Collaborative Development of Biomedical Ontologies and Terminologies

Stanford University, Stanford CA

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): The construction of ontologies that define the entities in an application area and the relationships among them has become essential for modern work in biomedicine. Ontologies help both humans and computers to manage burgeoning numbers of data. The need to annotate, retrieve, and integrate high-throughput data sets, to process natural language, and to build systems for decision support has set many communities of biomedical investigators to work building large ontologies. We developed and evaluated the Collaborative Prot¿g¿ system in the first phase of our research project. This software system has become an indispensable open-source resource for an international community of scientists who develop ontologies in a cooperative, distributed manner. In this competing renewal proposal, we describe novel data-driven methods and tools that promise to make collaborative ontology design both more streamlined and more principled. Our goal is to create a more empirical basis for ontology engineering, and to develop methods whereby the ontology-engineering enterprise both can profit from data regarding the underlying processes and those processes in turn can generate increasing amounts of data to inform future ontology-engineering activities. Our research plan entails three specific aims. First, we will enable ontology developers to apply ontology-design patterns (ODPs) to their ontologies, and we will measure the way in which these patterns alter the ontology-engineering process. Second, we will analyze the vast amounts of log data that we collect from users of Collaborative Prot¿g¿ to understand the patterns of ontology development. We will use these patterns to recommend to developers areas of ontologies that may need their attention, facilitating the process of reaching consensus and making collaborative ontology engineering more efficient. Finally, we will use the extensive data collected by our group and others to understand how scientists reuse terms from various ontologies and we will use these emerging patterns to facilitate term reuse. Each of these analyses not only will increase our understanding of collaboration in scientific modeling, but also will lead to new technology within our Collaborative Prot¿g¿ suite that will improve the ontology-development process and make collaboration among biomedical scientists more efficient.