Tech Cos Unite To Solve Problems Of Science-Data Sharing
.
By Carmen Fleetwood
Of DOW JONES NEWSWIRES
NEW YORK (Dow Jones)--Some scientists scribble results on paper, others type them into 15-year-old computers. Meanwhile, their technophile colleagues rely on cutting-edge machines.
Individuality is great, but how do you share results from experiments with your colleagues?
That's why Microsoft Corp. (MSFT), of Redmond, Wash., was among the companies to start BioITAlliance, a group which wants to improve the current situation in computational biology. Other members include drug makers, medical device companies, software makers and researchers, including Accelrys Software and Amylin Pharmaceuticals.
In 1665, the Journal des scavans, of France, was the first widely known scientific journal to publish results from physics and chemistry experiments. But with the massive amount of information spawned by recent projects such as mapping out the human genome, the data produced exceeds anything ever known in terms of quantity or complexity.
Don Rule, a platform strategy advisor for Microsoft, said "based upon conversations with customers and analysts, at least 15% of research is redundant because researchers either cannot find prior work that is relevant to them or they cannot interpret the results."
Other scientists and experts agreed that advances in many areas are slowed because of redundancy. The constant repetition means that both money and time are wasted.
Current costs to develop a new drug for the market now exceed $1 billion, compared with about $200 million in the 1990s. That doesn't even guarantee a viable drug. And even if a drug is approved, it could still be a target of a lawsuit, so the pharmaceutical industry is keen to cut costs whenever possible.
In science, "being able to replicate an experiment, is the first step" said Fiona Murray, a professor at the MIT Sloan School of Management, Cambridge, Mass. The second step is then to build on the existing data.
If the duplication can be eliminated by making different software and tools compatible for all researchers, the results would have a wide impact.
"These tools aren't discovery," but they allow us to drive science faster, said Peter Kuhn, professor of Scripps Research Institute, a biomedical non-profit research institute in La Jolla, Calif., and member of the BioIT Alliance.
Right now, Microsoft and Scripps are working on a first project, Collaborative Molecular Environment, which will help manage, archive, annotate and capture molecular data using several software programs such as Microsoft Office. Microsoft expects to have a pre-release version in late May.
Other possible BioIT projects planned include using technology to help compare genome types with clinical data to find out if a certain gene is more susceptible to certain ailments or reactions.
The alliance would like to help reduce the dependancy on recording data on paper by encouraging researchers to keep it electronically on such devices as portable notepads so that annotations and other information better organized.
The creation of a common data exchange format would make it easier to load data from different software into one format that can be widely used, said Anthony R. Kerlavage, a senior director, global service development and support for BioITAlliance member Applied Biosystems Inc. (ABI).
While it's a niche audience at this point, with about 50,000 people worldwide working in the computational biology sector in mostly a research setting, Microsoft sees a potentially lucrative market for its data management software business.
"When you start to see applications moving into the clinical setting, those numbers will grow," Rule said. But the real opportunity will come when you begin to see personalized medicine reaching individual doctors and patients. While we don't have exact numbers, we recognize this market is too large to ignore."
For example, if someone were to develop lung cancer, a test would then reveal what treatment is likely to work best on that specific patient and the drugs would be made with those specifications. The capability to do just that would be across the board.
Charles P. Waite, a general partner, at OVP Venture Partners estimated that this field right now is worth about $500 million annually, but that includes both in-house and outside companies as well as a number of specific fields such as genomic software.
Different segments have various growth rates ranging from 11% to the 50% expected for genomic software, according to Waite.
Other rivals are in the health-care database software business including International Business Machines Corp. (IBM) and Oracle Corp. (ORCL). Contacts for both companies weren't immediately available for comment.
-By Carmen Fleetwood, Dow Jones Newswires; 201-938-5216
(END) Dow Jones Newswires
May 01, 2006 14:55 ET (18:55 GMT)
Copyright (c) 2006 Dow Jones & Company, Inc.- - 02 55 PM EDT 05-01-06 |
