Drug interaction ontology (DIO) for inferences of possible drug-drug interactions

Stud Health Technol Inform. 2004;107(Pt 1):454-8.


Drug Interaction Ontology (DIO) was developed for formal representation of pharmacological knowledge. It provides a fundamental framework for accumulation of reusable knowledge components in molecular pharmacology. Ontology was employed and implemented as a relational model. Some features include: 1) Drug-biomolecule interaction was assumed as a primitive knowledge element. 2) Symbolic representation was developed for drug-biomolecule interaction. Consequences of two conjugated units of interaction were defined by using symbols. These are applied for query development for identification of possible drug-drug interaction. 3) The triadic relationship model was developed as a ground model for bio-logical interactions and/or function, including semantic ones. One application of DIO is to support hypothesis generation of drug interaction by providing new hypotheses from a structured database storing literature information on known drug-biomolecule interactions. A knowledge base using DIO that contains information beginning with anti-cancer drugs is now under development. Detection of possible drug interaction was tested and its capacity to lead clinically known ones was confirmed. The system generated theoretically possible drug-drug interactions, which implies potential usefulness of new drugs to be tested before actual clinical application. In this paper, sorivudine and 5-fluorouracil mediated by dihydropyrimidine dehydrogenase are presented.

MeSH terms

  • Arabinofuranosyluracil / analogs & derivatives*
  • Arabinofuranosyluracil / pharmacology
  • Artificial Intelligence
  • Computational Biology
  • Dihydrouracil Dehydrogenase (NADP) / pharmacology
  • Drug Interactions*
  • Fluorouracil / pharmacology
  • Humans
  • Vocabulary, Controlled*


  • Arabinofuranosyluracil
  • sorivudine
  • Dihydrouracil Dehydrogenase (NADP)
  • Fluorouracil