The DARPA Agent Markup Language (DAML) is not only gaining support from the World Wide Web Consortium as it promotes its vision of the “Semantic Web” — an extension of the current Web that will use software agents to add a new level of machine-readable meaning to Web pages — but it is also gaining a foothold in bioinformatics.
The Bio-Ontologies Consortium recently voted to recommend the latest DAML release, which includes the Ontology Inference Layer (OIL), as its language of choice for the exchange of bio-ontologies.
The BOC’s exchange language working group considered both DAML+OIL and XOL (Ontology Exchange Language) in its search for a language that would provide frame-based semantics and an XML-based syntax. While XML is an improvement over HTML because it allows information to be described unambiguously using tags, it has limited ability to describe ontologies and other relationships.
The representational capabilities of DAML+OIL and XOL were determined to be equivalent, so the BOC opted for DAML+OIL largely due to its support by the W3C and the Semantic Web community, including such heavyweights as Tim Berners-Lee.
“Our conclusion is that DAML+OIL is a strong technical candidate, but also is the most viable language with respect to adoption and tool development,” wrote BOC committee member Robin McEntire in his endorsement of the language prior to its adoption by the consortium.
Jim Hendler, a DARPA researcher who heads up the DAML initiative, told BioInform that the language offers several advantages for bioinformatics. “More precise searches will help in finding relevant papers and web sites, intelligent database access is crucial in genomics and other complex biological data sources, and many of the devices used for capturing biological information could be made available to others over the web,” Hendler said.
Hendler added that the DAML approach would allow more precision in the description of entities on the web. “Thus, where XML would allow a user to say that a certain part of a paper is about a particular chemical, DAML allows the user to express more precise relationships. This would let a user distinguish between articles that describe a particular use of a chemical in some experiment, that describe how to purchase the chemical, that explain some use of the chemical as a catalyst in some experimental system, and so forth.”
McEntire said that some tools already exist that allow developers to build ontologies in DAML+OIL “and it is our expectation that more tools will be developed in the near term.”
Hendler said that DAML has seen strong support from the IT community. In addition to the W3C’s involvement, a joint US/European Union committee on agent markup languages has been formed. He said that a number of prototype applications for DAML have already been developed.
DARPA is interested in working with the bioinformatics community, Hendler noted, “due to the [Department of Defense’s] critical interest in this area.”