontologies are increasingly being used to define the basic terms and relations in biological domains, often as the foundation for search, integration and exchange of biological data. How to present ontological information and make it easy for users to search and navigate has been a challenge for large and complex ontologies. In this thesis, a novel web portal for large biological ontology was developed using the Protein Ontology (PRO) as an example. ☐ Protein Ontology (PRO) provides an ontological representation of protein-related entities and relationships between them. Previously, the heterogeneous data of PRO was stored, accessed, and presented using a relational database as the backend. Recently, PRO data has been converted into RDF triples and integrated into a Virtuoso RDF graph database that provides the foundation for building a unified and scalable data retrieval solution. For example, hierarchical data can be efficiently retrieved from the RDF graph database. ☐ Biological ontology often consists of a set of concept names or terms, the definitions of terms, synonyms and the relationship between terms. The relationship between each ontology terms are usually hierarchically organized. Our study of top 20 BioPortal ontologies indicates that users could hardly browse the information associated with an ontological term and the hierarchy the term is organized in at same time, especially for large and complex ontologies. Paradoxically, large ontologies often have abundant contents for a term, as well as extraordinary large hierarchy. Some ontologies separate hierarchy and term information into different web pages, enable users to view the contents and the whole hierarchy independently. Thus, how to present integrated view in a user-friendly way is challenge for large biological ontologies. ☐ The contribution of this thesis is to address the problem of how to make better representation of ontological information for large biological ontology. This work focuses on the development of a novel web portal for PRO using RESTful APIs, SPARQL query and Virtuoso triple store as the backend. The entry page of the new web portal for PRO is designed with three panels that enable user to browse hierarchy and term information simultaneously, therefore providing the users with both global and local views at the same time. In addition, we also give users freedom to control what contents to show or hide on the webpage. As a result, this thesis work uses PRO to demonstrate the new web portal and proves the design could improve user experiences.