Stanford, UCSD Biologists Take Plunge Into Arpanet-Style Project With Sage Bionetworks
[Updated: 6:40 am Pacific] Stephen Friend had plenty of doubters last year when he quit a high-powered Big Pharma executive job to start a nonprofit seeking to spark an open source style movement for biology.
One year later, the former chief of cancer research at Merck has pulled off another unlikely feat. He’s corralled four world-class biologists who have agreed to pool their raw experimental data and models on the connections between genes, proteins, drugs, and disease states into a public database. This is the stuff scientific careers are made of, and can lead to blockbuster papers in top journals like Science and Nature, which is why it’s usually held close to the vest.
The biologists willing to take this plunge are Atul Butte of Stanford University, Trey Ideker of UC San Diego, Andrea Califano of Columbia University, and Eric Schadt of Pacific Biosciences (and soon to be at UC San Francisco). Getting stars like that on board is a critical step in the quest that Friend embarked on more than a year ago when he founded the nonprofit Sage Bionetworks. Seattle-based Sage, after some iffy early moments, has stitched together a support network that has committed more than $20 million, more than enough to fund operations for two years. Major drugmakers (Pfizer, Merck), government agencies (the National Cancer Institute, the Washington state Life Sciences Discovery Fund), and philanthropies (the CHDI Foundation) have made some of the earliest investments. They have all lined up to support the vision of allowing biologists on a grand scale to pool data and their collective brainpower in open databases. The hope is to better connect the dots between malfunctioning DNA, RNA, and proteins, to see how those things get manifested as symptoms of disease that a doctor observes in the clinic.
It all looks great on a whiteboard, but without real participation by key researchers, it’s hard for something like this to ever move beyond the theoretical stage. With stars like these contributing their data, Friend has said, the ultimate result would be more effective drugs, just as programmers contributing open-source code can create better software. On my most recent visit to Friend’s office in Seattle last week, he said these four biologists are carrying out a potentially groundbreaking experiment for biology analogous to the federal government’s investment in the Arpanet of the ’60s, which gave rise to the Internet of today.
This experiment is “really modeled” after the Arpanet, Friend says. “Four universities decided to send data back and forth. This is an equivalent project. What would happen if four groups at top universities wanted to share their data? We’ve asked what does it really take to do that?” [Updated: the Sage network actually has five sites, since the four additional universities have joined the founding group in Seattle.]
This isn’t the only effort that attempts to get biologists to pool data and experimental models in some kind of consortium, although others tend to be more specific to a certain disease, or type of data. The bigger vision is part of what Ideker says enticed him to get serious about Sage.
“After talking with Stephen, and later working together with him, it became quickly apparent that he had the right combination of vision, assembled expertise, and contacts to make this work,” Ideker says. “Athough Sage may not be the only game in town, it is likely to move the ball furthest downfield.”
Butte added: “The real open question is whether working with the Sage Federation will be ‘more’ (in some way) than just working with a group of collaborators… in terms of scalability, data, transparency, or something else…it’s just not clear what yet. But of course, it’s worth trying this model out.”
One year into this experiment, Sage is now at the point where the questions aren’t about what it’s trying to do, but how it’s doing it. Friend racked up more than 250,000 miles of travel last year, as he was on the road every week, and out of the country at least once a month, he says. He says he’s pushed himself harder this year than at any point in his medical residency, fellowship, or his 8-year executive stint at Merck. “Every time I interact with a patient group I’m reminded of the fact that it’s worth maintaining this intensity,” Friend says.
Now that Sage has received some degree of financial security, and some validation from academics willing to give it a shot, comes the hard work of operations. Sage has grown from an original team of 14 people, housed at the Fred Hutchinson Cancer Research Center, to a little more than 20 today, Friend says. There are huge technical challenges to dive into now. One of the first is in working on what he calls “disease maps” in which biologists can better predict what will happen when a given drug interacts with a molecular target, particularly how that will affect disease symptoms and side effects. Unlike building an airplane, where engineers can predict what will happen by running aerodynamic simulations on a computer, today’s disease models are “incomplete.”
As if that weren’t daunting enough, that’s really just the first challenge. Sage needs to convince more and more biologists to dump their data into the open. Friend has spent much of his time jawboning Big Pharma execs, seeking to talk them into handing over huge clinical trial databases that match up genetic data and clinical results, without handing their competitors data they consider proprietary. Friend says a few pharma companies have already gone along with this, by agreeing to hand over partial databases in which results on their innovative new drugs is withheld, but in which data from patients in control groups is made available.
Once the data arrives at Sage, there’s still more fundamental work to be done before it’s useful to anybody. Sage is building a computational platform, in which software engineers are writing new algorithms, which help put these diverse sets of data into formats that are useful. For now, Sage has found a place to store all this data on servers at the University of Miami, where the nonprofit is getting a bargain rate from its host by allowing scientists there a certain amount of access to the database, Friend says.
Writing the software to analyze all this data, and figuring out how to store it are clearly a couple of the big operational challenges ahead as Sage hopes to entice more companies like Merck and more researchers like Ideker and Butte to get on board. Sage is looking at options to use cloud computing options like those offered by Amazon as a way to store the data in an easily scalable and accessible way as it grows. And Sage is actively hiring software engineers in anticipation of this growth—Friend said he expects Sage to have 35 employees a year from now. Getting the right people on board, the same kind of risk-seekers who helped build the organization in its first year, will be key.
All the technical stuff is important, but won’t mean much if Sage doesn’t win the cultural battle. He’s been reading up on Nobel laureate Elinor Ostrom‘s work on how game theory can be used to spur collective action. He’s incorporated some of what he’s picked up when he takes his case to university presidents and tech transfer offices. The hope is to help set up policies that reward scientists for pooling their data in the open, not necessarily by filing patent applications, Friend says. “You have to build shining examples,” he says, on getting people to convert.
Not everyone, of course, buys in right away for Friend’s argument for urgent, revolutionary change in how biologists generate, share, and hope to potentially profit from their data. It is, after all, challenging the status quo of academic research and the for-profit biotech and pharmaceutical industries. Groundbreaking ideas in biology-based industries take years to prove their mettle in experiments in everything from mice to real human beings, who need to be tracked for years of follow up in order to answer basic questions.
This isn’t something that will happen overnight. That’s not easy for Friend to accept, but he understands it’s just reality. One of Sage’s directors, Hans Wigzell of the Karolinska Institute in Sweden, has advised Friend to “rush slowly.” While that sounds like an oxymoron, Friend explained what it means.
“It means keep the momentum, keep the energy up, don’t get complacent, and don’t give up,” Friend says. “But at the same time, acknowledge that the time this will take is measured in half decades, or decades. We’ve wandered into an area where it’s not about quarterly progress reports on the big mission. Fortunately, this is a nonprofit where you can do that.”