DIO: Drug interaction ontology - Application to inferences in possible drug-drug interactions

We employed ontology for encoding basic information regarding drug-protein (or endogenous substances) interactions. Those are developed as Drug Interaction Ontology (DIO), by which primitive knowledge components are encoded as symbolic representations: 1) interaction orientation, 2) interaction modality, and 3) interaction effect. We also schematized consequences of theoretical combinations between two sets of drug-protein interactions in the form of symbolic representations. Those can be applied for identification and reasoning possible drug-drug interactions. Although actual drug-drug interaction events in individual body is complex and need various parameters for sophisticated predictions, this ontology is designed to reflect basic cognitive process. The knowledgebase using this ontology implemented by relational model is under development, starting drugs in cancer therapy. Example of identification of possible drug-drug interactions in case of irinotecan is presented.