Will Open Source “SAGE” platform lead to more transperency to MERCK’s research and clinical trials process?
Consumers can only hope. Of course if MERCK corporate lawyers figure out that such an open source platform may lead to a clear chain of custody of the type of information they would rather not have readily available during a trial environment, we will see who prevails, the scientist or the lawyers
Reported by Kevin Davies
March 3, 2009 | SAN FRANCISCO – Merck scientists and executives Stephen Friend and Eric Schadt unveiled their plans for Sage, an open-access platform for sharing and disseminating complex data representing disease biology, in a major announcement at CHI’s Molecular Medicine Tri-Conference last week.*
In a joint presentation, Friend, Merck senior vice president and former oncology chief, and Schadt, an outstanding researcher based at Merck’s Seattle subsidiary Rosetta Inpharmatics, reviewed the successes and outstanding challenges that prompted them, with Merck’s blessing (in the form of money and resources) to entertain a bold new approach to improving the expense, time and productivity of drug development.
The benefits of analyzing complex bionetworks are very good, said Schadt, but “more expensive than any one company can afford.” The vision of Sage was “to create open access, integrative bionetworks, evolved by contributor scientists, to accelerate the elimination of human disease.” An all-star advisory team includes Nobelist Leland Hartwell, Sir David Lane (A*STAR Singapore), Navigenics co-founder Dietrich Stephan, Merck research chief Peter Kim, Yale’s Rick Lifton, and John Wilbanks (Science Commons).
“We need massive amounts of information appropriately integrated to build models that are predictive,” said Schadt. “Scientists across the globe involved in different areas of research need to be actively engaged in accessing these networks and contributing information back.”
The transition from a linear to a network mindset would require the generation of coherent datasets, the development of predictive models to design novel therapeutic approaches, and the leveraging of social networks and other means to foster a contributor network. “Watching the trends of public data access, we anticipate a transition of disease biology into the precompetitive space,” said Schadt. Friend added, “The concept of making disease biology a pre-competitive space is… something that we feel in the long run has an opportunity [to succeed].”
Schadt said Merck’s leadership had recognized an opportunity where donating some of its data into the public domain, forming an open access platform that will emerge from an incubator phase, would provide a potentially significant long-term understanding of disease biology.
As for why scientists should include their own data, Friend said, “picture chemistry, picture physics. The people who were originally trying to mix compounds didn’t get very far until they found molecular structures… This is the analogy for what’s going to happen in biology.” New representations of disease allow for data to be shared and layered.
The hardest part in making Sage successful may not be the technology, Friend concluded. “It’s either going to be … our institutions … that have a certain culture about what we do with data. Or it’s going to be the clinicians,” who aren’t used to presenting clinical data using defined standards.
Decade of Discovery
Over the past decade, Friend said Merck has introduced numerous bold technologies that have been successful to a degree. Widescale RNA expression profiling in tumors (in conjunction with the Netherlands Cancer Institute) led directly to the development of Mammaprint and Oncotype diagnostic tests for breast cancer metastases. But such measurements are confounded by multiple variables, making it impossible to infer causality.
Merck also championed whole-genome RNA interference (RNAi) screening. “Often we use it to choose what drug should be combined with what standard of care, and what patient is likely to respond to what therapy,” said Friend, such as identifying gene networks that influence the activity of cisplatin. But heterogeneity of samples made it almost impossible to put the results into context. “It’s like looking at a single frame in “Slumdog Millionaire” and going, Ah, that’s what that movie was about,” he said.
A third initiative, beginning around 2002, was to merge databases of clinical information and genetic information. Merck forged collaborations with European and Chinese cancer institutes, as well as the Moffitt Cancer Center, which enables Merck researchers to direct patient selection in clinical trials based on molecular signatures in the database. But Friend said that the volume of disease data amounted to “a clinical/genomic Tower of Babel” problem.
More recently, Merck has been riding the success of Schadt’s team in Seattle, which has taken major steps to harness the explosion of data and analyze biological networks to predict the physiological state of the system. “Drug companies were betting the farm on seeing things correlated with disease and beautiful patterns of expression and then developing drugs, without having any real idea of the casual nature of those patterns,” said Schadt. The key was to leverage DNA information and environmental effects. “To be competitive in the future and to impact human health, we must become masters of information,” Schadt said, displaying a picture of Aria, the all-seeing master computer from the film Eagle Eye.
Ed’S NOTE: Eric Schadt will keynote the 2009 Bio-IT World Expo on Tuesday, April 28.
*CHI’s Molecular Medicine Tri-Conference. San Francisco, February 23-26, 2009.