Beyond The Genome The next goal of DNA research makes the breakthroughs of the past few years look like high school biology. Some even say it's impossible--which is one reason Lee Hood is determined to go for it.
By Erick Schonfeld

(Business 2.0) – As science celebrates the decoding of the human genome, the man whose invention made it all possible isn't cheering. At the moment, in fact, he's tapping intently on a laptop in his office, trying to show a visitor why the long-awaited payday from the genome project is still a long way off.

"Here," says Leroy Hood, MD, Ph.D., director of the Institute for Systems Biology, inventor of the automated DNA sequencer, and a gale-force presence in modern molecular biology, "I have a beautiful example of a system." He flips the laptop around so the visitor can see a screen image that looks like overlapping spiderwebs, with circles representing yeast genes crisscrossed by lines representing gene and protein interactions. Hood brings up another screen shot illustrating what happens when a certain yeast gene is disabled by a drug. Several circles in the system are red, signaling that more genes than just the targeted one have been switched off--unintended collateral damage. "This shows you how one little change has very far-reaching effects," he says.

Don't get him wrong. Lee Hood was a principal scientist on the Human Genome Project and would never deny that the undertaking was a towering achievement. Thousands of researchers worked for more than 10 years to map the chemicals that make up each of humankind's 30,000 genes. The only problem is, all that work has little immediate medical or commercial value. "The Human Genome Project has given us a genetic parts list," Hood sighs. Before the wonder drugs expected to sprout from genomics research can arrive, he says, science must learn how all the parts on that list work together.

It sounds like a logical next step--and in a sense it is, notes J. Craig Venter. Venter led the private effort to sequence the genome in parallel with the government-funded Human Genome Project. "Now that we have the genetic code," he agrees, "for the first time in history we have the responsibility to look at how all the components interact to create life."

But talk about a towering achievement. Imagine multiplying the spidery networks on Hood's laptop by 30,000 genes and perhaps 300,000 proteins, and you begin to grasp what Hood is proposing: the most audacious biology experiment ever attempted. "This is orders of magnitude more ambitious than the Human Genome Project," says Michael Phelps, director of UCLA's Center for Molecular Medicine.

That may be putting it mildly. Realizing Hood's vision will require as-yet-uninvented scientific instruments throwing off unimaginable amounts of data. That data will have to be analyzed by as-yet-unwritten software. Some skeptics point out that the project could very well demand more computing power than exists on the planet.

On the other hand, if Hood succeeds, it could finally usher in the biotech golden age that the Human Genome Project once seemed to promise. When they understand the networks of genes and proteins that govern all cell functions, companies will be able to develop tests for genetic system errors that might lead to disease. Others could develop drugs to switch off those errant systems in advance and keep them off--as long as the drug is being taken. For the pharmaceutical industry, this would be the ultimate growth strategy: to make lifelong customers of people with no symptoms.

Even before that point, if Hood's history is any indication, discoveries at his lab could create all sorts of minor growth markets. During his 30-year career, Hood's research and inventions have spawned at least 10 biotech companies, including Amgen (the industry's largest) and Applied Biosystems, now a $1.6 billion division of Applera. Hood and a VC partner named Carl Weissman have already raised $15 million for a business "accelerator" whose sole aim is to spin off companies built on interim breakthroughs at Hood's institute. How do you raise $15 million to commercialize discoveries that haven't yet been made from a project that could very well collapse under the weight of its own ambition? The VC shrugs. "A lot of people realize that Lee is often around the next big idea," he says.

Hood's Institute For Systems Biology perches on the edge of Seattle's Lake Union, along what is fast becoming a biotech corridor. Inside the airy, five-story building, workbenches are lined with microscopes and multihued experiments. Rooms full of DNA sequencers, mass spectrometers, and cell sorters hum continuously, measuring the gene and protein makeup of the samples cooked up in the workbench beakers. The digital output of those machines feeds into local IBM and Sun servers or into the Arctic Region Supercomputing Center in Alaska for a number-crunching comparison against the entire human genome. The idea is to capture as much data as possible about cellular systems so that software can create detailed models of how they operate.

Much of the excitement that genomics currently generates is based on an overly simple premise. Now that we have identified and cataloged each human gene, all we need to do is find out what each one does. If we can find a gene that causes a disease and turn it off pharmaceutically, the thinking goes, we will cure the disease. Hood's research suggests it won't be that simple.

A gene, after all, does little more than encode the recipe to produce (or "express") one or more proteins. The proteins then do the genetic grunt work, interacting in complex networks with each other, with genes, with messenger RNA (the molecules that read and copy DNA), and with other parts of the cell. As a result, a drug that interferes with a single gene often fails to alter the network as a whole, or it causes side effects that the one-gene, one-target model could never anticipate. Unable to see the big picture, drug researchers are shooting in the dark. Hood wants to paint that picture.

To pull it off, he is essentially helping to create a new multidisciplinary field of study called systems biology. His nonprofit Institute for Systems Biology--co-founded with Ruedi Aebersold, the father of the protein science known as proteomics, and South African immunologist Alan Aderem--employs proteomists, computational biologists, chemists, mathematicians, computer programmers, physicists, and engineers. It draws its $130 million in funding from government agencies like the National Institutes of Health, corporations like Merck and IBM (which collaborate on research), and former junk-bond king Michael Milken.

Milken's presence among the sponsors is no surprise, since Hood seems to have a knack for parting wealthy brainiacs from their charitable money. During a four-hour Seattle dinner in 1992, he persuaded Bill Gates to put up $12 million to fund a biotech department, chaired by Hood, at the University of Washington. Seven years later, after the university balked at the multidisciplinary ambitions of systems biology, Hood quit and started the institute. When the world's richest man objected, Hood countered with a blunt analogy: "If a startup stays in Microsoft, it does not have a chance," he said, "because all it tries to do goes against what Microsoft is about."

At 64, Hood has gone white at the temples, but he is still charismatic and athletic--fitting for a man who quarterbacked his Shelby, Mont., high school football team to an undefeated season. As he clicks through the slides modeling the network of yeast DNA and proteins, he explains that the institute has mostly confined its systems approach to simple organisms, since there isn't enough data to fully model human cells yet. "We are at the very beginning, and we have an enormous amount to do," he says.

Visual models like the software behind the slides are central to systems biology. To make sense of the genetic networks they're studying, Hood's scientists have to see them--or rather, a metaphor for them. By way of illustration, institute co-founder Aderem points out the window to the Seattle skyline. "If the lake, the city, the cranes you see there were mathematical equations," he notes, "you could never make sense of it. There are simply too many variables." Visualization software will become even more important once new instruments begin, as Hood envisions, to record everything happening to every gene and protein in a cell.

No such measuring devices and no such software have been invented yet, of course, but that doesn't seem to bother Hood. As a young biologist at Caltech in the 1970s, Hood was frustrated by the limitations of the instruments of the time. So he divided his lab in half. One part was dedicated to pursuing cutting-edge biology, the other to creating cutting-edge technology to drive that biology forward. One of the machines that emerged from his lab--the automated DNA sequencer--made the Human Genome Project possible. It also became the backbone of Applied Biosystems, which licensed the patents to his instruments and built itself into the leader in its industry.

Still, no scientific invention in Hood's history is quite as ambitious as his latest project: the nanolab. This device would squeeze the equivalent of 1,000 laboratories onto a microchip. The idea may sound like science fiction, but Hood has already signed on a trio of the world's top nanotechnologists from Caltech, as well as Phelps from UCLA, to help bring the nanolab into being. They've designed a system of microscopic pipeworks to shuttle cells and chemicals around between hundreds of thousands of nanosensors. (See "The Biology Department on a Chip," page 100.)

If all goes according to plan, nanolabs will begin to produce oceans of data in three to five years. Systems biologists will then finally have enough information to create the ultimate visualization of the ever-changing workings of cells. Imagine the static diagram of yeast cells evolving into an animation of gene and protein interactions. "The output of the nanolab will be a videogame called Let's Play Biology," Phelps says, only half joking. In a constant feedback loop, scientists would use the software to turn genes on and off, or introduce drugs into the system, and see what happens.

How will all this change medicine? A scientist could use the nanolab to study the interplay of genes and proteins that gives rise to, say, prostate cancer. But instead of looking for a lone culprit gene, it would allow scientists to study a diseased cell as a whole. Once the system is understood, doctors could compare a patient's cells with the model and arrive at an early diagnosis or even a predictive one.

Then the nanolab could help guide the creation of a treatment. Through a series of software-controlled experiments, the nanolab would be able to test drugs on infected cells and, if a drug works, to create a blueprint to perform the same tests in animals and people (no model, after all, is perfect). Thus, in the future, a 40-year-old man's genes might be tested in the nanolab, and if a predisposition to prostate cancer were found, doctors could prescribe a drug that shuts off the aberrant system.

Before systems biology can deliver on any of these promises, however, a great many obstacles must be overcome, not all of them scientific. One is the huge management task of coordinating work in so many unrelated fields. James Heath, one of the Caltech trio working on the nanolab, recalls utter confusion at his first meeting in Hood's institute. Heath didn't know what gene expression was, and the biologists knew little about microelectronics. "Sometimes you talk to people for hours and all you come away with is a noun," he says.

And then, of course, there's the sheer volume of data to be generated. There won't be a fully functioning nanolab before 2006, at the earliest. If it works as promised, computers may not be powerful enough to make sense of all the data. "Some people feel it will be too complicated for mathematics," Hood acknowledges.

So why bother? Hood's answer is one that scientists have given for centuries: We won't know the task is truly impossible unless we try. And it's important to remember that Hood has been here before. The nanolab may not exist, but neither did the automated DNA sequencer in 1985 when Hood and other scientists first met to discuss the Human Genome Project.

Besides, Hood's systems biology experiment can create new technologies and real wealth even if it never succeeds at completely analyzing the trillions of possible interactions in genetic networks. Hood's lab should spark advances in nanotechnology, drug discovery, and other fields with commercial applications. The nanolab alone could conceivably spawn another biotools company like Applied Biosystems. The institute's proteomics expertise and ongoing research into diseases such as cancer and autoimmune disorders could lead to lucrative drug-licensing deals. In fact, Hood sees spinoffs as another way to "transfer knowledge to society." (And proceeds from a home-run discovery would give the institute a much-needed endowment.)

In any event, Hood isn't in this for the money. The Human Genome Project took the first step toward understanding how life works at the genetic level; a scientist like Hood can't avoid the next step, however improbably ambitious. "Life is a process of evolution," Hood says, "and anyone who thinks the current world order is OK does not get what evolution is all about." In other words, the urge to find out is irresistible. It's built into our system.

Erick Schonfeld (eschonfeld@business2. com) is editor-at-large for Business 2.0.