Joshua LaBaer, director of the Harvard Institute of Proteomics (HIP) at Harvard Medical School, is a bit like that little old lady who used to ask, "Where's the beef?" in television commercials. He keeps asking about all the genes mapped by the Human Genome Project, but the question is rhetorical. He knows no one has them.

"The outcome of the Human Genome Project is sequence information in computers, and that's it," says LaBaer. "One of the mistakes that some people make . . . is they think we have the genes already in hand. That's not true. It's all in silico. It's all terabytes of information."

For LaBaer, the logical next step is a $100 million-or-so project called the FLEXgene Consortium - FLEX being an acronym for full-length expression. If LaBaer has his way, public and private research groups around the world will clone all human genes - 35,000, more or less - and put the physical clones into a public repository open to any scientist who wants to use them.

So far, response has been enthusiastic, according to LaBaer, with representatives of the National Institutes of Health, the Wellcome Trust, the Ludwig Institute for Cancer Research, and about 20 private companies participating in the planning. "Everybody needs the clones. That's the fundamental thing," he says, describing the shortage of genes for research as a bottleneck that needs to be cleared before many promising projects, including those at his institute, can go very far.

A self-proclaimed "protein chauvinist," LaBaer might seem an unlikely champion for a gene project of this scale. Though he holds a Ph.D. from the University of California at San Francisco, where he studied with Keith Yamamoto, he is a practicing oncologist, who says gene therapy "ain't happening yet." His prime research interest is using high-throughput technology to purify and study hundreds, even thousands, of proteins in a short period of time. "It's the proteins that we're made of," he says. "It's the proteins that do the business of biology. It's the proteins that go bad in disease, and it's the proteins that are the target of all pharmaceuticals."

So why does he care about genes? Because genes make proteins.

In the twentieth century, scientists could devote their careers to the study of a single protein. They would spend years cloning the parent gene, purifying the protein, and studying it in all sorts of formats. That's too slow for proteomics advocates, who hold that finding the functions of all 150,000 or so proteins is the key to finding new therapies. They have the tools to proceed faster, LaBaer explains, but they need cDNA to make the genes and then the proteins on a large scale. What's more, he adds, they need to be able to move those genes quickly from bacteria to mammalian cells or to any other vector in which they want to experiment.

Edward Harlow, chair of Harvard's department of biological chemistry and molecular pharmacology, picked LaBaer to head the proteomics institute in 1999. Using a new technique called recombinational cloning, the new institute began building a clone collection for storage in a repository initially called "the Wall." The name has since been dropped as sounding too obstructionist, and the project went into a holding pattern when it became clear that the task was too big for one wall of test tubes in one Harvard laboratory.

Encouragingly, the first problem was the strong positive response from the private sector. Not only were companies willing to contribute money, but they also wanted to talk to LaBaer on a regular basis. He didn't have time to manage the project and interest in the project - and do more fundraising as well.

Advisors recommended a consortium modeled on the SNP Consortium, the group of public institutions and private companies that banded together to identify the common DNA base-pair variations known as single-nucleotide polymorphisms (SNPs). Arthur L. Holden, chair of the SNP Consortium, agreed to head the FLEXgene organizing committee.

He says the group is "solidly progressing through a number of issues," including the final budget and size of the project, with an eye to starting in the first quarter of 2002.

"To me this is a program that can be very helpful in translating genomics into applied research," says Holden, who is also chairman and CEO of First Genetic Trust ("The World's First Online Genetic Banking, Education and Advising Service") in Deerfield, Illinois. Even for private investigators, the project would be prohibitively expensive, according to Holden, who predicts that other SNPS-like consortiums will be formed to tackle common problems. "Money is a huge issue. . . . We're still at the beginning of having a full sequence of human genes," he says. "We're just at the beginning of the process."

Once the technical, legal, and administrative details are worked out, LaBaer estimates the repository could be assembled in three years with labs around the world taking on part of the load. These "details" are considerable, however: What cloning system to use? Who's going to fund the project? What criteria will be set for the quality of these clones? And so on.

Start with technology: recombinational cloning uses a polymerase chain reaction to make multiple copies of a gene and store them in a master clone format that, in turn, allows researchers to move genes into other vectors as needed. Initially, LaBaer's group used recombinational cloning technology called Gateway from Invitrogen Life Technologies in Carlsbad, California. In the meantime, a second company, Clontech of Palo Alto, California, has developed another system called Creator, and LaBaer expects that there may be others. Consequently, FLEXgene's organizers will draw up specifications in a request for proposals, so that everyone gets a fair chance to bid on the project.

Whichever system is chosen, automation will be an important part of the process. The work so far has shown that although robots can relieve technicians of a lot of menial work - such as loading gels - they demand meticulous programming. "Robots are very good at doing repetitive tasks and doing them one after the other without losing their places," LaBaer says, but he adds from experience: "If a robot is taught to do something a little bit wrong, it will do that thing wrong over and over and over again, and then your whole thing is screwed up."

As of September, the Harvard group had completed a first pass at the genes in Saccharomyces cerevisiae to test the concept, and captured several thousand human genes. This work has turned out to be more expensive than anticipated, however, so determining who will pay for the project has moved to the foreground. Despite the huge cost and the multitude of interests, both public and commercial, however, LaBaer does not anticipate a problem.

At an August 2001 conference on drug discovery technology, for example, he sat on a media briefing panel with the heads of two companies that are active in proteomics: UK-based Oxford GlycoSciences and Paris-based Hybrigenics. Both described the work ahead as so vast that it will require a multitude of players to get done. Hybrigenics CEO Donny Strosberg also praised Harvard's plan to resequence every gene in the repository, so that all scientists who use it will be working on genes that meet the same standards. "Quality control is not associated with biotechnology," he noted.

The panel discussion revealed a potential obstacle, however. Strosberg and Oxford GlycoSciences CEO Michael Kranda both said that they are aggressively patenting protein discoveries that show promise in development of new drugs. Moreover, patent attorney Scott A. Chambers, an expert on intellectual property rights in biotechnology, suggested that pharmaceutical companies wouldn't use biological discoveries that are not covered by patents. "If it costs 200 to 500 million dollars to bring a drug to market, unless you have the intellectual property position that you know you can protect it, you're not going to put that kind of money into it," he said.

Holden acknowledges an element of risk, and says the consortium can request licensing rights if patent issues arise. The possibility that patent disputes could keep researchers from working on a gene or protein only makes LaBaer more determined that the FLEXgene clones be a public resource, without encumbrances. "If this gets trapped behind a commercial wall, then it's going to slow research, not expedite it," he says, challenging the private sector to save the competition for finding clever ways to exploit the clones.

That's what he would like to do. After more than two years of campaigning for FLEXgene, LaBaer does not see himself leading the consortium. Rather, he proposes that the CEO be someone with experience in running large projects, and that the consortium be housed at many centers, not just at Harvard. "My goal is not to run that sort of project anyway," he says. "My goal is to eventually use these clones to do real science."