Achieving interoperability of genome databases through intelligent Web mediators

Resolving the tension between the flexibilities offered by a database query language and the so-called "user friendliness" supported by a form-based or menu-driven database query interface is undoubtedly a daunting task. However, researchers believe that the flexibilities offered by a declarative query language should be fully exploited to harvest genetic solutions within the wealth of information offered by genomics and proteomics. Additionally, the need to query databases at different abstraction levels, in a set-based fashion, and to process the queries efficiently by resolving the structural and semantic disparities among the genomic databases in a user-transparent way has never been more evident than in recent years. Inspired by such needs, we have developed an SQL3-compliant database query language called the Genomic Query Language (GQL) for manipulating globally distributed genomic databases that are publicly accessible. GQL, an SQL-like language, is given a source-independent syntax and exploits a high-level global view of the underlying databases. The uniqueness of this language relies heavily on the supporting interface through which all communications take place and the interoperability of databases behind firewalls is realized. In this paper, we describe how the Web-based interface we have built for GQL makes it possible to understand the scheme of any participating database, assimilates the syntactic and semantic information into the global view, and responds to user queries. The interface is autonomous for many genomic databases and is also fully self-maintainable.