As a nurse Mary Ann O'brien had done it hundreds, if not thousands, of times before, but this was a new experience for her husband, Timothy Meyers.* Sitting at the dining room table in their Cape Cod-style house in Central Ohio, Timothy quickly jabbed the needle into a vein in his wife's left arm. Maybe it was Mary Ann's good veins or Timothy's beginner's luck. Either way, it worked like a charm as about three teaspoons of blood glided effortlessly up the hypodermic needle and quickly filled a slim, glass vacuum tube. O'Brien whisked the "vacutainer" into an express mail package specially designed for biological materials like blood and shipped it overnight to Boston. Less than 24 hours later, Mary Ann O'brien's blood plus other assorted bodily fluids (serum from a blood sample drawn a few weeks earlier and a urine sample) were under lock and key in a super-cooled freezer in the basement of the Channing Laboratory. The Channing Lab is where Dr. Frank Speizer, E. H. Kass Professor of Medicine at Harvard Medical School and Professor of Environmental Science at the School, along with his colleagues at both the medical school and the School, collect and manage most of the data for the School's largest epidemiologic studies.

For O'Brien, 45, a little controlled bleeding is a small price to pay to do her bit for health and medical research--though as a subject in the Nurses' Health Study for almost ten years, she has spent hours filling out questionnaires about everything from how much broccoli she eats to oral contraceptive use. When she was in high school, O'Brien explains, she took an aptitude test and her score suggested she might pursue a career in research. She wound up being a nurse instead. "I guess if I can't do the research, I might as well be a subject," says O'Brien, who was contacted independently by the Review and not by study researchers. "This is my way of participating."

About 55,000 nurses have given blood to both phases of the Nurses' Health Study, although most of them have depended on the skill of trained colleagues rather than the luck of spouses for collecting the samples. For the nurses, it is a dollop of blood and, if they think like O'Brien, an act of altruism. For researchers, it is a wide and almost bottomless pool of information for them to plumb and use for the kind of genetic research that is sweeping through not only the School but the life and health sciences in general. Indeed, just a few years ago, it might have been possible to draw tidy lines around genetic research at the School. Now genetic research isn't so much a discrete category as it is a matter of doing research, period. Far from being a fad, this turn to genes is a "fundamental transformation in the nature of science," says Acting Dean James Ware. Max Essex, chair of the Department of Immunology and Infectious Diseases, agrees: "I can't think of anyone who does cutting-edge research on infectious diseases who isn't dependent on knowledge of the genetic sequence of whatever pathogen they are studying."

Although there are certainly those historic, paradigm-shifting, great conceptual leaps forward, much of scientific progress tends to be driven by technological advancements. In genetics several technologies have combined to grab the steering wheel and punch the accelerator to the floor. The development of the polymerase chain reaction (PCR) by Kary Mullis in the mid-1980s; the fast and always-getting-faster techniques and machines for sequencing the genome; various cloning capabilities that let scientists splice, dice, and generally manipulate DNA--all have combined to set a pace that seems to reflect the geometric, generative powers of DNA itself. Researchers have already identified over 6,600 human genes. The federal government's massive Human Genome Project is now competing with several private groups over who will get credit for identifying all 80,000 or more genes.

In medicine, in agriculture, in the food and pharmaceutical industries, the hope is that genetic discoveries will yield fresh and clever ways of intervening and manipulating all sorts of biological processes. It may be a cliché, but DNA is nature's blueprint. If that's the case, then on one end of the spectrum are the remodeling jobs, such as the development of "smarter drugs" that know how to home in on a particular gene or a particular mutation in a particular gene. On the other end is the total redesign of reproduction and all the troubling brave new world motifs of clones, organ factories, and selecting offspring for desirable traits.

For public health-oriented researchers, the genetics era of research poses several special challenges and opportunities. It takes some strategizing to find the right role to play in an area of science that is so hotly contested and highly commercialized. The promise genetics holds for improving medical treatment and for treating relatively rare, single-gene disorders seems pretty clear. Just how genetics will fit into broad, prevention-based health approaches and policies is far less certain. Moreover, some researchers at the School worry that the tried-and-true messages about not smoking, eating healthily, and staying active will get drowned out by all the buzz about genetics. "There are probably some folks out there waiting for a gene to be found that will suddenly make them thinner. But that may never happen. Meanwhile you can go out and try to start exercising," says Walter Willett, Fredrick John Stare Professor of Epidemiology and Nutrition, chair of the nutrition department, and professor in medicine at Harvard Medical School. "We don't want to forget what we already know."

Yet Willett is no public health Luddite. "All of this genetic research is legitimately interesting, and the fact that we don't know where it's going is what makes it interesting," he says. Edward Giovannucci, assistant professor of medicine at Harvard Medical School and assistant professor of nutrition at the School, says it's sometimes difficult to square genetic research with the commonly understood public health agenda. After all, one of the pillars of public health, at least in the affluent West, is to target behaviors that can be changed. Says Giovannucci, "But at the same time, as a scientist, you have to try to understand something as well as you can." Giovannucci had a study published in the Journal of the National Cancer Institute in August that linked selenium dietary intake to prostate cancer. "If I think selenium is important, and someone tells me that there is an important gene that is related to how we metabolize proteins that contain selenium, you can't ignore that. It's another potential avenue you have to take."

It's not hard to grasp why genetic-level research might be enticing to infectious disease researchers. As New York Times science reporter Nicholas Wade wrote recently, knowing the genome of a pathogen is like knowing the battle plans of the enemy--the tactics it might use to dodge the immune system and the tactics that might be used against it in the design of a vaccine or more-effective treatment drug. The relatively small genomes of viruses have been known for some time. Four years ago J. Craig Venter, head of one of the private groups racing to map the entire human genome, wowed the scientific world by being the first to completely sequence the genome of a free-living organism, the Haemophilus influenzae bacterium, which can cause meningitis and deafness. In June a French and English team announced that they had sequenced the entire genome of Mycobacterium tuberculosis, the tuberculosis pathogen. A month later the genetic code of the syphilis spirochete was cracked. Essex, whose lab recently acquired a second gene sequencer, is amazed at today's sequencing technology: "In the middle or late 1980s, for someone to get the whole sequence of a type of HIV might have taken a year with 25 people working on it. Now one postdoc can do a complete sequence of 10 African viruses in three weeks."

In the labs of infectious disease researchers at the School, the sequencer that can identify the base pairs that make up the genetic code and the PCR machine that can "amplify," or multiply many times over, a segment of DNA are as much a part of today's equipment as the microscope and petri dish were generations ago. For example, Professor Tun-Hou Lee's de-glycosylated gp120 strategy for making an aids vaccine is based, in part, on the ability to manipulate the DNA for gp120 so as to produce a very selectively de-sugared form of this HIV surface protein. In looking for ways to defeat the protozoan that causes a serious form of dysentery, Associate Professor John Samuelson has cloned genes for enzymes on which the colon-dwelling germ depends for survival. Clone the gene, and perhaps you can figure out a pharmaceutical way to thwart the action of that enzyme. Professor Donald Harn's lab has developed a so-called naked DNA vaccine against schistosomiasis, a snail-borne, tropical parasitic disease that kills approximately 200,000 people each year. Traditionally, vaccines have been constructed out of a killed or weakened version of the pathogen. Naked DNA vaccines, on the other hand, consist of just snippets of DNA rather than the whole germ. The idea is that the DNA will be "taken up" by the inoculated person's cells, and these cells will then express, or produce, the protein or protein fragment (the immunogen) necessary to produce a protective immune response.

Professor Ming Tsuang: looking for schizophrenia genes.

Other researchers are working on ways to foul up the genetic and cellular machinery of pathogenic microbes. Professor Dyann Wirth, the School's top malaria expert, gets on the Web almost every day to check out the ongoing effort to map the genome of the malaria-causing Plasmodium falciparum. She is looking for clues for designing drugs that will jam the cellular "pump" that P. falciparum uses to expel treatment drugs. Late last year Dean Designate Barry Bloom and his colleagues at Albert Einstein College of Medicine reported that they had invented a "phasmid": a cross between a phage (a virus that infects bacteria) and a plasmid (a small, circular piece of DNA). The phasmid will make it easier to create mutant versions of the pathogen that causes tuberculosis. Creating mutants is a key way to identify the genes that confer drug resistance to a pathogen.

So far, there is every reason to hope that the compilation of thick genetic dossiers on pathogens, as well as work delineating the genes behind the immune response to pathogens, will pay off with more-effective drugs and vaccines. But Essex, who, as an expert on the genetic instability of HIV, knows how protean pathogens can be, sounds a note of caution. "[Pathogens] are independent, evolving agents. They find a way to adapt. New ones emerge. Infectious disease is a moving target."

Of course, infectious disease research hasn't been the only beneficiary of genetic-level poking and prodding. One by one, scientists have nailed down the genes that previous research on family patterns had strongly hinted were responsible for inherited diseases. The list includes Huntington's disease, muscular dystrophy, cystic fibrosis, and some cancers. This kind of research depends on genetic markers- bits of repeated sequence for which the genome can be scanned- and a statistical technique called linkage analysis, which measures how tightly joined the markers are to the putative gene in diseased people.

Associate Professor Xiping Xu and postdoc Mark Seielstad: plumbing polygenetic disease in China.

Several faculty members are involved in research using linkage analysis. Professor Ming Tsuang, Stanley Cobb Professor of Psychiatry at Harvard Medical School and a professor in the epidemiology department, is a lead investigator on two key schizophrenia studies that use the sib-pair method to locate genes. The sib-pair method uses brothers and sisters with the same disease to find genes. One of Tsuang's studies consists of 120 sib-pairs, the other, 160; for this kind of research, this is an extraordinarily large and statistically powerful pool of subjects. Like many diseases, schizophrenia seems to result from a combination of inherited genetic and environmental factors, says Tsuang, with the genetic component accounting on average for about 50 percent of the risk. So Tsuang and other schizophrenia researchers have two major jobs ahead of them. The first is sifting through environmental factors. Tsuang suspects that in utero exposure to malnutrition, trauma, or infection that results in abnormal brain development may be contributing factors in people who are genetically predisposed to schizophrenia, a disease that generally doesn't mani-fest itself until people are in their late teens or early 20s. The second is pinpointing the genes that predispose people to developing schizophrenia. Four such genes are thought to exist, and Tsuang and others are collaborating with Millennium, a Cambridge, Massachusetts, biotech company, to identify them. "The pharmaceutical industry is interested in finding the genes because they are interested in developing medications to treat schizophrenia," says Tsuang. "From our point of view, it is not just treatment. We are talking about prevention. If we could find people who were susceptible to the disease, we might be able to protect them somehow."

Just how important the statistical side of this research can be was shown this summer when an analytical method developed by Nan Laird, chair of the Department of Biostatistics, and Deborah Blacker, an assistant professor in the Department of Epidemiology, proved crucial to the discovery of a new gene for Alzheimer's disease. Rudolph Tanzi, a top Alzheimer's gene researcher at Massachusetts General Hospital in Boston, was hot on the trail of a gene called A2M on chromosome 12 as a causative factor in Alzheimer's. But standard analysis didn't show any connection between A2M and Alzheimer's, and Tanzi's lab had hit a frustrating impasse. In a collaborative effort, Laird, Blacker, Marsha Wilcox, an epidemiology postdoctoral student, and Steven Horvath, a biostatistics doctoral student, developed a new statistical method for using DNA from unaffected siblings that showed that the Alzheimer's patients were three or four times as likely to have a mutant form of the A2M gene as their un-affected brothers and sisters. As the story on the discovery in the July 28th Science magazine said, the new method "struck pay dirt." Now Laird and her colleagues are trying to figure out ways to modify the method so cousins and other more distantly related people can be used in family-based gene studies.

The Program for Population Genetics, founded two years ago by Associate Professor Xiping Xu, also does research using families and sibling pairs. The School's largest single research endeavor focused solely on genetics, the program has about 20 researchers and DNA from over 10,000 people at its disposal. Xu has depended on contacts in his native province of Anhui, China to launch research projects there looking for genes that might cause hypertension and asthma. He has also recently started to work with collaborators in Lebanon. Unlike some single-gene disorders that have been identified so far (such as Huntington's disease), hypertension and asthma are thought to be caused by an interplay of environmental factors, such as diet, and several genes. In polygenetic disorders such as these, each gene may be contributing a "weak signal," thus creating huge "signal-to-noise" problems for researchers. Mark Seielstad, a postdoctoral student in Xu's lab who is working on the hypertension studies along with Xin Xu and John Rogus, says that the large number of subjects enrolled by the program is a way of over-coming this problem: with a large sample, researchers are turning up the volume, so to speak, and increasing their chances of picking up on a gene with a "weak signal."

Anhui Province is in a remote part of China. The fact that the population is relatively homogenous--the consequence of a group of people having descended from a few common ancestors--is also a boon to researchers. Genetic homogeneity increases the likelihood that the genetic markers used in linkage analysis will be shared by more people. Seielstad also points out that hypertension's low treatment rate in rural China eliminates a complicating factor in doing similar research in an American population.

Gene discovery is exciting science, the stuff of headlines and awards. But many researchers at the School are doing a brand of genetic research that takes the difficult but perhaps less glamorous step of plugging exposure into the genetics equation. They are asking what the health consequences will be if people with a particular gene--or a particular mutation in a particular gene--smoke, get exposed to toxic chemicals, or eat high-fat diets. Karl Kelsey, a professor in the Department of Cancer Cell Biology, says emphatically, "I am at the School of Public Health, and I am interested in diseases caused by exposures." Furthermore, says Kelsey, there are many questions about disease causation that the gene hunters simply can't answer. "When you have a disease that's caused by an exposure, penetrance [the frequency with which a gene manifests its effect] is necessarily going to be affected by exposure. Consequently, it becomes problematic and almost impossible to do classic linkage analysis to find genetic determinants."

But just how, and at what level, researchers go about studying the gene-exposure relationship is as varied as the answers they might find. Kelsey, for example, depends on tumor cells from Mass General lung cancer patients, whom he and others in his department have been following for years. Through PCR and other techniques, he is able to identify the genetic alterations in these tumor cells (they wouldn't be tumor cells if they weren't genetically fouled up in some way). Then he uses regression analysis to relate this genetic information to both disease and exposures. In a sense, what Kelsey is trying to do is recreate the crime--the crime being the transformation of a normal cell into its cancerous counterpart. "What I am interested in sorting out is the pattern of genetic alteration and how that gives rise to this disease," he explains. "How do you get there? What mechanistically occurs? And you can do that if you build a model, in an epidemiologic sense. But you have to know everything about the person. You've got to know when they started smoking, when they stopped, and all of those nice epidemiologic determinants." A paper that Kelsey published in Cancer Research in May is an example of the epidemiologic reasoning and design he is talking about. Combing through the genes of cells collected from 40 lung tumors, Kelsey and his colleagues showed that the tumor cells were missing chunks of a tumor suppressor gene called FHIT, an omission that was quite different from the reshuffling pattern, or "point mutations," seen in another tumor suppressor gene, the p53 gene. They were also able to show that asbestos exposure and smoking were significantly associated with these fhit deletions.

Associate Professor Claire Doerschuk: it makes sense to use antisense.

Yet there are other ways to skin the gene-exposure cat. Associate Professor Claire Doerschuk in the environmental health department is an expert on how lung inflammation at both the molecular and cellular levels is orchestrated in response to tobacco smoke, ozone, and bacterial infections. Although the inflammatory response is usually beneficial, it can also cause harm, so lung disease researchers have been intensely interested in how the inflammation is regulated and how, in certain situations, it might be switched off. Like many other researchers these days, Doerschuk uses mutant mice in her labs--25 varieties in her case. Although this is admittedly an oversimplified explanation, mutant mice experiments can be explained as disabling, or "knocking out," a gene that makes a molecule deemed crucial to some biological system of interest--and then watching to see what happens during the mouse's exposure to infection or other stimulus. It's common to buy off-the-shelf versions of knockout mice. But often researchers want to tailor make a knockout mouse to their investigation. Making a knockout mouse from scratch, so to speak, is a painstaking process, one that involves manipulating the genes of tiny mouse embryos consisting of just a few cells. For example, it took Brian Glassner in Toxicology Professor Leona Samson's lab over a year to make a knockout mouse for experiments that Samson is doing on DNA repair genes.

Mutant mice have nonetheless revolutionized disease and biological research. Samson says that when she started out as a young researcher in her native Great Britain 20 years ago, she was working with bacteria. Then she moved up to yeast. "The mouse gets us closer to the human situation," she says with obvious delight. Very close, indeed: Samson and her colleagues are now working with "humanized" mice, animals that have human genes spliced into their genomes. Laurie Glimcher, Irene Heinz Given Professor of Immunology, has made extensive use of mouse models in her investigations of T helper cells. Her experiments have shown that differing subsets of T helper cells- and the ratio of these subsets to each other- are crucial to a well-regulated immune response.

Still, to avoid working with mutant mice and the obstacles they pose, Doerschuk and others are now experimenting with antisense oligonucleotides--tiny bits of cDNA (the c stands for complementary) that, in effect, stick to mrna (the m stands for messenger) molecules inside a cell. Antisense is relatively easy and cheap to make, and "cells don't like it when cDNA and mRNA get together" the complexes get chewed up," says Doerschuk. Because mrna is crucial to the multistep process that results in the expression of a protein by a gene, the fact that the cDNA-mRNA combination is destroyed results in the same kind of knockout phenomenon as mutant mice but for a fraction of the cost and trouble.

The research that Doerschuk, Samson, Glimcher, and to some extent Kelsey are doing is sometimes called "bench science"--research done in biology's Lilliputian world of cells and molecules. The School's epidemiologists, on the other hand, attack health questions at the population level, and what they need are large numbers: studies with hundreds, if not thousands, of people. Because of their access to both phases of the Nurses' Health Study and its 230,000 women; the Health Professionals Follow-up Study and its 45,000 men; and some other large prospective cohort studies centered at the School and other Harvard-affiliated institutions, the School's epidemiologists are indeed richly blessed with large numbers. "It is important to have good ideas," says Giovannucci, "but you need to have good data."

These large cohort studies allow the School's epidemiologists to take two approaches to genetic research. First, they can examine the genes of several hundred people with a disease and look for mutations--sometimes called polymorphisms--that are associated with the disease. These are not gene-hunting expeditions, which most researchers would prefer to do in family studies, but rather a hard epidemiologic look at so-called candidate genes- genes that are under suspicion either because they previously have been identified in family studies or because they express a protein known to be a key player in a metabolic pathway or some other system. "Our strength is interpreting the significance of the variation in the gene once the gene has been identified with respect to disease risk," says David Hunter, associate professor of epidemiology. For example, Hunter is currently collaborating with Mary-Claire King, the renowned breast cancer gene researcher at the University of Washington, in a study of 600 or so breast cancer cases in the Nurses' Health Study. Using DNA extracted from the women's blood, they are sequencing the BRCA1 and BRCA2 genes, previously identified as inherited breast cancer genes, looking for different mutations. At first glance this kind of research bears a resemblance to Kelsey's work with lung cancer cells; the key difference is that Hunter and King are looking for patterns in inherited DNA from blood samples, not genetic changes in diseased tissue. Hunter says the BRCA1/BRCA2 project has proved to be a painstaking one because there are no obvious "hot spots" in these genes, so effectively the whole gene has to be sequenced.

Similarly, Giovannucci had 591 cases of prostate cancer to work with from the Physicians' Health Study when he launched a study of how a certain genetic pattern, called a CAG repeat, in androgen receptor genes might relate to prostate cancer risk. The androgens (such as testosterone) and androgen receptors control cell division in the prostate gland, so it seemed biologically plausible that the genes expressing androgen receptors might have an influence on cancer risk. In a paper published in the Proceedings of the National Academy of Sciences in April 1997, Giovannucci and his colleagues did, in fact, conclude that men with shorter CAG repeats were at a higher risk for prostate cancer.

Large numbers are particularly important to epidemiologists at the School who are pursuing gene-exposure research. Often the hypothesis in these studies is that a given gene or mutation doesn't trigger a disease unless a particular exposure comes along and makes the combination of that gene and that particular slice of the environment dangerous. Hunter says the rough rule of thumb is that such gene-environment interaction studies need four times the number of subjects as other epidemiologic studies. "With all the excitement about genes, you'll find in the literature studies where people have gone and looked at 50 cases and seen evidence of gene-environment interaction, when the standard theory tells us we really need to look at no fewer than 500 cases," he says. Of course, it's not just their size that makes these large prospective cohort studies valuable. They are also fairly bulging with exposure information from years of questionnaire responses.

Any kind of genetic research, even when ballasted with environmental factors, tends to conjure up ideas of biological on-off switches that a person or society can do little about. But as Hunter points out, this generation of genetically oriented epidemiology should be equally useful in pinpointing environmental causes of disease. He uses the red meat- colon cancer association as an example. While there is a fairly large body of epidemiologic and other evidence showing an association between red meat consumption and colon cancer, no one can say what exactly it is about red meat that causes the disease. Is it the fat? A subgroup of the fat? And it's a safe bet that no amount of dietary questioning is going to provide the answer. But in a study recently published in Cancer Research, Jia Chen from Hunter's research team showed that people who, because of a genetic alteration, tend to activate heterocyclic amines (carcinogenic compounds in cooked meat) are more prone to colon cancer. Thus a genetic study, ironically, points to an environmental exposure--heterocyclic amines--as the health culprit. In Hunter's words, Chen's study "pulled a specific carcinogen out of that complex mixture we call red meat." The public health message may now be not to overcook meat- though of course that advice flies in the face of worries about Escherichia coli contamination from undercooked meat. With a wan smile Hunter says, "Well, we so often have to deal with tradeoffs in public health."

Though they certainly get most of the spotlight, the School's prospective cohort studies are not the only game in town, genetically speaking. Kathleen Egan, S.D.' 96, a research associate in the epidemiology department, is in the midst of collecting cheek cell samples for DNA analysis from the women in a huge but largely unheralded breast cancer case-control study that was started 10 years ago. In contrast to the 600 breast cancer cases in the famous Nurses' Health Study, Egan says that this lesser-known case-control study has accrued 14,000 women with the disease. So far, Egan has seen a response rate of 70 to 80 percent among cases and 60 to 70 percent among controls. One gene the study is looking at- the N-acetyltransferase (NAT2) gene- plays a crucial role in the metabolism of a carcinogenic constituent of tobacco smoke. In most large epidemiologic studies, smoking has not been found to be a risk factor for breast cancer. But Egan says that results from a NAT2 analysis in one study showed that women who smoke and have a particular form of NAT2 were in fact at a higher risk.

Over a century ago, the development of the germ theory ushered in a new era of public health and medicine. The enthusiasts (and there are many) for this new era of genetic research say it could produce comparable gains for human health and well-being. Better treatments seem likely as researchers figure out ways to jam a genetic stick into the spokes of the disease process. The hope is that a disease will be stopped at the genetic level before it even gets a chance to really get started. Egan says that if the defective protein that causes breast cancer in women with the BRCA1 and BRCA2 genes could somehow be altered or replaced, "you are talking about potentially eliminating many familial breast cancers." Genetic research should give scientists a new edge on pathogens, the quick-change artists of the disease world. In terms of prevention, perhaps people who are susceptible to a disease could be treated or counseled to avoid certain exposures. Tsuang, for example, says that finding a schizophrenia gene would be extremely helpful in designing teenage intervention programs that he believes might head off or ameliorate the disease.

Yet the era of genetic research is full of pitfalls too. Researchers are facing thorny issues concerning the informed consent standards for genetic research. Is it sufficient to get sweeping consent from a subject to do genetic-type research, or must the researcher get consent "one gene at a time"? And what are the obligations of the researcher to inform subjects who are found to have disease genes, particularly if those genes are "high penetrance" genes that carry a high likelihood of disease? For their part, research subjects are worried that their genetic profiles will fall into the wrong hands. "You have to remember that we are doing this research against a backdrop of all these scary headlines that frighten women who are worried, that no matter what we say, the results are somehow going to end up in the hands of the insurance companies," says Egan. But it's not just insurance companies they are worried about. Xu's work in China was briefly imperiled in 1997 when he was the subject of a short-lived but rancorous controversy after some accounts in the Chinese press, based on bad translations, portrayed him as shipping 200 million Chinese blood samples to the United States. Xu was the victim of some bad journalism, but the furor it touched off clearly tapped into some widely held fears in China and elsewhere that Western governments and pharmaceutical companies are exploiting other people's genetic heritage. It's because of this fear that controversy has dogged the federal Human Genome Diversity Project, an evolutionary biology study that has sought to collect genetic samples from people worldwide.

Outside of methodology issues, public health-oriented researchers are also concerned about how genetic research will affect one of the dominant themes in public health for the last 20 years--namely, that risk factors for chronic diseases can be modified through behavior. Due in no small part to the flood of findings from the School's large cohort studies, the refrain has been that if you don't smoke, eat right, and get plenty of exercise, your chances of good health will improve (putting aside for the moment the arguments that larger societal forces have been underplayed in the rush to identify modifiable risk factors). In some ways all this genetic research could mean just a further refinement of that message: people who are genetically susceptible to a given condition or behavior might benefit from targeted programs or biomedical interventions, just as people with high cholesterol levels are treated to stave off heart disease. Yet genetic susceptibility (and its flip side, protection) may be used to muddy the behavior message because health risk could come to be seen as more a matter of what's in the DNA than what's on the dinner plate or how often you take a walk. A fair number of people already believe that "it's all in the genes."

In fact, even when the genes themselves have not yet been identified, the very concept of genetic susceptibility is creating waves. For example, Marianne Berwick, an epidemiologist at Memorial Sloan-Kettering Cancer Center in New York, has caused a stir with her assertion that sunburn is not the direct cause of skin cancer but perhaps merely a marker for genetic susceptibility. It follows from Berwick's view that the broad-based advice that everyone should slather on sunscreen is misguided. This opinion hasn't made Berwick too popular at the American Academy of Dermatology, a strong proponent of sunscreens. Or consider the statistic often cited by smokers in denial: only one in ten smokers gets lung cancer. Says Hunter, "We may unravel the reason why only ten percent of smokers get lung cancer. It might have something to do with genetic susceptibility. But if that distracts from the immediate goal of reducing cigarette smoking, that would be unfortunate." Acting Dean Ware says that in the final analysis, genetic research will inform rather than undermine today's public health. "Environmental factors play a key role in diseases such as heart disease and cancer. It is clear that most of us have some degree of genetic susceptibility to those diseases," he notes. "For most chronic diseases and for most people, the role of genes will be a complicated, multifactorial story. Yes, they will be important. Their role will be interesting. But behavior and environment will still be important."

Back in Ohio, Mary Ann O'Brien says she hopes the fruits of genetic research could be used to guide and counsel women like her sister, who had several children with spina bifida. O'Brien also has a friend who, after having an anencephalic baby, got genetic testing and counseling. She subsequently took the risk and had more children, who fortunately were healthy. In O'Brien's opinion it was helpful for her friend to know where she stood, genetically speaking. "At least she knew the odds." As for herself, O'Brien says she is not overly concerned about insurance companies or employers using genetic information against her; breast cancer, she notes, does not run in her family. If she were tested and found to have a breast cancer gene, O'Brien says she would be inclined to see the up side: "I might be a little more compliant with self-examination if I knew I had the gene.

*pseudonyms to protect confidentiality

Immaculata De Vivo's formula for making some great DNA: turning blood into C, T, G, and A.

The Nurses' Health study and other large, perspective cohort studies used by researchers at the school and other Harvard-affiliated institutions are vast store-houses of information about health, disease, nutrition and other kinds of exposures. Now researchers are adding genetic-level research to their repertoire using blood and other tissue samples collected as part of these studies.

Instructor in medicine at Harvard Medical School Immaculata De Vivo, a researcher in the laboratory affiliated with the school's Center for cancer Prevention (headed by Center Director David Hunter), described the steps necessary to turn a blood sample into usable information for a case-control study "nested with the Nurses' Health Study.

1: A Nurse Draws about 10 milliliters of blood--about 3 teaspoons' worth--from ma vein in the arm. Some studies are now depending on a swab of tissue taken from inside the cheek rather than blood.

Samples get sent by overnight mail service to the Channing Laboratory on Longwood Avenue. Many of the School's epidemiologists are affiliated with the Channing Lab.

2: Blood is processed chemically and mechanically into three parts: red cells, plasma, and the "buffy coat," the thin, dirty-white layer that contains the white blood cells. Each subject in the Nurses' Health Study has been assigned a number; samples are bar-coded with this number and then put into a freezer cooled to -80 degrees Celsius (or -112 degrees Fahrenheit).

3: A committee that oversees the Nurses' Health Study must give researchers the okay to use blood samples. "Those are very, very prized samples," says De Vivo. "It is very well, and very tightly, regulated."

Approval is also needed from a separate blood committee and from the School's Human Subjects Committee, which reviews research projects to make sure they meet ethical guidelines.

4: frozen White Blood Cells are sent over to the researcher. A typical, 50 microliter sample has about 10 million white blood cells. Each of these cells has the complete genome of the person from whom the blood was taken-- All 46 chromosomes and 80,000-- 100,000 genes.

Researchers used to extract DNA from white blood cells with a phenol chloroform solution; it would take about three weeks to process 100 samples. Now everything is automated and the job can be done in four days, De Vivo says.

With DNA in hand, she exults, "Now we are ready to get down to business!"

5: Polymerase Chain Reaction (PCR) is used to isolate, then "amplify"--or multiply many times over--the DNA. With heat and bits of DNA called primers, PCR "unzips" the double-helix and separates it into single strands. Then, by attaching molecular mirror images to these single strands, the PCR process creates copies.

Each cycle of PCR doubles the amount of genetic material to be studied. In a matter of hours, a PCR machine can generate one billion copies.

6: In a process called single-strand, conformation polymorphism analysis, the PCR-generated DNA is "pre-screened" using an ABI 377 automated sequencer. The idea is to pick out differences (variants) between the DNA that came from the study subjects and a "reference strand" form a national, standardized gene bank.

7: Researchers look for the stretches of DNA that are different than the norm and then sequence them. Sequencing means to find the exact order of adenine (A), cytosine (C), thymine (T), and guanine (G) bases. These sequences of bases are the defining characteristic of a stretch of DNA. They determine what kind of protein will be generated by a gene.

These bases, and the order in which they come, are to DNA what the alphabet and letters are to the written word; if the bases are out of order or some are missing, the DNA can't be "read" properly. De Vivo describes this step as "defining the variant."

8: With the Variants Identified, epidemiologic analysis and reasoning comes into play: is the variant more common in cases (sick people) compared to controls (well people)? How much more common? Other factors such as diet are folded into the equation to see if a given genetic variant, in combination with some exposure (a high-fat diet, for example) might lead to higher risk. "The strength of the Nurses' Health study and the Channing is the ability to do a lot of gene- environment work," comments De Vivo. The Channing Laboratory also has a corps of statisticians that sometimes helps with analysis.

9: The researchers submit a manuscript to the committee that gave the study the initial go ahead. The committee provides what is, in effect, and in-house peer review, checking the number and looking for any holes in the study's argument and presentation.

10: Voila! Regardless if an association is found between the disease and the gene, if the study is well-designed and executed, and the research gets published, then the formula is a success.