Heightened competition, globalization, and spiraling R&D costs are causing many companies to rethink innovation. Firms that once developed all of their products and services behind a veil of secrecy are beginning to reach outside their walls to source ideas, inventions, and uniquely qualified minds from a vast global talent pool. The result is a new model of innovation where companies conceive, design, develop, and market products with a diverse community of external participants.

Nowhere are these new collaborative models potentially more transformative than in the partnerships between firms and scientific communities to advance the basic sciences. Though fundamental to the long-term capacity of industries to remain innovative, such research is fraught with uncertainty and the risk that costly, long-term projects may fail to churn out a marketable product. High costs and uncertainty, in turn, have forced many companies to scale back their investments in basic research. Though innovators still need to know the underlying sciences, their primary aim in-house cannot be to further them. For that they will increasingly rely on partnerships with universities and other research organizations, while corporate research teams use their skills and resources to move quickly to practical application.

Public-private partnerships sound simple in theory, but in practice they’re difficult. When Don Tapscott and I went looking for a good model to include in our book Wikinomics we came across the Alliance for Cellular Signaling (“AFCS”), a bold new open science initiative. The project exemplifies the emerging scientific paradigm and provides a template for how open source tactics can be leveraged to solve complex biological problems.

The project began with a compelling idea: build a fully functional computerized model of the cell that could provide researchers and companies with an incredible new engine for research and pre-clinical testing. To do that, scientists would need to map what they call complex signaling networks – the pathways that transmit different molecules as a form of communication within the cell. Robin Irvine, a pharmacologist at the University of Cambridge, equates these microscopic communications networks to “a thick soup of proteins talking to each other in ways we just don’t understand.”

The prize for decoding this Byzantine language offers considerable incentive to try. Signaling networks are at the root of the unintended side effects that cause promising drugs to fail in clinical trials. A computerized model of cellular communication could give researchers the ability to reliably predict the responses to treatments that have yet to be tried in the clinic or laboratory. Considering that clinical studies represent about 40% of average cost of developing a drug, a virtual testing capability would be a major boon to the industry.

The problem is that cells have about as many moving parts as a jumbo jet but are only the size of, well, a cell. US-based jet manufacturer Boeing has just recently reached the point where it can simulate all of the electro-mechanical parts in an airplane, so just imagine the intricate complexity of building a comparable model with nuclei, ribosomes, mitochondria, and thousands of different protein molecule

To succeed, the project needed an organizational approach that mirrored the complexity of its goals. Identifying all relevant signaling molecules, determining how they are connected, mapping and quantifying how information flows among them, and building mathematical models that simulate cellular signaling in action is a complex undertaking. It requires a range of expertise and resources beyond the reach of a single research group or institution. Meeting those goals called for the direct involvement of hundreds of scientists and technicians and indirect input from thousands more.

The ambitious multi-million dollar project got underway in 2002 with the leadership of Alfred Gilman, a Nobel prize-winning pharmacologist at the University of Texas. The National Institute of Health provides the bulk of the annual budget of ten million dollars. The remaining thirty-five percent comes from companies that are eager to tap the insights, such as Lilly, Johnson & Johnson, Merck, Novartis, Chiron, and Aventis.

To make some sense of this vast uncharted territory, Gilman arranged the Alliance into three groups. The first group consists of fifty lead investigators who monitor and direct progress of the Alliance laboratories. The second group includes an army of PhD students and technical personnel who do the grunt work in the eight universities around the US. The third group is a vast worldwide network of researchers that helps the AFCS accomplish its goals by lending their specialized expertise in particular molecules.

The overall effort is orchestrated with a mix of local control and large-scale self-organization. Of the eight labs, seven host scientists who work with living cell specimens. Each of these “wet labs” specializes in a distinct aspect of the cell-signaling network. The eighth lab is the context provider; it aggregates and integrates all of the data produced by the others and is building the computational model.

Most impressive is AFCS’s ability to tap the knowledge of over 100,000 cell signaling researchers worldwide. As the AFCS researchers do experiments on various pieces of the cell signaling puzzle they post their data and results in a publicly accessible databases. Each discovery then has to be validated. “We don’t begin to have the manpower or the talent to do that,” says Gilman. So Gilman has been enlisting the whole signaling research community to participate in advancing research in their related areas of expertise.

So far the AFCS has recruited over 1,500 external experts, each of whom takes responsibility for monitoring and updating research on two or three molecules. These experts host “molecule pages” on the AFCS Web site, which provide dynamic repositories for information on each of the nearly 4,000 protein molecules involved in cell signaling reactions. Over 100,000 additional researchers receive weekly project updates and use the AFCS data in their own research. Fluid interchanges across the entire community provide the basis for a real-time knowledge creation and peer review engine of unparalleled size and diversity.

In this respect, the AFCS is a model for the new Web-enabled paradigm of scholarly research and communication. The community of interested researchers maximizes the speed, efficiency and veracity of its efforts though large-scale, synchronized collaboration. All of the data and publications are shared freely. And results are published immediately on the Web, where they are vetted, cross-referenced and reported on voluntarily by peers.

Eschewing conventional peer-reviewed scientific journals in favor of a more expeditious Web-based publishing model was controversial for some members. But getting the data in the hands of interested researchers within days instead of months took precedence over academic conventions.

AFCS participants also agreed to give up any intellectual property rights in their research. Gilman felt this would maximize access to the project and eliminate potential sources of friction. “We forgo all intellectual property rights because we totally appreciate that researchers in the signaling community at large will follow up on the leads that we provide only if they are assured of a level playing field,” says Gilman, “Insiders have no special advantage. After all,” says Gilman, “the real intellectual property ultimately lies in identifying the genes and gene products that will hopefully become drug targets, and then finding the drugs that will alter the behavior of these molecules appropriately.”

How did the private supporters react to this decision? “We have tried to convince the pharmaceutical industry that our results will be enormously useful to all of them and that if they all were to participate their financial contribution could be relatively modest,” says Gilman. “After all, a proper virtual cell will be an incredible drug discovery engine.” And, with over 100,000 people collaborating, the project’s output will far exceed what any individual company could do alone.

Private sector partners do get some private benefits. Gilman says, “Companies who are not sponsors will still have access to our data on our Web site. But we will work hard with our sponsoring companies to help them interpret our data and realize value from it. They also get preferential access to new Alliance technology.”

Gilman is confident that the investments will pay off. With the project still about five years from completion, it’s too early to judge. But the AFCS is a useful model for public and private sector players seeking to leverage the benefits of large-scale collaborative efforts in embryonic research domains. “If everybody will put in just a little bit of time,” said Gilman, “the community will have a great resource.”

Related posts

• No related posts.