The use of Foundational Ontologies in Bioinformatics

Abstract Background: The FAIR principles recommend the use of controlled vocabularies, such as ontologies, to define data and metadata concepts. Ontologies are currently modelled following different approaches, sometimes describing conflicting definitions of the same concepts, which can affect interoperability. To cope with that, prior literature suggests organising ontologies in levels, where domain specific (low-level) ontologies are grounded in domain independent high-level ontologies (i.e., foundational ontologies). In this level-based organisation, foundational ontologies work as translators of intended meaning, thus improving interoperability. Despite their considerable acceptance in bioinformatics, there are very few studies testing foundational ontologies. This paper describes a systematic literature mapping that was conducted to understand how foundational ontologies are used in bioinformatics and to find empirical evidence supporting their claimed (dis)advantages. Results: From a set of 79 selected papers, we identified that foundational ontologies are used for several purposes in bioinformatics: ontology construction, repair, mapping, and ontology-based (federated) data analysis. Foundational ontologies are claimed to improve interoperability, enhance reasoning, speed up ontology development and facilitate maintainability. The complexity of using foundational ontologies is the most commonly cited downside. Despite being used for several purposes, there were hardly any experiments (1 paper) testing the claims for or against the use of foundational ontologies. In the subset of 49 papers that describe the development of an ontology, it was observed a low adherence to ontology construction (16 papers) and ontology evaluation formal methods (4 papers). Conclusion: Our findings have two main implications. First, the lack of empirical evidence about the use of foundational ontologies indicates a need for evaluating the use of such artefacts in bioinformatics. Second, the low adherence to formal methods illustrates how the bioinformatics field could benefit from a more systematic approach when dealing with the development and evaluation of ontologies. The understanding of how foundational ontologies are used in bioinformatics can drive future research towards the improvement of ontologies and, consequently, data FAIRness. The adoption of formal methods can impact the quality and sustainability of ontologies, and reusing these methods from other fields is encouraged.