W3C's Semantic Web initiative holds promise for life science's data-integration challenges

By Salvatore Salamone

December 15, 2004 | The World Wide Web Consortium (W3C) has targeted the life sciences as an early and ideal application area for its nascent Semantic Web technology. "The challenges posed by drug discovery can be solved only if we can integrate data across the many fields of life sciences," says Tim Berners-Lee, founder and director of the W3C.

Many researchers agree, and last month met to discuss how best to proceed at a W3C Workshop on Semantic Web for the Life Sciences, held in Cambridge, Mass. Proponents hope that the Semantic Web will unsnarl key data-sharing traffic jams. Consider these extracts from a position paper discussed at the meeting:

• "How can experimental protocols, descriptions of model systems, statistical criteria for data acceptability, and many other critical elements be effectively communicated between technology silos?" asked Ted Slater, associate research fellow at Pfizer, whose comments were part of the paper.
  
  
• "The second issue is that of synthesizing results from the various technology silos into a holistic picture of physiology," Slater said. "What does it mean when RNA profiling results don't correlate with proteomics results? How can we talk about genes, their various RNAs, proteins, and their post-translational modifications, small molecules, and the myriad processes that involve all of these things simultaneously, succinctly, and without semantic errors?"

Tags Defined 

Tim Berners-Lee

Solving the kind of data-integration challenges described by Berners-Lee and Slater is precisely the idea behind the Semantic Web. The proposed standard would add "defining tags" to data within Web pages, provide links to that data, and enable applications to find the data and make associations between different data elements.

Success or failure, of course, will depend on how widely the Semantic Web is adopted and incorporated into applications and databases. The early interest seems high. Workshop participants included representatives from AstraZeneca, Aventis Pharmaceuticals, Elsevier, HP, IBM, Jackson Laboratories, the National Cancer Institute Center for Bioinformatics, Oracle Life Sciences, the Swiss Institute of Bioinformatics/UniProt, and the University of Michigan School of Medicine.

Think of the Semantic Web technology as a way to create a globally distributed database that can be used easily.

An often-cited example of Semantic Web technology relates to a prospective conference attendee. This person typically seeks the conference's Web page, and looks up details such as date, location, and list of speakers. He or she might open a calendar application, cut and paste session dates and times into the calendar, and perhaps cut and paste the event location into a mapping program to obtain directions.

Standards Rule Semantic Web

The Semantic Web defines a new language for computers to exchange data. At the heart of the language are several Web technologies that make it easier for applications to share information. Read More

That's a lot of manual work, which would be eliminated by Semantic Web tagging, its supporters say. However, for this approach to succeed, posted information (data) must be tagged or identified as a date, location, or speaker, in the case of a conference. Applications that collect and use such information must be able to recognize the tags and associated data. Life science data would have its own set tags.

To facilitate data exchange, the W3C has developed several standards such as the OWL Web Ontology Language and the Resource Description Framework (RDF). W3C leverages XML descriptions and Life Sciences Identifier (LSID) data for tagging.

Semantic Web technologies are also being incorporated into existing products. IBM, HP, Adobe, and others now have toolkits to help create Semantic Web applications. Support is growing in the open-source community as well.




PHOTO BY EPA PHOTOS