GenXref. VI: Automatic generation of links between two heterogeneous databases.

Abstract

A large proportion of the information found in public databases is not sufficiently cross-referenced. We developed genXref, an automated system for link inference, because embarking on a manual cross-referencing of genome data would require too much expensive human expertise. It uses information retrieval technology to generate links between objects of heterogeneous databases. GenXref was used to generate links between GDB genes and Genbank human sequences. It resulted in > 10,000 links with a precision of 83% and a recall of approximately 32%.