Creation Of Linkage Map Falters, Posing Delay For Genome Project

Researchers, discouraged by mapping's drudgery, doubt that a 5-year plan to finish high-resolution image is now feasible. WASHINGTON--After three years of escalating expectations, rising financial support, and congressional accolades, the Human Genome Initiative has encountered its first major hurdle. A key element of the project has fallen several years behind schedule, in part because peer review panels at the National Institutes of Health decided that some incoming grant proposals on the topic were "boring." Of those scientists who were funded, many have diverted part of their research toward more tantalizing pursuits. The process of catching up may require a dramatic rethinking of many of the strategies of the 15-year, $3 billion project. The immediate obstacle stems from a lack of progress on the so-called linkage map of the genome. The term refers to a list that links human traits and their associated genes to locations on the human DNA molecule. Matched with a detailed physical map of the consecutive chemical "base pairs" that make up all genetic material, the linkage map serves as a guide to the human genome and to the secrets of genetic diseases. But before scientists can read this "Encyclopedia of Man," the cryptic sequences must be checked and put in context. That task requires a high-resolution linkage map perhaps five times more detailed than the best one currently available. Although two years ago such maps were thought to be well within reach, progress has slowed to a crawl and only one major lab now gets NIH funding for large-scale linkage mapping. Some scientists worry that this situation bodes ill for the future. "Linkage mapping was the first test. We have come up against a repetitive task and we're not doing it," says David Botstein, a geneticist and vice president at Genentech Inc., a San Francisco-based biotechnology firm. "If we get slowed down by policies and politics on this relatively simple-minded task, I'm beginning to have doubts about the entire effort." Botstein made his remarks during a meeting last month of an advisory committee, of which he is a member, that helps NIH direct the project. The 12-member panel of scientists is reviewing the first five-year plan for mapping and sequencing the human genome for the newly created Center For Human Genome Research. Although the center has existed officially only since October 1, NIH has been issuing grants aimed at a human genome project for several years. But if the linkage map is any indication, the agency is still wrestling with basic issues in funding large-scale, long-term research. A 1988 report by the National Research Council (NRC) recommended that the project shoot first for a linkage map with a resolution of one centiMorgan, a unit of genetic distance equal to about a million base pairs, or 0.03% of the entire human genome. "It should be possible," the report states, "to complete a [1-centiMorgan linkage map] within five years with the investment of more money and the encouragement of coordinated efforts." The goal was heartily endorsed by James Watson, director of NIH's component of the project, as he took office a year and a half ago. But although the project's annual NIH budget has grown from $17 million to $60 million since then (the Department of Energy will spend another $27 million this year on genome-related work), scientists have made little progress toward the linkage-mapping target. Two years after the NRC report, Washington University geneticist Maynard Olson, who was a member of the NRC committee and now serves on the NIH panel, has become far less optimistic. "I think that there is zero probability that we will develop 1-centiMorgan maps in the next five years," he told fellow panel members last month. The reasons why genome mapping has stumbled point to the pitfalls of big science. Some scientists have rebelled at the numbingly repetitive nature of the work, while others who had actually embarked on large scale mapping projects have redirected their research at the first sign of a possibly interesting disease gene. Still others claim that NIH review panels have discouraged researchers by rejecting linkage- mapping grant applications in favor of studies of specific genes. And as other mapping technologies have matured, the picture has become increasingly clouded. The critical role of linkage mapping in the genome project has heightened tensions about its progress. Linkage maps are assembled by correlating the relative positions of certain genes with the probability that they will be inherited together. So, for example, if blue eyes tend to be inherited with blond hair, it is reasonable to expect that the genes for blue eyes and those for blond hair are located near each other and are likely to remain together during the natural breaking and reassembling process of DNA recombination. This ability to relate a location on the genome to physical traits makes linkage maps invaluable in extracting biological meaning from the 3-billion-long string of base pairs that make up the physical map of the genome. "We feel that the physical mapping alone is going to be so difficult that we'll need every possible check," says Botstein. "When doing physical mapping, the question is: `How do I know that I've done some very complicated calculation correctly?' You know it because it agrees with the predictions that come out of linkage mapping. But the physical map has yet to be tested against any such measure." Nor is that likely to change anytime soon. At last month's NIH meeting, the panel members agreed to rewrite the draft five-year plan to call for a considerably coarser linkage map than that recommended by the NRC committee. The final NIH plan is expected to recommend that at least a five centiMorgan-resolution map of the human genome be completed in five years. Because the amount of work that goes into a linkage map is roughly proportional to the square of the resolution, the fivefold decrease in the linkage map goal will translate to a mapping effort 25 times less arduous, scientists say. Although some researchers argue that a five-centiMorgan map is nearly as useful as a one-centiM genome, most agree that the shift is an genome, most agree that the shift is an embarrassing retreat for the NIH project and a poor omen for the similar tasks that still lie ahead. More embarrassing yet is the fact that linkage mapping is based on techniques that have changed little since the optimistic NRC report. As opposed to physical mapping technology, which is far from a plausible embarking point, the machinery needed for linkage mapping "is today's technology," says Botstein. Indeed, a 10-centiMorgan map of the entire human genome has existed since 1987. Because the work is relatively straightforward, although time-consuming, scientists see NIH's struggle with the linkage map as an indicator of future problems for the genome project. One hurdle that is already apparent is the human factor. "It's very laborious, hard work," says Benjamin Barnhart, director of the Department of Energy's portion of the project, about mapping entire chromosomes. "It's many 12-hour days and seven-day work weeks." While people are willing to do that for a certain length of time, he says, one can't expect them to do it forever. "Graduate students or postdocs tend to put in more time than they should on things because they're idealistic and keenly motivated, but even they burn out," he says. Adds NIH panel member Phillip Sharp, a cancer researcher at the Massachusetts Institute of Technology, "young people have to do something interesting within five to seven years [of receiving their degree]. If they don't, they'll get bored, they won't get promoted or get tenure, and they won't get grants." Another looming problem is the natural tendency for scientists to grab at the brass ring of any disease gene that they may come across in the process of mapping a chromosome. No matter what their grant may say, says Caltech geneticist Leroy Hood, scientists "get a fascination with disease-causing genes and tend to focus on them." One possible solution, says University of Michigan geneticist Francis Collins, who led the team that recently discovered the gene believed responsible for cystic fibrosis in most people, is to offer scientists a reasonable time delay before reporting results. That would allow them to go back and explore their data after finishing mapping the chromosome. Although such a move is likely to slow the project by keeping important information out of the public domain, the risks are acceptable if it results in retaining top scientists, Collins says. "We will have to offer the best researchers a real opportunity to follow up on what they find," he says. "It would be ridiculous to insist that the genome project proceed to completion before anybody does any biology. The goal may be to get the Encyclopedia of Man, but let's read it a little bit as we go over it." Other scientists point the finger squarely at NIH when discussing the flagging linkage map. "I would have thought that by now you'd have three or four big grants to do this, but I don't see them," says Botstein. "The NRC report two years ago said that now is the time for substantial spending on high-resolution mapping. I won't speculate on why it wasn't done, but it wasn't." Although the NRC report recommended a $40 million commitment over several years, NIH spent only $5.4 million on linkage mapping last year, most of which went to the pursuit of specific genes. Indeed, the only major NIH grant for chromosome linkage mapping was to geneticist Ray White at the University of Utah. And even part of White's team has been spun off to pursue particularly interesting genes. Mark Guyer, assistant director for program coordination of the NIH genome office, acknowledges that NIH may not have spent enough on the linkage map. But he says that tight budgets have often tied the program's hands. At the same time, he adds, "I think the discussion at the advisory committee meeting will prompt us to go back and reassess what the situation is with regard to linkage mapping." Another worry is that NIH peer review panels may have discouraged the few researchers who had actually submitted proposals to map full chromosomes. In 1987, Helen Donis-Keller, a molecular biologist at the Bedford, Mass.-based biotech company Collaborative Research Inc., released the first linkage map of the entire genome. But since then she has had several proposals rejected for alleged lack of originality. "Time after time the NIH study sections told me that [large-scale linkage mapping] is not exciting, not innovative, boring," she says. "I'm becoming very cynical about how [the genome project's] mandate is being carried out." Now at Washington University, she has pieced together with mostly private funding a linkage map at an average resolution of six centiMorgans. Guyer acknowledges that peer review decisions may have hindered linkage mapping in the past, but he says that the new genome office has solved the problem. "Early review committees were not given instructions about policy matters and sometimes made inappropriate decisions about what was interesting or not," he says. Over the next year project officials will likely reach a meeting of minds on the future of linkage mapping. One solution is to give physical mapping the highest priority and to return eventually to linkage mapping with refined technology and a better understanding of the genome. Officials are also considering a move away from traditional grants and toward contracts with private companies. Both possibilities have supporters and detractors. "If you had made the decision to go out and give some company $10 million a year for the next seven years to do a linkage map, they would get a linkage map," says MIT's Sharp. "However, if the physical map is the right way to get the linkage map, which it now appears to be, you would have frozen $70 million into some program that science would have eclipsed in the first couple years," he warns. But others believe that too much flexibility is inefficient. Argues Botstein, "What we're doing is a lot of starting and making very little progress towards finishing. If scientists won't get us this data, then we should buy it from industry. The critical thing here is closure."

Creation Of Linkage Map Falters, Posing Delay For Genome Project

Interested in reading more?

Become a Member of