It started as the seed of an idea at ISMB in Copenhagen in 2001, and sprouted a bit at ISMB 2002 in Edmonton. But BioPAX, an initiative to create a data exchange standard for pathways databases, didn’t fully flower until a few weeks ago, appropriately enough in the fecund atmosphere of PSB 2003, held on the Big Island of Hawaii.
BioPAX, short for “biopathways exchange,” is led by Chris Sander at the Memorial Sloan-Kettering Cancer Center in coordination with the BioPathways Consortium. A core group of around eight people from MSKCC, the University of Colorado Health Sciences Center, SRI’s Bioinformatics Research Group, Argonne National Laboratory, Harvard Medical School, and the Samuel Lunenfeld Research Institute have already laid out the groundwork for the effort, as well as an initial ontology, which they detailed in a presentation during the biopathways session at PSB.
The format is being designed to combine the strengths of existing pathway databases such as BioCyc, BIND, WIT, and KEGG, with the goal of improving data exchange and integration between these resources. The group hopes to include support for all types of biological pathway data, including metabolic pathways, signaling pathways, protein-protein interactions, and genetic regulatory pathways.
The initial focus of the group was the exchange of chemical compound information, so the group evaluated the Chemical Markup Language (www.xml-cml.org) as one possible model for the BioPAX format. However BioPAX also plans to be compatible with a format for proteomics data exchange that the Proteomics Standards Initiative is developing (http://psidev.sf.net). An initial implementation of BioPAX will soon be available as an XML schema document, according to the group, and BioPAX is also evaluating DAML-OIL and OWL as possible description languages.
BioPAX workgroup member Joanne Luciano of Harvard Medical School told BioInform that the next short-term goal for the group would be the development of software that exports EcoCyc to CML format. “Since we’re starting with small molecules, this seems to be the first logical step,” wrote Luciano in an e-mail. BioPAX will develop the specification and related documents for the format in parallel, she added.
The BioPAX organizers hope to attract interest from the broader bioinformatics community in carrying out the next stages of the initiative. Further information is available at the group’s website: www.biopax.org.