Despite the I3C’s assurances that it is working to benefit the entire informatics community, some open source developers remain skeptical that the consortium’s final specifications will be in line with their work. Moby, a project led by Mark Wilkinson of the Plant Biotechnology Institute of the National Research Council of Canada, is gaining ground among researchers seeking an open source solution to the database interoperability problem.
Moby wasn’t launched as a direct response to the I3C’s activities. It had its beginnings in the model organism database community, which held the whimsically named “Model Organism Bring Your-own Database Interface Conference” (MOBY-DIC) last September in an effort to improve data sharing between projects. The web-services-based specification that grew out of the meeting happened to develop in parallel to similar efforts within the I3C’s technical architecture group.
Researchers from the Arabidopsis Information Resource, Berkeley FlyBase, WormBase, the National Center for Genome Resources, the GO Consortium, and several other groups are currently involved in Moby’s development.
A prototype of Moby-Central, the system’s registry for cataloging available services, was recently released at the project’s website, biomoby.org. A prototype Moby-Client, which discovers and executes web-based biological services was released last week.
Other components of the system include Moby-Servers, which register with Moby-Central and provide services such as database lookup or sequence alignment. Data is passed in the form of XML-based objects.
Wilkinson said the system is compatible with other projects based on data-only object models, such as MAGE or standards that may be developed by the I3C. “We’re not trying to re-invent the wheel,” he said. “We want to be as compatible as possible with as many other projects as we can.”
While stressing that he’s not overtly opposed to the I3C, the possibility of unforeseen restrictions in the group’s upcoming standards makes a guaranteed open source option a valuable resource, he said. In addition, he said, “We are exploring the possibility that a different approach to the problem might work better.”
And while it seems unlikely that the two efforts will merge in the near future, the possibility may not be that remote. Tim Clark, interim chair of the I3C board, told BioInform that the group intends to “preserve freedom for open source developers and academics” by releasing an open source reference model for each use case and keeping the interfaces that are developed in the public domain. In fact, OmniGene, which acts as a “control center” in the I3C’s current demonstration, is an open source project (omnigene.sourceforge.net), and its developer, Brian Gilman, co-chair of the I3C technical architecture working group, is also a Moby developer.
Gilman said he thinks Moby would fit nicely within the I3C’s current architecture, but the two projects should also be able to co-exist separately as long as they remain compatible.
Moby was recently awarded Can$340,000 (US$221,000) from the Genome Prairie division of Genome Canada as part of a larger bioinformatics project led by Christoph Sensen of the University of Calgary. The award is contingent on a matching grant, so Wilkinson is currently seeking corporate sponsorship for the project.