Elsevier position paper - W3C workshop on Semantic Web for Life Sciences 
Anita deWaard (a.dewaard@elsevier.com) 
Ellen Hays (e.hays@elsevier.com)


Elsevier is naturally focused on improving the quality and usefulness of the services we offer our customers. In support of that aim we are engaged in a number of efforts involving both the technical mechanisms and the content of the Semantic Web. Here are a few:

1) Data - (XML) - Virtually all (and soon to be all) of our content is now in XML. Elsevier has a data standards group that tracks W3C standards, OASIS, etc., for application in our business. 

2) Web services - We have developed a web services architecture, which is at the core of our electronic offerings, and are helping the MedBiquitous Consortium (http://www.medbiq.org) to define a similar architecture, exploiting our experience with web services for making medical journals available online. 

3) Data integration - Ontologies: Elsevier is doing experiments using RDF and open standards for representing and using thesauri/ontologies. For this work we are especially interested in RDF Schema and OWL. 

We have done one project using RDF(S): DOPE (Drug Ontology Project for Elsevier), which uses EMTREE encoded in RDF as its thesaurus. In this implementation, data sources had RDF wrappers, and we used Sesame (from Aduna) to tie it all together. A paper on DOPE was submitted to Semantic Web Challenge (2003), and was published in the IEEE special issue on the Semantic Web Challenge. 

The DOPE project made use of one thesaurus and multiple data sources. Future efforts will focus on expanding our experience with thesaurus mapping, using, for example, RDF(S) and some variant of OWL, to integrate multiple thesauri and any number of data sources. 

What we would like from the standards community: 
* stable standards;
* a scalable architecture for searching large volumes of data.

We are also quite interested in initiatives such as LSIDs (life science IDs) and Gene Ontology.