(FORTUNE Magazine) – BEHIND the red brick walls of two unprepossessing buildings in a science park in Rockville, Maryland, l35 scientists and entrepreneurs are laying the groundwork for a new epoch in biology and medicine. Computer-assisted robots, the galley slaves of the 21st century, work around the clock in a spotless, brightly lit bay, doing the researchers' bidding. The object: nothing less than to decipher and commercialize chemical sequences that make up human genes, the molecular arbiters of health, intelligence, and behavior. The researchers, robots, and computers are all part of Human Genome Sciences (HGS), the largest and the most lavishly financed of about a dozen new companies racing to crack the human genetic code. Some analysts think HGS is so far ahead of the field that it may be the Microsoft of genetics in the making. William A. Haseltine, who took a leave from a professorship at the Harvard Medical School to serve as CEO, boasts that HGS and the research foundation it supports, The Institute for Genomic Research, will have isolated and partly deciphered most of the important human genes within two years. Developing medical applications could take years more, but simply having the knowledge will give HGS what Haseltine calls the ''gotcha'' advantage. HGS expects to profit mightily by licensing the secrets it uncovers to drug companies and exploiting them itself. It should be able to patent many of those discoveries. So far, the most visible manifestation of the gotcha advantage is a chauffeur-driven limousine, parked in front of HGS headquarters, at Haseltine's beck and call. The company turned its first profit last year, $1.8 million on $22 million in revenues, and went public in the fall at $12 a share; at a price of $16.75 a share in early May, Haseltine's 8% stake was worth about $20 million. The public offering succeeded in part because HGS has the backing of drug giant SmithKline Beecham, which committed $125 million a year ago for product rights and a 7% stake. HGS's 1993 revenues consisted entirely of payments from that arrangement. Seldom has a major corporation put up so much money at such an early stage for a technology being developed by someone else. SmithKline Beecham isn't alone. In recent months, platoons of scouts from big drug manufacturers were seen sliding on Southern California mud and slipping on Boston-area ice as they sought out startups with names like Millennium Pharmaceuticals (Cambridge, Massachusetts) and Sequana Therapeutics (La Jolla, California). Among the most hotly pursued: Myriad Genetics (Salt Lake City), co-founded by Walter Gilbert, the Harvard Nobelist; and Darwin Molecular (Bothwell, Washington), co-founded by Leroy Hood, who invented a gene-decoding machine as central to the new industry as the cotton gin was to + textiles. In March, Hoffmann-La Roche agreed to put more than $70 million into Millennium for the right to turn genetic data about obesity and adult-onset diabetes into pills. Venture capitalists say more such deals are imminent. The big boys have also started a brisk trade in what might be called gene futures. Eli Lilly paid Myriad Genetics $2.8 million for rights to a gene implicated in breast cancer. Genentech, the $650-million-a-year biotech pioneer, invested $17 million in GenVec (Rockville, Maryland) and set up an affiliate called Genomyx. Baxter International, SmithKline Beecham, Hoffmann- La Roche, and other large companies are also rushing to build up in-house gene-hunting capability. Says Kevin Kinsella, a veteran venture capitalist and founder and CEO of Sequana Therapeutics: ''As a venture capitalist, I've started seven biotech companies since 1982, but I haven't seen anything like it. This is biotech's counterpart of the Oklahoma land rush.'' WHAT LURES the pioneers is the promise of a bonanza that could make the Sooner land rush pale by comparison. An almost biblical event -- the final decoding of the fundamental secrets of life -- will pay off, participants expect, in novel and uniquely effective medicines. With genetic codes in hand, scientists will be able to design drugs to attack the causes of disease rather than the symptoms, as most medications do now. Afflictions that have crippled and killed for millenniums -- some cancers, rheumatoid arthritis, heart disease -- will become amenable to treatment. So will such miseries as migraine headaches and obesity. Even baldness may be treatable -- Kinsella's company hopes to develop a shampoo that would deliver genes for promoting hair growth directly to the cells of the scalp. Diagnosing people genetically prone to a given illness and treating them preemptively could reduce national health costs, perhaps by a lot. Exults Haseltine: ''We are witnessing the beginning of a new age in biology and medicine.'' The trip from the lab bench to the bank window is almost sure to turn out to be more difficult and time-consuming than scientists and investors expect. But for pharmaceutical companies, holding back could mean missing a historic opportunity. Juergen Drews, president of international R&D at Hoffmann-La Roche, hints why when he says that the new exploration of genetics ''will revolutionize our approach to drug discovery.'' Scientists will avoid many of the blind alleys of conventional R&D, still largely a hit-or-miss process that involves concocting new compounds and then testing them for therapeutic effects. Instead, researchers will be able to precisely design drugs aimed at specific targets. The promise of dozens of potent new medications comes at a time when the R&D pipelines of most big companies are almost empty. In the past three years, according to the Food and Drug Administration, big companies won approval for only 81 truly new drugs, including treatments for high blood pressure, epilepsy, and infections. The rest -- 998 medications -- were variations on existing drugs. THE MAN largely responsible for this historic upheaval is J. Craig Venter, 47, director of HGS's non-profit affiliate, The Institute for Genomic Research (TIGR) in Gaithersburg, Maryland. A soft-spoken molecular biologist, Venter presides over the most advanced gene-discovery laboratory in the world. In a cavernous room, fresh-faced young people in white lab coats poke at computer keyboards and ease tiny glass tubes into the maws of big, gray gene-reading machines. The vials are filled with molecules of human DNA and four fluorescent dyes, each designed to mark one of the four basic chemicals that pair up to form rungs of the famous double helix. Lasers activate the dyes and generate vivid, stained glass-like images on computer screens. The images correspond to the sequence of so-called base pairs in the DNA being tested. Data from the machines course through thick black cables into a Maspar supercomputer, equal in power to 4,000 PCs, on the floor above. It compares the new codes with millions of genetic sequences from dozens of species. Because the genetic code is universal, a human sample may match DNA from another human, a rat, a bat, a mouse, a worm, a fruit fly, or even a microbe. The human genome -- the entire DNA code for a human being -- differs from that of a chimpanzee in only about 1.5% of its content; chimps, fruit flies, and other creatures make proteins identical to or similar to those of humans. If the new sequence matches a known one, which currently happens about half the time, researchers can infer its genetic function; otherwise they carefully catalogue the sequence for further investigation. Venter's procedures help scientist-entrepreneurs sidestep the biggest obstacle in biogenetics: the sheer vastness of the genome. The human body consists of more than 75 trillion cells, each of which, except for red blood cells, has a full complement of chromosomes nestled in its nucleus. Each chromosome is a wadded-up strand of DNA made of hundreds of millions of base pairs; stretched out straight, it would measure anywhere from three to nine feet long and about 20 atoms across. The chromosomes -- 46 to a cell, in 23 pairs -- constitute a complete set of instructions for the making and functioning of a human being. Those instructions take the form of genes, submicroscopic chemical sequences scattered along the chromosomes. Magic words in the book of life, genes direct the production of proteins that make up cells' structure and run their vital chemistry. Of all the DNA in the genome, genes account for scarcely 3%; the other 97%, sometimes called junk DNA, is now thought to serve structural and other purposes. The Human Genome Project, a $3 billion international effort to map the entire genome that was launched in 1990, involves 350 labs and isn't expected to complete its task until 2005. Since genes are distributed randomly, and sometimes in pieces, among the six billion base pairs in each set of chromosomes, isolating them represents one of biology's great challenges. Working in the 1980s at the National Institutes of Health, Venter brilliantly combined three insights. He found a way to harness living cells to isolate genes. ''A cell can do much more than any supercomputer,'' he explains. ''It knows how to extract the information it needs from the chromosomes.'' Inside the cell, a remarkable mechanism transcribes DNA into a concise blueprint for a protein, called messenger RNA, by editing out all the junk DNA. The cell works much like an insomniac movie buff taping a film from late-night TV and zapping the interminable commercials. Venter decided to fish out the fragile messenger RNA molecules from cells, and by applying special enzymes, to transcribe them back into sturdy DNA that represented pure, edited genes. This so-called complementary DNA could be analyzed and held for future reference in vials in a freezer. The discovery helped open the new frontier. Government agencies, universities, and private institutions have used Venter's method to build entire biological libraries. They keep samples of healthy and defective genes and make the deciphered codes available via Internet to researchers all over the world. Scientists use the data to investigate genetic defects, even the tiniest of which can cause devastating illness. For example, sickle cell anemia, the crippling blood disorder, results from a single error in a single gene. Other diseases, such as asthma and diabetes, are polygenic -- associated with multiple genetic defects, often on more than one chromosome, interacting in ways not yet understood. Venter was among the first to see that by marrying computers and biotechnical instruments, scientists could speed the search for the genetic causes of disease. He championed the use of robots to perform routine experiments in the lab; by hooking together machines that read genes and computers to process the resulting data, he was able to streamline the isolating and decoding process. The field he helped pioneer is now known as bioinformatics. His third, most controversial, breakthrough accelerated the gene hunt by as much as a thousandfold. Rather than struggle to piece together entire genes, he decided to isolate and investigate gene fragments. Some scientists scoffed, arguing that analyzing fragments, especially those of genes whose function was not known, would be meaningless. But Venter was betting that the fragments would yield valuable clues to the meaning of other genetic data accumulating in computerized libraries, just as a scrap of a secret message can be used to crack a much longer message written in the same code. Soon Venter was decoding DNA fragments by the thousand. His work exploded onto the commercial scene in 1991, when NIH attempted to patent the fragments as a public service. Bernadine Healy, the NIH director, intended to encourage commercial applications of Venter's research by licensing the data at nominal cost to private companies. She argued that companies would not be interested in developing gene-based medications unless they could be assured of patent protection. But James Watson, the legendary co-discoverer of the double helix and a director of the Human Genome Project at NIH, blasted the plan, arguing that patenting the secrets of life would hinder research. Venter soon found himself in bureaucratic Siberia. When he applied to the Genome Project for funding to expand his lab, he was rejected twice; prominent university geneticists, such as David Botstein of Stanford, criticized his work as narrow and shortsighted. The hubbub set off sensitive seismographs in the offices of biotech venture capitalists. Before long, Venter struck an innovative deal with Wallace H. Steinberg, a former Johnson & Johnson executive who heads HealthCare Management Investment Corp., a venture fund in Edison, New Jersey. They / founded Venter's institute and HGS in 1992. In exchange for exclusive rights to the institute's discoveries, HGS is obliged to pay $85 million over a ten- year period. Venter and his 70 researchers are free to publish their findings after HGS and SmithKline Beecham have mulled them over for at least three months. Though he wound up with a stake in HGS that is now worth $12 million, Venter says he wasn't looking for money. What makes him happy is the freedom to pursue whichever genetic research he wants. Says Venter: ''Now I'm the bureaucracy.'' Venter's research was vindicated in March when a team from the Johns Hopkins Oncology Center in Baltimore used his institute and HGS to help pinpoint genes on Chromosome 3 that play a role in cancer of the colon. HGS, which has invested heavily in finding and cataloguing defective genes, had already singled out those in question. The colon cancer genes are particularly important because they hint at a mechanism that may be at work in many forms of cancer. In their healthy state they help direct the repair of damaged DNA within the body's cells. When they themselves are damaged, they no longer function. The defective DNA accumulates, and cancer soon starts to grow. HGS and TIGR are racing to refine their gene-hunting technology and expand their competitive edge. But the work Venter did at NIH remains in the public domain, and many other genomics companies are using it too. HGS's biggest competitor, Incyte Pharmaceuticals in Palo Alto, is compiling data on genetic differences between healthy and diseased immune-system cells. The information could help companies design drugs against allergies, asthma, rheumatoid arthritis, and other immune-related disorders. Incyte's strategy, says director of business development Lisa Peterson, is to market its data to major drug companies: ''Since HGS is largely tied up with SmithKline Beecham, we plan to serve the rest of the industry.'' Like HGS, Incyte went public last fall. Other genomics companies are betting on an approach that is the reverse of Venter's. Rather than match gene fragments with diseases, they are working backward, trying to trace inherited disease to its genetic cause. Millennium, for example, has tackled three tough polygenic disorders: obesity, diabetes, and asthma. Myriad Genetics is probing the causes of heart disease and cancer by analyzing the vast records that the Mormons have kept of generations of faithful in Utah. Sequana hopes to solve the mystery of baldness by comparing human genes with those of hairless mice; it is trying to isolate obesity genes by studying mice that naturally get fat. Such quests will go on for many years. But the knowledge genomics researchers are accumulating may pay off in a surprising variety of ways, some quite immediate:

-- DIAGNOSTICS. Detecting disease is the arena in which many genomic startups are looking for their first commercial success. Analysts see a $7-billion-a- year industry taking shape in the next few years. One leader, Genica Pharmaceuticals of Worcester, Massachusetts, sold $7.2 million of tests for neurological disorders last year. With nearly 30 diagnostic tests on the market or under development, the company claims to have commercialized more discoveries stemming from the Human Genome Project than anyone else.

The promise of genetic diagnostics lies in the certainty with which such tests can detect disease. Matritech, a publicly traded company in Cambridge, Massachusetts, for instance, is developing a foolproof test for colon cancer. It checks for six proteins that turn abnormal only when cancer is present. Genetic tests may also help doctors gauge the severity of a disease, which can vary depending on which part of a gene is defective. For example, in cystic fibrosis, a congenital disorder affecting the lining of the lungs and pancreas, the gene in question on Chromosome 7 can be damaged at more than 300 sites. The severity of the disease depends on the locations of the flaws. Tests that pinpoint some of them are now being sold by Integrated Genetics of Framingham, Massachusetts. The ultimate diagnostic tool may be the laboratory on a chip: tiny devices that combine a chemical process and electronic sensors to test for a specific disease. Such a product is under development at Lawrence Livermore National Laboratory in California. Support for the project comes from Roche Molecular Systems, a Hoffmann-La Roche subsidiary in Alameda. The chip is a square of silicon less than a half-inch across; it incorporates a reaction chamber in which DNA fragments are heated and exposed to reagents to reveal the genetic code. The hope is to use the chip in a hand-held, battery-powered kit for diagnosing diseases in hospitalized patients. Meanwhile, Affymatrix in Santa Clara, California, is working on a dime-size lab-on-a-chip. It is designed to compare a DNA specimen squirted onto its surface with embedded test samples of DNA. Defects show up as mismatches, which chemicals on the chip cause to fluoresce. A laser device scans the chip and displays the result on a computer screen.

-- GENE THERAPY. Diseases such as muscular dystrophy, hemophilia, and some forms of heart disease and cancer are characterized by the body's inability to manufacture a vital protein. Soon doctors may be able to administer new genes that restore protein production. Early attempts at such treatments have worked. W. French Anderson, a former NIH researcher now at the University of Southern California, has healed a small number of children with rare genetic diseases by injecting them with undamaged genes. Preliminary trials on patients with cancer, cystic fibrosis, and other diseases are already under way. The question of how best to administer replacement genes is one of the most intriguing puzzles in medicine. Ideally, defective genes should be detected and replaced in childhood or even in the womb, before they cause disease or get passed along to future generations. But technology is nowhere near that stage, and researchers are pressing ahead with the tools they have. To treat cystic fibrosis, GenVec of Rockville, Maryland, is testing a liquid it hopes to use in spray form to deliver healthy DNA to the lungs. In March, researchers at the University of Pennsylvania Medical School reported that they had introduced healthy genes surgically into the liver of a young woman from Quebec who suffers from an extremely high cholesterol level. The researchers removed tissue from her liver, introduced intact genes into the cells, and then reinserted them. Her cholesterol quickly dropped 20% as the new genes began to function; it is too early to tell whether the improvement is permanent. Genetic Therapy, of Gaithersburg, Maryland, is among the companies positioning themselves to commercialize such research.

-- REGULATING GENES. A gene that fails to function doesn't necessarily have to be replaced. Sometimes all that is defective is its on/off switch -- the sequence of DNA that starts or stops the protein production process. Replacing command sequences, which are as short as 50 base pairs and are readily synthesized, could represent an entirely new way to treat disease. Transkaryotic Therapies of Cambridge, Massachusetts, recently showed in the lab that it can insert a new switch into a cell and turn on a balky gene. In the future, expensive manufacturing plants may no longer be needed to produce proteins like erythropoeitin (EPO), which restores red blood cells in anemia; the patient's own cells may be stimulated to do the job.

-- PROTEIN THERAPY. Genome companies also hope to use their burgeoning knowledge to produce medicinal proteins in the lab, using well-established genetic engineering techniques. Haseltine notes that biotech powers like Amgen and Genentech have built sizable businesses on just a few such drugs. Example: Genentech's tPA (tissue plasminogen activator), which dissolves blood clots in heart arteries. ''We already have 200 to 300 genes for candidate proteins,'' says Haseltine. Because HGS and other genome startups are allied with major drug manufacturers, new proteins could find their way to market within the next few years.

-- SMALL MOLECULE DRUGS. Test tube chemistry can be harnessed to make small molecules for the treatment of many diseases including AIDS and some cancers. Unlike proteins, which are effective only when injected directly into the bloodstream, small molecules pass easily through the lining of the stomach and can be made into pills. That's where the big money is in the drug industry. Darwin Molecular of Bothwell, Washington, is staking its future on a daring strategy to convert genetic sequences directly into candidates for pills. In a novel approach called man-made molecular evolution, it uses chemistry and computers to design and test millions of molecules for their ability to find and repair disease-causing genetic defects. In a series of screening steps, the candidates are gradually weeded out; only the fittest make it through the entire process and go on to clinical tests. So promising is the method that Darwin recently attracted a $10 million investment from Microsoft billionaires Paul Allen and Bill Gates.

Like dirt farmers who became rich in the land rush by discovering oil on their property, some genomics entrepreneurs will likely hit the jackpot in genes in the next decade. Says Sequana's Kinsella with total conviction: ''The payoff of genomics will be bigger than that of the Manhattan Project or the space program.'' But there are epic risks as well: Many startups will fail as founder-researchers prove to be less adept at management than at science, and giant companies that move too slowly may see their markets eaten away. To fully exploit the promise of genomics will require effort on a scale that is hard to comprehend. James Watson of double-helix fame predicts that it will | take geneticists 10,000 years to fully understand the workings of the genome. Still lacking, for example, is an explanation for how genes dictate the three- dimensional shape of proteins. Within cells, proteins act like biological micromachines. Some form valves that let nutrients enter through cell membranes; others serve as catalysts in chemical reactions. Yet no scientific language today can comprehensively relate this 3-D reality to two-dimensional codes of DNA. That is only one of the untamed frontiers that still confront genomics pioneers.