MyMED is a database system that mirrors MEDLINE頤ata and that provides access to the data locally, without the restrictions involved in accessing the National Library of Medicine Web site. The implementation of MyMED is based on the IBM DB2 Universal Database鬠the DB2ML Extender and the DB2 Text Information Extender. The resulting system supports advanced queries, including wildcard searching abilities, proximity searching, scoring functions, and capabilities to build a thesaurus, and constitutes an important research tool for text mining of biomedical-citation data. The paper describes the challenges encountered in the use of DB2 Extenders頰roducts, the methods used to overcome these challenges, and the implementation of MyMED. The benefits of MyMED are illustrated through some use cases, and directions for future work are outlined.