With growing needs of microarray data sharing, there are efforts for the development of microarray standards. The standard data exchange model, MAGE-OM (Microarray Gene Expression Object Model) is an object-oriented conceptual model for microarray expression data. MAGE-OM database system is applicable for storage of the associated XML data exchange format MAGE-ML (Microarray Gene Expression Markup Language) and for higher level analysis and integration with biomedical resources. We have implemented MAGE-OM in both frame-based ontology and relational database to exploit the great modeling power of MAGE-OM and compared them in terms of consistency, efficiency and flexibility to the data model. Two implementations showed considerable difference in representing relationships among classes. The ontology in the frame-based system nearly matched the object-oriented model, but performance may become problematic as the database grows. The relational database schema was preferable for performance but it is difficult to guarantee the consistency to the conceptual object level. Our relational schema is also shown to be simplified and provide improved efficiency in comparison with recently published database Array Express at the European Bioinformatics Institute. These design approaches would be helpful to understand the suitability and limitations of each implementation in the context of building standard-compliant database for microarray.