Ontologies became increasingly important in recent years, especially in life sciences. Typically, they consist of a harmonized vocabulary of terms, the so called concepts, describing and structuring a domain of interest. A prominent example in life sciences is Gene Ontology, a heavily used ontology source providing sub ontologies for molecular functions, biological processes and cellular components. Other life science ontologies are collected and made centrally available by the OBO (Open Biomedical Ontologies) Foundry. In life sciences, the biological objects, such as genes and proteins, are typically associated (“annotated”) with ontology concepts to consistently and semantically describe their properties, e.g., molecular functions and biological processes where proteins are involved in.

Usually, life science ontologies are explicitly modeled by ontology developers and scientists. Hence, these ontologies are mainly influenced by a specific community agreement (at least among the ontology developers) on the one hand and a specific state of domain knowledge an ontology represents, on the other hand. Therefore, ontologies evolve whenever they are adapted to implement new or changed community agreements, and new research results or insights influencing the covered domain knowledge, respectively. Since ontology changes can lead to outdated annotations of biological objects, particularly when ontology concepts have been deleted or became obsolete, it is of interest how stable ontologies are and how many and what type of changes have been made.

The analysis covers an interval of 45 months (May 2004, Feb 2008), i.e., we consider only ontology versions within this interval and at most one ontology version per month. For more than one ontology version per month, we use the first available version for our analysis.