A pattern-driven approach to biomedical ontology engineering

TL;DR

This paper presents a pattern-driven methodology for developing biomedical ontologies, demonstrating how reusable design patterns and programmatic tools can reduce complexity and improve efficiency in ontology engineering.

Contribution

It introduces a novel pattern-based approach using Tawny-OWL for creating and refactoring biomedical ontologies, enhancing automation and reusability.

Findings

Successful creation of a karyotype ontology using patterns

Refactoring of SIO ontology to incorporate patterns

Demonstrated reduction in development complexity

Abstract

Developing ontologies can be expensive, time-consuming, as well as difficult to develop and maintain. This is especially true for more expressive and/or larger ontologies. Some ontologies are, however, relatively repetitive, reusing design patterns; building these with both generic and bespoke patterns should reduce duplication and increase regularity which in turn should impact on the cost of development. Here we report on the usage of patterns applied to two biomedical ontologies: firstly a novel ontology for karyotypes which has been built ground-up using a pattern based approach; and, secondly, our initial refactoring of the SIO ontology to make explicit use of patterns at development time. To enable this, we use the Tawny-OWL library which enables full-programmatic development of ontologies. We show how this approach can generate large numbers of classes from much simpler data structures which is highly beneficial within biomedical ontology engineering.