(FORTUNE Magazine) – Loss threatens young biotech companies in more forms than any other kind of business. Investors can lose millions when a promising drug fails to work or funds run out before testing is complete. Researchers can lose prestige and precious years by chasing up blind scientific alleys. Doctors can lose patients--and people clinging to the hope of a cure can lose their lives.

Those daunting risks did not stop Michael Blaese, the widely respected chief of gene therapy for the Human Genome Project unit at the National Institutes of Health, from moving to biotech startup Kimeragen full-time early last year. His decision stunned colleagues in the arcane, intense world of gene-therapy experimentation. A small, privately held firm near Philadelphia, Kimeragen was a scientific maverick. It had staked everything on an unorthodox technique that differs significantly from most gene therapy. The conventional approach tries to replace faulty genes with new, healthy versions. Kimeragen instead aims to fix addled genes, using a patented form of molecular microsurgery that corrects tiny misspellings in the DNA codebook, like a sort of genetic White Out.

This technique, dubbed gene repair, was far from proven. Many scientists questioned the published papers showing that it worked--and wondered why, if it was so promising, more researchers weren't embracing it. Suggestions of data doctoring in the early career of gene repair's inventor only heightened their doubts.

What's more, Kimeragen was in turmoil. The company had forced out its chairman in 1998, and the post was still vacant. Half a dozen employees, including senior scientists and administrators, had been fired. Money was running low. Since its founding four years earlier, Kimeragen had raised just $22 million--a pittance with which to try to develop a new medical technology and bring it to market. And while Kimeragen's press releases promised the sky, progress toward a first human trial of gene repair was going painfully slowly.

To Blaese, none of this mattered. The 60-year-old pediatrician saw in gene repair a chance to cure dozens of genetic diseases that he had spent his career battling. "I wanted to make [gene therapy] something that I could actually treat my kids with," he says. "This is the only technology I've seen that has the opportunity of actually providing a successful gene therapy to the vast bulk of genetic disease." For some researchers, Blaese's move cast Kimeragen in a new light. Says William Haseltine, CEO of Human Genome Sciences, an industry-leading gene-hunting firm in Rockville, Md.: "Aside from the technology, that's the most interesting thing about this company--that someone with his knowledge, his background, would make this bet."

Nine months after Blaese landed at Kimeragen, a tragedy rocked the broader gene-therapy world. Last September, Jesse Gelsinger, an 18-year-old research volunteer, died in a gene-therapy experiment at the University of Pennsylvania. His was the first death known to result directly from gene therapy. Congress demanded explanations. The NIH and the Food and Drug Administration promised to toughen their oversight. And though Kimeragen tried hard to distinguish its repair technique from the conventional replacement one that killed Gelsinger, his death badly hurt its chances of survival as an independent company.

This is a story with no winners--at least not so far. But some big names have begun betting that Kimeragen's technology will beat the odds. Last May, G. Steven Burrill, a wily San Francisco venture capitalist with long biotech industry experience, bought 28% of the company and became chairman of the board. This April he merged Kimeragen with ValiGene, a genomics company in Paris. Its chief scientific adviser is J. Craig Venter, the genome- sequencing guru who is leading Celera Genomics in its historic quest to sequence the entire human genome. Venter will stay on as chairman of the scientific board of the merged company, now called ValiGen.

Kimeragen bills the merger as a triumph that will enable it to reach its goals much more quickly. The company will also put gene repair to work in a potentially lucrative new area--the hunt for drug targets, which was ValiGene's focus. But there's little doubt that the merger was a marriage of necessity for a 37-employee company that had perhaps a year of cash left when it started looking for suitors.

Kimeragen's story shows how ferociously difficult it can be for a biotech startup to do ambitious, unproven science. Forces that should be controllable, like executive egos, can sabotage a company's best interests. Forces beyond the company's control, like those buffeting gene therapy, can deal a wicked blow at the worst moment. And as is true with Kimeragen, the lives of very sick people may hang in the balance.

Derick and Amy Martin, ages 10 and 8, have never slept under covers, worn pajamas, or stayed overnight at friends' houses. The children live in a big, old, tan clapboard farmhouse in the rolling farm country of central Pennsylvania. They attend a one-room schoolhouse nearby. After school they help their father sweep the barn and feed the family's 30 dairy cows. Their parents, Floyd and Katie Martin, are Mennonites who have known each other for most of their lives.

In 1990, when Derick was a days-old baby, his parents noticed that he was extremely yellow--it was a jaundice that, unlike typical newborn jaundice, didn't quickly disappear. That evoked a dismaying memory. Eight years earlier a nephew of Floyd's had had the same condition. He died when he was 3.

Derick, the Martins learned, has an inherited disease called Crigler-Najjar syndrome. (Amy was diagnosed with the same thing when she was born.) Known by those who live with it as Crigler's disease, it is extremely rare. But it occurs often among Amish and Mennonites of the Midwest because the members of their small population aren't allowed to marry outside their religion. This causes genetic errors to propagate, turning up in children whose parents both carry the gene for a given disease.

The children's disease is caused by a single-letter misspelling in a gene that makes an important liver enzyme. In children with Crigler's, this enzyme is missing or scant. Without it the liver can't process bilirubin, a byproduct of the body's constant breakdown of old red blood cells. When bilirubin levels get too high, the substance is toxic to the brain, causing damage much like a stroke. Bilirubin levels rise during illness or fasting, or after injury, making life for Crigler parents an anxious exercise in constant vigilance.

Derick and Amy's physician, Holmes Morton, has seen what can happen when bilirubin levels rise too high. A 7-month-old boy with a high fever was brought to a local hospital. "By the time I saw him," Morton says, "he had no head control, no control of his arms and legs. He couldn't swallow, is deaf, and has virtually no eye control." The boy, now 3, has to be fed by a tube. "He may live to be 40 years old," says Morton. Some children don't survive these episodes. Katie Martin, who is in touch with affected families all over the world, says that to her knowledge, the oldest living Crigler patient is in her mid-30s.

Other things make the disease hard to live with. Exhaustion is a frequent symptom. So is jaundice. Amy, a shy, slender girl with waist-length blonde braids, is already sensitive about her yellow skin and the yellow whites of her almond-shaped blue eyes. (Derick's glasses hide his.) Strangers approach her mother on the street, asking if she is aware that Amy's eyes are yellow and whether Amy is contagious. The bilirubin also causes gallstones; both Derick and Amy have had their gallbladders removed.

Amy and Derick also have a bad case of media fatigue. Time, 60 Minutes, and National Geographic have been among their pursuers. Photographers have spent hours at their house. When FORTUNE visited, the kids were extremely reticent. Derick answered questions mostly in mono-syllables, Amy not at all. It took a major bargaining session before they consented to be photographed for this story.

The worst thing about the disease for the kids, though, is the lights. From birth, Crigler children must lie in diapers or underwear under blue lights for eight to 14 hours a night. These lights help break down bilirubin, but the kids hate them. The lights mean you never get to wear pajamas or a nightie. They are hot in summer and draw flies. In winter it's cold. "I could sleep better under covers," says Derick, a solid boy with a cleft chin. He wakes up at 4 A.M. daily and can't get back to sleep.

As Crigler children hit adolescence, phototherapy becomes less effective. By early adulthood a liver transplant, which is fraught with risks, is their only recourse.

Through the 1990s, Dr. Morton, who had given up a promising academic career to care for the Crigler kids and others, worried about Derick, Amy, and 14 other Crigler patients. The kids seemed headed inexorably toward liver transplants. Then, in early 1998, Morton got a call from Mike Blaese. Morton knew of Blaese as a gene-therapy pioneer who had run the first human experiment in 1990, treating two girls with a rare immune disorder. (The treatment was partially effective.) Explaining that he was consulting for Kimeragen, Blaese asked if he could come talk to Morton.

On an April morning Blaese arrived at Morton's clinic, set in the middle of a cornfield on an Amish farm near Lancaster. The men talked for hours. Afterward Morton began contacting the Crigler families. He told them a doctor named Blaese was planning an experiment they might want to hear about.

Mike Blaese is thoughtful, soft-spoken, and not given to gambling, stock speculation, or self-promotion. A lumbering man who kicks around in old, paint-stained trousers that make his wife and two grown daughters groan, he likes woodworking and fishing for striped bass at his second home, on Chesapeake Bay.

For 32 years, until January 1999, he worked at the National Institutes of Health, caring for kids with incurable inborn diseases of the immune system. He amassed more than 300 research publications and an enviable professional reputation, both as a scientist and as a caring doctor. Through most of the '90s, as chief of the clinical gene-therapy program at NIH's human genome research institute, he held what he has called "the best job in the world." But as the decade passed, Blaese grew increasingly doubtful that gene therapy would ever help his small, suffering patients. During ten years of human trials with thousands of patients, it failed to achieve a clear success. The FDA still has not approved a single gene therapy.

Its technical problems were profound. Gene therapists try to deliver healthy new genes to replace misbehaving counterparts. But new genes insert at random anywhere in the body's genetic instruction book, which can throw off their regulation, causing them to turn on or off at the wrong time, or not to work at all. Delivering genes to cells is also a daunting problem. Scientists mostly use disabled viruses to ferry them. But viruses are huge in cellular terms. Blaese says using them to deliver genes is like sending enormous delivery vans down narrow streets in Italy. The viruses can also cause harsh immune reactions, like the one that killed Jesse Gelsinger.

The more Blaese wrestled with these problems, the more intractable they seemed. So when, in 1996, a former medical device industry executive and NIH hematologist named Gerald Messerschmidt came to him with an unusual pitch, he was intrigued. Messerschmidt wanted Blaese to work for Kimeragen, where he was CEO.

Kimeragen's technology had already caused a stir in gene therapy circles. Gene repair was invented in 1993 by Eric Kmiec, a molecular biologist then at Philadelphia's Thomas Jefferson University. The university patented the invention and licensed it exclusively to Kimeragen, which Kmiec helped launch in 1994.

Kmiec had adapted a naturally occurring process in which separate DNA strands sometimes spontaneously swap small, similar sections. In Kmiec's lab procedure, scientists mix cells with tens of thousands of identical short pieces of DNA carried in microscopic droplets of fat that fuse readily with cells' outer membranes. Each DNA piece contains the correct spelling for a misspelled, disease-causing gene. These DNA snippets--guided by sticky flanks of a cousin substance, RNA--home in on the misspelling in the DNA coiled in each cell's nucleus. They trick the cell's DNA-repair machinery into replacing the misspelled letter with the snippet's correct one. The snippets themselves degrade within 48 hours.

Messerschmidt laid out for Blaese the technique's enormous possibilities. Though the repair molecules were hard to make, early studies suggested that gene repair was going to work in every DNA-carrying species, from plants and bacteria to animals and humans. Blaese shouldn't think gene repair would be confined to rare genetic diseases of limited business potential, Messerschmidt added. It was possible that down the road, tweaking genes could help everything from Alzheimer's to heart disease. Of course, he added, there remained a huge gulf between the lab and the bedside. Making repair molecules in great enough quantity and sufficient purity to put into humans was a daunting task. But that was where Blaese's experience in both the lab and the clinic could really help the company.

Blaese recalls that as he listened, he felt a rush of excitement. Gene repair dodged a lot of the problems of conventional gene therapy. The repair molecules were tiny, and delivering them didn't require viruses. What's more, the corrected genes remained in their original places in the genetic instruction book--the right spot for the genome's on and off switches to play their roles. If the process could be harnessed for human beings, it could work a sort of typographical magic, allowing repair of dozens of genetic diseases in his young patients. "This can make a very significant difference in people's lives," says Blaese. "You don't have that chance many times in your life."

The decision to quit his high-profile government job didn't come right away, or easily, for Blaese. As he consulted for Kimeragen during 1997, some troubling things about the company emerged. Kmiec, the inventor of gene repair, had a worrisome history. In the mid-1980s, while he was a post-doctoral student at the University of Rochester, Kmiec published four papers on another subject that reported exciting and unexpected findings. As is typical in science, his mentor, Abraham Worcel, was listed as a co-author, although Kmiec actually performed the experiments. When other scientists failed to reproduce the results and suspicions regarding their validity arose, Worcel at first defended Kmiec. But in 1988 he retracted the findings in a letter to the journal Cell--a crushing admission for a highly acclaimed scientist. (Fifteen months later, Worcel, 51, jumped to his death from an eight-story parking garage.)

Kmiec says he addressed the reasons for the unreproducible findings in four papers that have since been published in scientific journals. Even so, in 1996, when he published the first paper about gene repair to draw a lot of attention, his past made scientists who knew him especially skeptical. Writing in the journal Science, Kmiec reported using the technique to correct the single-letter DNA misspelling that causes sickle-cell anemia, in test-tube samples of cells from people with the disease. Exciting as that sounded, the paper raised eyebrows because its data suggested that the genetic flaw had been corrected in as many as 11% of the cells--a success rate high enough that, if it could be achieved in patients, promised big benefits. In experiments with an older, related method of DNA repair, scientists had been able to change genes in just one in 1,000 cells.

Skeptics assailed Kmiec's work. They suggested that the "repaired" cells might have been healthy ones that contaminated the experimental mix. They criticized Kmiec for failing to demonstrate that the repair remained stable for succeeding generations of cells. They suggested that simply by bombarding cells with so many repair molecules, he could have corrected genes just by chance. "If you have 1,000 monkeys typing, eventually one will type the Bible," says Mario Capecchi, a DNA repair expert at the University of Utah who wrote a letter to Science questioning the experiment's methods. In principle, there is nothing wrong with correcting genes by chance. But Capecchi is concerned that the bombardment could upset the cell's DNA-repair machinery and cause as many mutations as it corrects.

Blaese started his consulting work with Kimeragen in 1997, in the midst of this controversy. Kmiec has never wavered from his position that time will prove his discovery. "Any quantum leap is judged harshly because these leaps occur only rarely," he says. But as months passed, scientists stood up at conferences and declared that they had tried to apply gene repair in other experiments and failed. Of the handful of labs that succeeded, the majority had links to Kimeragen.

The most noteworthy--and controversial--research was by Cliff Steer, a liver specialist at the University of Minnesota. Blaese knew Steer, a former NIH researcher, as a very careful investigator. But Kimeragen sponsored the work, and Steer's brother, Randy, is a founding board member and a six-figure investor in Kimeragen. In a 1998 paper published in Nature Medicine, Steer reported using gene repair in reverse to cause a mutation in lab rats. The change gave them a disease much like human hemophilia.

Steer reported that he had induced this mutation in 40% of the rats' liver cells--a number that amazed unaffiliated experts. "I don't think there's skepticism that [gene repair] could happen. The skepticism arises from how well it seems to work in some people's hands," says Tom Rando, a Stanford University neurologist. Rando recently reported repairing 1% to 2% of treated muscle cells in mice with muscular dystrophy using repair molecules supplied by Kimeragen.

Other scientists criticized Steer for failing to do a direct DNA readout, a way of proving that the genetic change had been made in the cells. The indirect method he used, they argued, allowed too much room for error.

Blaese says Steer's numbers "astounded" him. Yet he thought if they were mistaken, it wasn't because of a shady motive. "Cliff is one of the most honest people I've ever met," he says. And Blaese soon found other studies that helped convince him that gene repair was real. He visited company-affiliated scientists at Cornell--including Charles Arntzen, one of the country's most respected plant molecular biologists--who were trying the technique in plants. Their project was clearly working. (Last July, they published their results in Proceedings of the National Academy of Sciences, showing that gene repair could confer herbicide resistance in tobacco.)

In another experiment, a scientist from Kmiec's original lab team at Thomas Jefferson University reported using gene repair to produce black pigment in skin cells from albino mice. By 1998, Blaese says, "there was no question in my mind that the phenomenon was true. It then became a matter of degree and application. Could we take it and develop it into a useful therapeutic?" Eager to try, he decided to use accumulated leave time he had amassed at NIH and work at Kimeragen nearly full-time.

The January 1998 press release in which Kimeragen trumpeted Blaese's arrival had all the hopefulness and hype you'd expect from a startup struggling to survive long enough to get products to the market. Kimeragen boasted that it would attack genetic diseases, viral diseases, cardiovascular diseases, and others. Human trials in genetic diseases, it declared, would be under way by that fall. Trials in acquired diseases like heart disease and leukemia would follow in 1999 and 2000.

Nothing in Blaese's long and successful career had prepared him for the stresses of life at a biotech startup. In his tenth-floor office at the NIH clinical center, he had been in a doctor-scientist's mecca. His funding was secure--congressional backing for NIH had been growing every year--and there was a constant stream of patients to inspire him.

The contrast at Kimeragen was striking. The company was housed in a one-story, 20,000-square-foot building in a nondescript industrial park in Newtown, near Philadelphia. Messerschmidt, the intense CEO, liked to boast that he ran a "frugal but focused" outfit that bought only used lab equipment and geared everything toward an eventual product.

Frugal was an understatement; Kimeragen was perilously cash starved. In the three years since its founding the company had raised $9 million from about 200 private investors and one venture-capital firm, Oracle Strategic Partners in New York. Leonard Shaykin, a veteran venture capitalist and LBO investor, was chairman of the board. The company had also collected $4.5 million from agricultural sublicensees that hoped to put gene therapy to work modifying plants: Pioneer Hi-Bred International and Hoechst Schering AgrEvo.

By almost any measure, these were terribly small sums with which to try to launch an unproven biotechnology. Haseltine of Human Genome Sciences calls Kimeragen's financing "spectacularly unsuccessful." "It's enough for about six months of a reasonable operation," he says, adding that he would employ at least 200 people for any similar endeavor, more than five times as many as worked at Kimeragen. Worse, the hardening environment that surrounded biotech in the late '90s made Kimeragen's chances of substantially increasing its funding slim.

For Blaese, the culture shock and the financial worries didn't take the luster off a project that seemed to get more exciting by the week. Further experiments by Cliff Steer were showing promising results. Kimeragen's young, khaki-clad scientists were making rapid progress on designing a repair molecule for a human trial. After his visit to Holmes Morton, the Crigler children's doctor, Blaese arranged to meet the families and make his pitch firsthand. He and Morton scheduled the meeting with parents for Aug. 7, 1998.

Despite its rarity, Crigler's was a compelling choice for Kimeragen's first human trial. Not all genetic diseases make good targets for gene repair, which in lab studies has been efficient at fixing only single-letter misspellings. Cystic fibrosis, one of the most common inherited disorders, is usually the result of a three-letter misspelling. In others disorders, long sequences of letters are missing or reversed.

Still, there are other, larger targets besides Crigler's. Sickle-cell disease, which afflicts more than 50,000 Americans, seems tailor-made for gene repair: The same one-letter misspelling causes every case. And some 7,000 of the country's 25,000 hemophiliacs are thought to carry a single-letter misspelling.

But Kimeragen chose Crigler's disease in part because, despite its rarity, an unlikely group of 16 affected Amish and Mennonite children was near at hand. All had the same misspelled letter, making them perfect subjects for a single repair molecule. If the therapy worked, the results would be unambiguous: The children's bilirubin levels would fall, and their jaundice would fade. Success would dispel critics' doubts about gene repair, show that the technique would likely apply to dozens of genetic diseases involving the liver, and put Kimeragen in a prime position to go public.

The scientists' optimism made it easy to overlook how sick Kimeragen really was. But that spring, as Blaese pushed ahead with early preparations for the human trial, a boardroom crisis nearly did in Kimeragen. What actually happened is in dispute. By one account, Shaykin, the chairman, had lined up a deal in which a venture-capital syndicate would commit $10 million to Kimeragen and raise another $20 million from other VCs--all on the condition that a seasoned biotech executive replace Messerschmidt as CEO, and that Messerschmidt's allies on the board yield their seats to the VCs. By this account, the CEO angrily rejected the plan and began marshaling board and shareholder support to force Shaykin out. Rather than fight, the chairman resigned.

Messerschmidt remembers the dispute differently. He says the VCs did not make his stepping down a condition of the deal. Instead, Shaykin quit to pursue other business opportunities, causing the deal to fall apart. The VCs had a lot of concerns, Messerschmidt says, "about the stability of the company."

As 1998 went on, Messerschmidt cleaned house. He fired administrators and scientists who questioned his moves after Shaykin left, including Ramesh Kumar, the leading scientist in the company. He also severed funding to Kmiec, who had moved back to his lab at Thomas Jefferson University. (He is now at the University of Delaware.) While the blowup consolidated Messerschmidt's power, it depleted the scientific staff and hurt Kimeragen's financial prospects; the now chairmanless company was unable to find a deal to replace the one Shaykin had engineered.

Kimeragen's turmoil left Blaese with "incredibly mixed feelings" as he worked to move gene repair into a human trial. "I was so unbelievably excited by the prospects of the technology," he remembers. "But it seemed it was going to get flushed down the toilet for lack of [investor] interest." At perhaps his lowest point in that summer of 1998, Blaese tried to enlist a VC who had backed a gene-therapy company where Blaese had served as a consultant 12 years earlier. The VC turned him down flat: He said he wasn't investing in any company that didn't project a positive cash flow in 18 months.

Blaese was shocked; for the first time he began to doubt the choice he'd made. "I had decided to make a big change in my life and commit myself to something because I really believed in it. Now it seemed as though my ability to do what I really wanted to do was becoming less and less certain," he recalls. But he no longer had a choice: Even though he was technically still an NIH employee, the door was closing behind him as officials moved to find a replacement. He later told a colleague, "If this move were not forced, I would probably not join, in view of [Kimeragen's] present condition."

Even worse, Blaese faced a quandary. He was scheduled to meet with the Crigler families in August--a meeting that would raise their hopes of a cure. Yet by the end of the year, Kimeragen could be bankrupt. Even now, though it was continuing to work with its agricultural partners, it was about to cut back all its other pharmaceutical goals to focus on launching the trial.

Blaese the risk-taking scientist was suddenly at odds with Blaese the caring physician--and the risk taker won out. The meeting with the Crigler families would proceed, Blaese decided, and while he would caution them that funding for the trial was uncertain, he'd also make a strong case for what Kimeragen hoped to do for the kids. "Denial is a wonderful thing," he says in retrospect. "It was my belief--my denial--that we were going to get through this somehow. The money just had to come. So I needed to be interacting with the families."

When Blaese and the Crigler families met at Morton's clinic that summer of 1998, Katie Martin was harboring news that made her especially attentive: She was pregnant again, six years after bearing Amy. She and Floyd knew that they had a 25% chance of having another affected child, but they had finally dispensed with caution, so much did they want another baby.

At the meeting, Blaese described what the company planned in a first human experiment. Three volunteers would be treated. On six days spread over a month they would be infused for several hours with an intravenous solution of repair molecules. The molecules would be delivered in fatty sacs coated with sugar molecules that would direct them to liver cells. With each treatment the number of repair molecules would increase, until at the end each child would have received trillions.

If gene repair could be made to work in 5% to 10% of the kids' liver cells, Blaese said, they should begin to produce enough enzyme to have a real effect. The children's bilirubin levels would fall, hopefully enough to free them from the hated lights. Their jaundice should also become much milder. Best of all, a liver transplant could be deferred indefinitely.

The oldest children--Derick among them--would be considered first as study subjects.

Blaese showed the families slides of two pairs of newborn laboratory rats. One pair was healthy; the rats were pink. The other pair was afflicted with a rat version of Crigler's disease; those rats were yellow. Blaese explained that Cliff Steer, the researcher, seemed to be getting gene repair to work in the diseased rats.

Katie studied the image and thought hard about what Blaese was saying. The moment "was just remarkable," she remembers. "It was, 'That could be our kids.'"

Kimeragen limped along, helped by a $2 million bridge loan whose terms would horrify any investor: It was secured with all the company's intellectual property. But afew good things happened too: In October 1998, Kimeragen struck a research agreement with the rock stars of the biotech world, the Scottish scientists who had made history by cloning Dolly the sheep. Their company, Rosalyn Bio-Med, wanted to use gene repair to help create animals whose organs would be suit-able for human transplants. In December, Kimeragen signed a new agricultural partner that kicked in $3 million of equity.

By spring 1999, a new financier had finally appeared--San Francisco venture capitalist Steven Burrill. Burrill was relatively new to the VC world, but he was a canny industry veteran. For decades he had directed Ernst & Young's services to biotech clients, helping nascent companies like Amgen and Genentech. In May, his venture fund put up $3.5 million for a 28% stake in Kimeragen. Burrill drove a hard bargain--the deal valued Kimeragen at only one-third of what the VC syndicate had decided it was worth a year earlier. Another $4 million came in from private investors. Burrill became chairman of the board.

The infusion came just before Blaese spoke at a big June 1999 conference of Crigler families, scientists, and doctors. The Martins attended, bringing their baby daughter, Katherine. She had been born that April, unafflicted with the disease.

Blaese told the assembled families that preparations for the gene repair trial were moving ahead. Kimeragen still faced months of tough work scaling up to produce repair molecules in the quantity and purity it would need, but, he announced, the trial should be ready to launch by year-end. The Martins were giddy. "I was ready to come home and repack and go to Lancaster and do the gene therapy. We were all excited," Katie says.

In August there was more good news. Steer's new research--conducted with others including Jayanta Roy Chowdhury, a world authority on Crigler's disease at the Albert Einstein College of Medicine--was published in the Proceedings of the National Academy of Sciences. The scientists reported that they had repaired 25% of the liver cells of the diseased rats. In a vital retort to critics, they had done a direct readout of the cells' DNA, showing that the sought-after change had been made. Steer also demonstrated that bilirubin was being excreted in an altered form in the rats' bile--proof that the missing enzyme was now present.

For some critics of gene repair the paper was convincing. Mark Kay, a pediatric geneticist who runs Stanford's human gene therapy program, had been a skeptic. The new data persuaded him. "They basically dotted their i's and slashed their t's in that study," says Kay. "It's hard to argue with."

Like the other Crigler families, the Martins were counting the months until the trial was supposed to start. But when Katie picked up the phone one day last September, it was not Kimeragen calling with news of further progress but her brother. He asked if she had heard that a teenager had died in a gene-therapy experiment at Penn. Jesse Gelsinger had gone into multiple organ failure, slipped into a coma, and died four days after receiving replacement genes encased in weakened adenovirus, the common cold virus. He had had a mild form of another rare liver disorder.

"It really scared us," says Katie. "He was just like our kids."

Katie called Blaese. She knew what he would say but found it reassuring anyway. Unlike standard gene replacement, gene repair doesn't involve viruses. And Kimeragen had found no evidence to suggest that its gene-repair technique might be dangerous. Still, Katie and Floyd didn't tell Derick about Gelsinger. "We just didn't want him to have the added worry," she says.

As the Martins mulled Gelsinger's death, Kimeragen's board did the same. Directors had already been arguing that Kimeragen needed to broaden its scientific pursuits if it was to survive. Now it would be next to impossible to persuade investors to shell out for gene repair, a technology that, no matter its uniqueness, was lumped with standard gene therapy. The Gelsinger death threatened to add more regulatory delays and hurdles--not to mention public scrutiny.

Kimeragen was also running into delays in the lab. The end-of-year launch date for the trial had slipped to spring of 2000, and even that was looking dubious. Blaese and his co-workers were struggling with scale-up problems.

The ValiGene merger represented what seemed to Burrill and the board the best way to extract value from Kimeragen's research. It positioned the company to address a huge, obvious application of gene repair. Pharmaceutical companies are starving for information about what newly discovered genes do--information that can be translated into drugs. Gene repair would be a valuable tool in that hunt; because of its ability to finely tweak genes, it could help scientists investigate how small changes affect a given disease in lab cultures and lab animals. The technology meshed neatly with that of ValiGene, which had developed other sophisticated tools for tracking down genes suspected of causing particular diseases.

The companies billed the deal, completed in April, as a merger of equals. But this time Messerschmidt--who had repeatedly locked horns with Burrill--was out. The CEO job went to ValiGene's chief executive, Jean Louis Pourny; the chairman of the board and the chief of R&D are also from the French company.

Pourny says that the new application of gene repair in no way means that the Pennsylvania trial will be abandoned. ValiGen, he says, intends to proceed with it as quickly as politics and prudence allow--hopefully this year. Treating Crigler's will never yield a profit, but ValiGen views the trial as a prelude to developing a commercial version of gene repair to treat hemophilia, for which ValiGen hopes to begin human trials in 2001.

Given this, ValiGen has a big stake in carrying through with the Pennsylvania trial. As in Crigler's disease, the target cells in hemophilia are in the liver. The same molecular packaging can be used; only the spelling of the repair molecule will need to be changed. Blaese, who heads human therapy at the merged company, says ValiGen's chiefs have used "all of the right words to make me feel comfortable" that the trial is still a priority. "We'll get there," he says. "I'm absolutely convinced."

Blaese was more circumspect, though, in a recent letter to one of the waiting Crigler's families. He assured them, as he had the Martins, that Kimeragen was committed to making certain the trial would be safe. But work toward the trial was progressing, he wrote, "much too slowly for my likes.... The primary problem has been related to [producing] a high-quality drug in sufficient quantities to use in a human subject instead of just a rat.

"I feel confident," he went on, "that we will overcome this problem, but it is just too critical to all of us that we do this with patient safety as our first concern. Unfortunately, this does make the timetable somewhat uncertain.... I now hope to be ready before the end of the year, but I can't give a better guess just now."

Recently, Katie Martin moved baby Katherine to a crib where she neatly arranged a little blue quilt with a ruffle. She'd had the quilt a long time: It was a baby present for Derick that Katie had never been able to use. Amy saw the crib and came to her mother, crying.

"She said that she just is going to stop using her light; she's going to push it away from her bed. She's going to put covers on her bed, and it's going to be normal," Katie recalls. She paused. "When she does that, I talk about the gene therapy."

FEEDBACK? mwadman@fortunemail.com