German English

Determining and Analyzing Semantic Ontology Changes with CODEX


Google Scholar

publication iconHartung, M.; Groß, A.; Rahm, E.
Determining and Analyzing Semantic Ontology Changes with CODEX
Demo @ Intl. Conference on Data Integration in the Life Sciences (DILS)


Life science ontologies continuously evolve to incorporate the latest requirements and insights of a domain. A key task to support ontology evolution is to find out what has been changed between an old and an updated ontology version (diff). Such a diff should not be limited to simple changes like concept additions and deletions but also cover more complex changes such as merging, splitting and moving of concepts or the addition of entire sub-ontologies. The typically large complexity and size of life science ontologies make such a semantic diff highly desirable as it can be relatively compact and well understandable. Such a diff can be valuable for both ontology designers and users. For instance, a developer team may be interested in a compact representation of the latest changes to better plan further modifications of their ontology. A semantic diff can help users to explore whether or not the ontology changes affect their analysis studies and results, e.g., for term enrichment.

We will demonstrate our new web tool CODEX (Complex Ontology Diff Explorer) to determine semantic changes between ontology versions. It is available at The ontology versions in OWL or OBO formats can either be uploaded or directly provided by users. Alternatively, one can use ontology versions available in the OnEX repository. The computation of the semantic diff is performed by our COntoDiff algorithm that involves an initial matching of the ontology versions. Users can explore the results in multiple ways. First, one can study the overall statistics of the diff via tables and pie charts which quantify the number of found changes. Second, to explore the diff in more detail CODEX offers a change explorer and navigator. On the basis of tag clouds one can identify the most frequent changes or elements that have been influenced by a change. For one complex change one can drill down into its details, e.g., exploring its associated simple changes. Finally, users have the option to export results for further processing.

In our demonstration we will show how to use CODEX to determine semantic diffs for ontologies such as the Sequence Ontology and the Gene Ontology. We will demonstrate the different input options and explain the various exploration/analysis possibilities.