We have participated in a number of projects to convert data to RDF. We are also providing an operational knowledge base that has a SPARQL endpoint accessible by Internet.
The knowledge base incorporates information marshalled from
- Medline 2007 baseline (for MeSH headings for articles)
- 2007 MeSH (conversion to SKOS thanks to Mark van Assem et al.)
- Entrez Gene (for PubMed references)
- Rat genome database
- Gene Ontology (GO) (for category relationships)
- EBI GO Annotations (GOA): genes annotated to GO categories
- Mammalian phenotype (MP)
- Semantic Web Applications in Neuromedicine (SWAN)
- BrainPharm and NeuronDB
- Brain Architecture Management System (BAMS)
- Neurocommons text mining pilot
For tractability, we limited the scope to the organisms of greatest interest to health care and life sciences research: human, mouse, and rat.
SPARQL Endpoint Availability
Science Commons will provide a SPARQL endpoint for experimental public use during the one-month period 16 May to 10 June, 2007. The servers are available before 16 May, but we suspect some serious bugs and omissions that may lead to query failure. These are known and should be fixed before the 16th. Availability after 10 June will depend on whether we can obtain hardware for continued use.
The SPARQL endpoint may be found here.