Integration of disease entries across OMIM, orphanet, and a proprietary knowledge base

Abstract

Integration of disease databases benefits physicians searching for disease information. However, current algorithmic matching is not sufficiently powerful to automate the integration process. This paper reports our attempt to manually integrate disease entries spread across public disease databases, Online Mendelian Inheritance in Man and Orphanet, with a proprietary disease knowledge base. During the process, we identified that relations between synonyms require special handling, and a set of resolution rules are proposed. Situations encountered throughout the integration suggested that variations in the cross-references would facilitate future integration of distinct disease databases.