Sometime within the last year, the Interoperable Informatics Infrastructure Consortium (I3C) quietly disappeared. Sadly, perhaps, almost nobody noticed.
Researchers and vendors launched the I3C with the noble goal of developing interoperability standards for the life sciences that would make it easier to access, exchange, and share data.
To its credit, the organization made some significant progress, chiefly in the development of the Life Science Identifier (LSID) specification, which tags data with a uniform URL-like address and descriptive name, making data easier to recognize and locate over the Internet. (See: LSID: An Informatics Lifesaver.)
So why did the I3C just vanish with so little fanfare? Opinions among some of the I3C founding members vary, but the consensus is that the work of the I3C is being carried out today in other standards bodies.
Back in Time
When the I3C incorporated several years ago, there were about 75 participants. Member organizations included IT vendors Compaq, IBM, and Sun Microsystems; life science and informatics vendors Accelrys, Affymetrix, and LION bioscience AG; and academic/ government organizations including the National Cancer Institute, Los Alamos National Laboratory, and the Whitehead Institute.
“What energized me [at that time] was the vast amount of public information that was becoming available,” says Andy Palmer, CIO at Infinity Pharmaceuticals and past president of the I3C. Public databases funded by the NIH and other government agencies contributed to the free data glut. “All of these databases were islands of information that were relatively disconnected,” says Palmer. The idea of developing standards to bind together data from public and private sources “resonated with me,” he says.
Vendors were also noticing the same issues. “There was a lot of attention focused on interoperability of systems,” says Howard Asher, founder and chairman of the Life Sciences Information Technology (LSIT) Global Institute. “The need for interoperability was stimulated a lot by database issues.”
In the early days of the I3C, Asher was at Sun Microsystems, where he served as group director of global life sciences. His former colleague Sia Zadeh, then group manager for life sciences, also helped drive the I3C as the organization was coalescing. (Zadeh has since left Sun.)
“The interest in interoperability was appropriate at the time,” says Asher. “Data and data management were at the forefront of the [birth] of genomics data.”
Trouble Signs
Interoperability and data management remain vital issues for life scientists today. So what happened to the I3C? Is there not still a need for interoperability standards?
In fact, there were some concerns about the viability of the I3C early on. As Bio-IT World noted back in 2002, at the time the I3C incorporated, (see “Chaos In, Order Out” April 2002), “Don’t start celebrating yet. Consortia don’t have great track records in these matters.”
But the consensus then was that the I3C was avoiding past mistakes. “We’re not trying to solve world hunger,” Zadeh said at that time. “We’re keeping the objectives feasible and achievable.”
Still, founding members had some agita over big pharma’s lack of involvement in the I3C. Caroline Kovac, general manager of IBM Healthcare and Life Sciences, once said: “They are a big missing piece. We absolutely must have big pharma in the I3C.” But big pharma never bought into the I3C concept.
Another issue was the difficulty in forging consensus on a single data representation standard. For example, an I3C working group on imaging was split into fundamentally different camps -- some wanted to take images and describe their metadata and relationships in XML, while others preferred to explore how the DICOM [Digital Imaging and Communications in Medicine] standard could be expanded (see “The Sharper Image,” Bio-IT World, May 2003).
The economic health of some vendors also eroded support. “Budgets were tight and it was hard to demonstrate the value of [I3C membership] to my company,” said one vendor representative.
Some vendor members simply went away through acquisitions. For instance, Compaq was acquired by HP And Optive Research, a cheminformatics analysis tools company that joined the I3C for its work on LSID, was acquired by Tripos.
Meanwhile, other groups such as Health Level Seven (HL7), the Clinical Data Interchange Standards Consortium (CDISC), and the World Wide Web Consortium (W3C) were separately developing standards with the same interoperability and data sharing issues in mind.
Much of this focused work was seen as complementary to the broader focus of the I3C. This non-competing view was evident as the I3C was developing the LSID specification. The I3C looked at some of the work being done by the W3C. The relationship between the I3C and the W3C grew over time and took on a new role last year when the W3C was developing standards for the Semantic Web.
In the fall of 2004, believing that the data interoperability challenge facing the life sciences was the poster child for this new technology, the W3C hosted a workshop on Semantic Web for the life sciences. Evidently the two groups were working on some similar problems. “I asked myself, was the mission of the I3C alive and well and better served in the W3C?” says Palmer.
Others agreed. It was becoming clear that “interoperability becomes less of a concern if Semantic Web comes along,” says Asher. Many I3C members have simply shifted their focus to Semantic Web efforts. Ironically, an important factor that simplifies use of Semantic Web technology on life science data is the adoption of the I3C’s LSID specification by many life science databases.
In some ways, the W3C has picked up where the I3C left off.
Related Stories:
Masters of the Semantic Web
Necessary Liaisons: Making Standards Work
LSID: An Informatics Lifesaver
Chaos In, Order Out
Getting the Gobbledygook Out of Data Sharing