almost three years ago from Xerox Palo Alto Research Center, where he spent eight years building a research program for microelectromechanical systems, or MEMS. He came to Intel to build the same kind of research program, making the move in part because he felt he could get access to better facilities at Intel. At Intel, Berlin's Precision Biology research team has demonstrated its ability to use microfluidics to isolate and interact with molecules, which it says is an important step toward better diagnostics. Ideally, the group will develop probes that could determine the molecular signatures of diseases.
| Andy Berlin
While Intel might seem like an odd place for a biotech researcher, Berlin says the company offers some of the most advanced chemistry facilities in the world. "Making billions of transistors defect-free is a chemistry problem," he said. The equipment Intel uses to keep its chips defect-free also applies to building diagnostic systems on a single chip. While companies such as Aclara BioSciences and Caliper Technologies are also developing "labs-on-a-chip," Intel is taking a different tack.
"The people who buy labs-on-a-chip are not doctors. They're biologists trying to do basic science," Berlin said in an interview at Intel Research's first-ever open house in March. What's different at Intel is that Berlin wants to put such tools in the hands of doctors treating patients.
Sensing a Problem
While Berlin cautions that his work is still very early research, observers say Intel has everything it needs to create new kinds of products for the healthcare industry. Martin Reynolds, research fellow at Gartner Dataquest, thinks Intel could drive development of DNA sequencing systems for diagnostics and silicon implants for activities such as monitoring the body's health.
"Intel has the money, the need, and obviously the technology," Reynolds said.
Even flow: Intel's "3-D hydrodynamic focusing" allows fluid flow to be shaped within a microfluidic channel on a chip (prototype shown). Biological molecules can be precisely positioned in three dimensions within the channel and moved past a molecular analyzer.
Now it has to prove itself. Berlin's team of scientists has published one research paper on its work developing new techniques for diagnosing diseases and has several in peer review. The group is close to delivering a prototype diagnostic device to doctors (Berlin declined to reveal specifics). "You'll see us delivering working prototypes to the medical community this year," Berlin said.
He would not indicate if Intel will test a lab-on-a-chip, commenting only that the company will pick from several kinds of diagnostic tools under development. The specific device will depend on what best meets the needs of the medical center the group decides to work with. The goal is to have the first device in the field in December, he said.
The device might require a bioinformatics component, Berlin said. Molecular analysis of blood made possible by his group's lab-on-a-chip research generates gigabytes of data, though other types of research focus on less compute-intensive applications.
Despite the forthcoming prototype, Berlin's quest to combine silicon with biology to make wetware could remain basic research.
"Andy Berlin is doing fascinating research, but there's no guarantee that anybody will ever see any of it become a product," said Tim Mattson, Intel's industry sector manager for life sciences. Mattson said Intel hopes Berlin's research will help create new markets for the company but recognizes that it won't know for at least three more years. At this point the company doesn't have a plan for how it would bring resultant products to market, and in particular how the data generated by those products would be managed.
"We know the problems we're trying to solve in life sciences are difficult, and we really don't know how we're going to solve them," Mattson said. If Berlin succeeds, though, his research at Intel may someday yield the tools to diagnose diseases like cancer years earlier than can be done today.
At least one other project at Intel Research may spawn demand for bioinformatics tools. Intel researchers have been studying the potential for networks of wireless sensors, called motes, to make it easier for Alzheimer's patients to live on their own or be cared for by family. These sensors could collect data, such as vital signs, and transmit the information to a caregiver or family member. The company is preparing to test these sensor networks in actual homes in early 2004. Ultimately, this research could lead to the deployment of tens of millions of motes for healthcare applications.
|"We know the problems we're trying to solve in life sciences are difficult, and we really don't know how we're going to solve them."
—Tim Mattson, Intel
"This is like standing on the bridge of the Starship Enterprise, except all of it is possible," said Dan Hutcheson, president of semiconductor industry tracker VLSI Research. Hutcheson thinks that combining wireless networks and MEMS-driven sensors will eventually remake how computers — and people — interact. Healthcare applications will drive this redesign, and Intel could be a major beneficiary of the shift.
"Intel is really good at packaging things in very small packages, and really good at building computer-based systems," Hutcheson said.
Healthcare, at times, has proven a reluctant adopter of IT. It may render Intel's wetware all wet, its sensors nonsense. But it's just as likely that Intel could wind up a prominent supplier of healthcare technology.
Michael Fitzgerald is a technology journalist based in Oakland, Calif. His stories have appeared inThe Economist, Computerworld, and other magazines.