A Basic Formal Ontology-Based Ontological Modeling for Plan and Occurrence, a Biomanufacturing Process Verification Use Case

Abstract Biomanufacturing has gained significant importance in recent years due to its role in developing new medications, handling pandemics, and increasing the well-being of human populations. The nature of biochemical processes requires complex planning and control, with many controlled and non-controllable variables that impact the quality of bioproducts. Representing biomanufacturing process knowledge, control models, and actual occurrences in coherent ontologies could aid both humans and computers in dealing with the complexity. However, there is a lack of such coherent ontologies. Even though the Industrial Ontology Foundry (IOF) Core ontology has provided a groundwork based on the widely used Basic Formal Ontology (BFO) for such ontological requirements, there are still insufficient constructs and clear guidance on the representation of digital artifacts and their correspondences to the physical counterparts. This paper presents a framework to extend the IOF Core to address the gap. The framework is founded on establishing a counterpart (CR) relation pattern presented in our previous paper. Counterpart relation was selected for its ability to facilitate a more intuitive and concise representation of many kinds of digital artifacts (e.g., planned, designed) and physical entities (e.g., planning process, manufacturing process). We validated the approach with a process verification of a fed-batch bioreactor operation. The paper started by defining the use case requirement, which was followed by an ontology development. A knowledge graph of the bioprocess plan and occurrences of processes in the plan was then instantiated. Competency questions were used to concretize the ontology requirement from the use case, and subsequently, an executable set of queries was created from them and was used to computationally validate the ontology against the requirement. The GraphDB tool was used to support the validation. The result of this research not only showed that the CR pattern described in our previous paper could satisfy the requirements related to the digital thread of digital and physical process information, but it also demonstrated that several visualization approaches on graph data can be used to address competency questions. These findings provide insights into the future of data integration and management within biomanufacturing, highlighting the role of ontologies for improved data interoperability and analysis.