Teranode Aims to Ease Data Modeling

 
 
By M.L. Baker  |  Posted 2004-10-18 Email Print this article Print
 
 
 
 
 
 
 

The informatics company's design automation software and equipment lets users create, reuse and tweak visual models tied to experimental data.

Teranode Corp., an informatics software company in Seattle, launched updated versions of software and new equipment on Monday, promising to tear down data silos and help companies turn data into information. Most life sciences companies use a hodgepodge of spreadsheets and other applications to store data, said Matthew Shanahan, Teranodes chief marketing officer. Typically, these arent searchable and dont let scientists build flexible models of the interactions they believe are occurring in cells or other body systems.
Making matters worse, the pace of data collection is rapidly increasing, forcing companies to hire more people just to deal with integration. Shanahan said some companies are "hiring a bioinformaticist for every two biologists—that doesnt scale."
Teranodes system lets users create, reuse and tweak visual models tied to experimental data. "You have the data from your experiments and the model that you built of what you think is happening, and you can see them side by side," said Jack Beusmans, principal scientist in AstraZenecas pathways group, which studies interactions between networks of cellular proteins and their effects on disease. If the data and the model dont match, he said, its very easy to look at the protocols used to collect the data to see if experimental conditions explain the disagreement. The software also lets scientists search the database for other data that might impact the model. "One of the main applications for the model is communication," Beusmans said. "It has a very simple and intuitive graphical interface that you can show to biologists and they will immediately understand it, but it also has good underlying mathematical simulation technology."
Also, the system is set up so that the model can be easily documented, he said. For example, a scientist can flag reservations about a particular part of the model and solicit feedback from other scientists. Beusmans said he has been using an alpha version of the software for about eight months and that the price allowed scientists to try the system out tentatively, "Other packages are too expensive to do on a trial basis, but Teranode is priced so that you can actually do that." Teranodes Web site also lists drug giants Amgen Inc. and Pfizer Inc. as clients. The new version of the design automation software, Teranode Design Suite V2, starts at $1,000; a Model Server, which provides "check-in" and "check-out" functions and enhanced search capabilities, starts at $25,000 per CPU (half that for some academic settings); and an API that can integrate laboratory equipment, domain-specific analytic tools and data repositories starts at $5,000. The data and model are stored in an XML model called VLX, which also works in a relational database. VLX is an open specification that other researchers can build to, said Shanahan, but the ability to integrate protocol data with the biological model is patented. The software also supports SBML (Systems Biology Markup Language) and KEGG (Kyoto Encyclopedia of Genes and Genomes). This helps researchers include information from public databases in their models. Derek Walker, program manager of clinical research informatics at the Fred Hutchinson Cancer Research Center in Seattle, said he purchased the system as a way to share data across various labs. "I had been looking at lab notebook products, but they werent good at integrating," he said. "Data from each lab was so inaccessible, and we didnt have the metadata that would let us compare data from one lab to another." He said the Teranode interface was flexible enough that it could work across labs without a lot of tinkering. "Weve written a couple specialized plug-ins that plugged in well," he said, but generally, "theres not really any code to write, it just sort of flows." Sometimes, though, the system could be too flexible, allowing scientists to add searchable attributes too easily. Beusmans and Walker said the data collection, integration, and modeling functions are easy to learn. "Im advising an undergraduate student at Brandeis," Beusmans said. "I showed her the software for an hour, and she was modeling away. Thats pretty amazing." Check out eWEEK.coms Health Care Center at http://healthcare.eweek.com for the latest news, views and analysis of technologys impact on health care.
 
 
 
 
Monya Baker is co-editor of CIOInsight.com's Health Care Center. She has written for publications including the journal Nature Biotechnology, the Acumen Journal of Sciences and the American Medical Writers Association, among others, and has worked as a consultant with biotechnology companies. A former high school science teacher, Baker holds a bachelor's degree in biology from Carleton College and a master's of education from Harvard.
 
 
 
 
 
 
 

Submit a Comment

Loading Comments...
 
Manage your Newsletters: Login   Register My Newsletters























 
 
 
 
 
 
 
 
 
 
 
Thanks for your registration, follow us on our social networks to keep up-to-date
Rocket Fuel