Experiences Using a Meta-Data Based Integration Infrastructure.

摘要

One of the classic challenges in bio-informatics has been the integration of multiple community data sources into a single, consistent data repository. Over the years, there have been several attempts to address this problem. So far, none have been entirely successful. The task is a daunting one: integrate large amounts of complex data, obtained from autonomous, heterogeneous, distributed data sources which change their data representation regularly, and present it in a consistent and intuitive way. In this paper, we discuss our experiences using the meta-data infrastructure developed as part of the DataFoundry project at LLNL. First, we present the DataFoundry architecture, including both the meta-data representation and how it is used, in Section 2. Section 3 describes our experiences using the meta-data infrastructure to integrate two genomic data sources into our warehouse, outlines some of the problems initially encountered, and contrasts the process to the traditional, manual approach. Finally, we briefly summarize the current status and future directions of the project.