New Brain Trust to Work Like the WebBy M.L. Baker | Posted 12-12-2005
A software vendor and a nonprofit group are teaming up to create NeuroCommons.org, a free, shared repository of data and other tools to speed research on brain function and disease.
Informatics company Teranode will provide an infrastructure and means to store disparate data in common formats. Science Commons, a project of the nonprofit corporation Creative Commons, will develop a community of users and experts, plus work to help create an intuitive interface to find and analyze content.
Science Commons is housed at the Computer Science and Artificial Intelligence Laboratory at the Massachusetts Institute of Technology. Teranode and Science Commons announced the partnership on Monday and plan to launch NeuroCommons.org in the second half of 2006.
There's a real need for a shared platform in neurology, said John Wilbanks, executive director of Science Commons. Separate research foundations exist to fund different rare diseases, but they cannot share information without running afoul of technical and legal complications.
One hope is that researchers can gather preliminary evidence for their hypotheses using other researchers' datasets. NeuroCommons.org should also allow researchers to readily compare proposed mechanisms about what, how, and when various genes and proteins interact.
Neurologists would use an interface much like a Web search engine, but instead of finding relevant Web sites, they would be able to find other researchers' datasets and protocols, as well as working models of how genes, proteins and brain regions interact.
Even better, NeuroCommons.org could automate such tasks and analyze the results. Researchers would not need to spend days doing literature searches or hunting with several available databases for useful data, said Matthew Shanahan, CMO for Teranode. That's especially important as the number of proteins and genes associated with diseases swells. "The thought that a scientist can do that manually efficiently doesn't make sense; you really need the aid of software now."
Teranode is responsible for figuring out how to get data from widely varying sources, from brain scans to gene chips, into a format that can be searched intuitively. Because XML is "insufficient to represent the complex data of life sciences," another mark-up language, RDF (resource description format), is used to support what is being called the "semantic Web," said Shanahan.
"All the Web can do is find a document for you and display it for you," said Wilbanks. "The semantic web marks things up in a more concrete manner; it says that there are relationships." For example, a scientist could search for peer-reviewed articles about a particular gene, data related to that gene, or models about how that gene might affect other genes and proteins.
Neurocommons.org is set up to be maintained by its community of users. Researchers will be able to annotate each others' data.
Wilbanks hopes that, eventually, researchers will see contributing information to the semantic Web as part of their scientific duty, much like peer review. But he admits that it isn't yet part of scientific culture. "It's hard to get someone to take the time to say, 'I'm going to make my data reusable by someone that doesn't know me.' "
He believes scientists will be converted once NeuroCommons.org demonstrates that it can help ask new kinds of questions.
Check out eWEEK.com's for the latest news, views and analysis of technology's impact on health care.