A recent “substantial” gift from Incyte Genomics will help Stanford’s Gene Ontology project expand its ontologies and make them more useful, said Stanford’s Michael Cherry, who heads the project.
Cherry declined to disclose the amount of the gift, which is the first step in a collaboration between Incyte and the Gene Ontology Consortium. Cherry said that Incyte’s expertise in annotating its own databases would be a significant factor in the collaboration.
Stanford is a founding member of the GO Consortium, which aims to create a common vocabulary to share information on the functional analysis of genes. The consortium also includes several academic, governmental, and commercial groups.
Other commercial participants include Proteome (which joined prior to its acquisition by Incyte), Celera, and AstraZeneca, which supplied the initial funding to launch the project in 1999, Cherry said.
Celera and Proteome have not contributed to the funding of GO, but are including Gene Ontology terms within some of their bioinformatics products, Cherry said. He noted that commercial users get the information for free, but they have an obligation to provide useful feedback on structure and form.
Cherry said that the project also anticipates a grant from the NIH’s National Human Genome Research Institute in the near future.
Incyte will also grant Stanford a royalty-free license for the GO Consortium to use Incyte’s proprietary Protein Functional Hierarchy, a vocabulary for the functional categorization of genes that the company has been using for annotation.
Cherry said that while the Protein Functional Hierarchy works on a different level from that of GO, “We’ll be able to link their information, not necessarily one-to-one, into the ontology.” He said the group would also benefit from studying how Incyte addressed its own ontological challenges.
Cherry described the gene ontology as a “network of biological terms that will act as a control vocabulary” for gene function.
He stressed the importance of network structure of the gene ontology, as opposed to a hierarchical structure. “Most functional description vocabularies that have been created until now have a strict hierarchical structure. And within biology, because the cell likes to use various proteins for various things, we’ve found that it really must have the flexibility of associating things differently,” he said.
The GO Consortium came together after the first Bio-Ontology Consortium meeting, which followed the Intelligent Systems for Molecular Biology conference in 1998. Three database groups formed the core of the consortium: the FlyBase database group from the Drosophila project, the mouse genome informatics group at the Jackson Laboratory, and Stanford’s Saccharomyces Genome Database.
The GO Consortium now meets every three to four months, Cherry said, and has expanded to include the Arabidopsis Information Resource database group, the Caenorhabditis elegans WormBase group, and several companies.
While some of the members of the GO Consortium are also involved with the Bio-Ontologies Consortium, Cherry said the two projects do not have a formal relationship.
However, Robin McEntire, principal computing scientist with GlaxoSmithKline and a member of the Bio-Ontologies Consortium, said there is a “strong connection” between the two groups.
McEntire said that the Bio-Ontologies Consortium has served as a driving force for efforts such as GO that are building ontologies within particular domains. The group facilitates use of developed ontologies by driving the commercial market toward their use, he said.
Formed with the original intent to build life science ontologies on its own, the Bio-Ontologies Consortium has evolved, McEntire said. It currently acts as an umbrella organization to support the efforts of individual groups building ontologies for a variety of specific fields, ranging from biochemical pathways to pharmacogenomics, cell signaling, and neuroanatomy.
“The ontology created by the GO folks is very good,” McEntire said, “and we’re very happy to see it there in the community, and actually being used, which is even better. But we feel it’s a fairly slim ontology right now, and there are some things that the Bio-Ontology Consortium can do to beef it up.”
The Bio-Ontologies Consortium is currently exploring involvement with the Object Management Group’s Agents Group, McEntire said. Consortium members will be present at a meeting of the OMG’s Ontologies Working Group at Irvine, Calif., in February.
Cherry expressed caution about aligning the gene ontology with a standards group, however. “GO was started by a bunch of academics and we all hate committees,” he remarked.
In an effort to avoid a “holy war” over nomenclature, he said, “we wanted to start small and see if even we could agree, and if we could then maybe others could come in.”
The growing project is currently facing a dilemma as it decides how to best implement the developed ontology. “We’re cautious about expanding because of the problems that that can bring, but at the same time we realize that for this to really be useful, others will have to be involved,” he said.
Cherry maintained that if GO ends up being the standard, it should be because it’s a useful tool, “not because a lot of people decide this is the way to do it,” he said.
“We’re hoping it will survive on its own merits,” he said.
— Bernadette Toner