Biological databases typically tag their records with unique identifiers called accession numbers. These locators simplify record retrieval, but they are also database-specific, presenting a headache for bioinformaticians who want to map records in one database with their equivalents in another. Typically this is done manually, a tedious, error-prone task, especially when applied genome-wide.
A few years ago, Robert Kincaid, a senior research scientist at Agilent Laboratories, Palo Alto, Calif., came up with another approach. Inspired by Internet domain naming services that map, for instance,
Kincaid's prototype derives its lookup information from NCBI's LocusLink database. So, for example, it can resolve gene aliases into their preferred names or GenBank accession numbers to a UniGene cluster. But BNS is more generic than that and can map other types of data, Kincaid says in an E-mail.