Realizing the Promise of Genomics in Biomedical Research

In April 2003, the Human Genome Project achieved all of its original goals, including production of a finished sequence of the human genome.¹ With that historic achievement, the Human Genome Project ended and the “genome era” began. Especially because this transition occurred only days before the 50th anniversary of Watson and Crick’s article describing the DNA double helix,² it led to much discussion regarding the future of genetics and genomics and how they would affect biological exploration, health, and even society.^{3 - 8} Now, 2 years into the genome era, is biomedical research any closer to that future, and, if so, what does it look like?

Several current genome initiatives afford a look at the future even as they bring it closer. One is an international effort to create a haplotype map (HapMap) to explore variation within the human genome.^{9 - 10} Now rapidly approaching completion, the HapMap Project¹¹ has already provided lessons about geographic distribution of human genome variation and has contributed convincing evidence that the variation in the human genome is organized into local neighborhoods, or haplotypes.¹² Because of this structure, knowledge of which base is located in a particular variable position of an individual’s genome sequence allows very good prediction of the variants present nearby, a phenomenon referred to as linkage disequilibrium. These regions of disequilibrium operate over variable distances, but often stretch across 10 to 30 kilobases. The HapMap project is defining the boundaries of these segments of linkage disequilibrium across the whole genome. This now allows researchers to choose a panel of a few hundred thousand single-nucleotide polymorphisms (SNPs) that ably represent variation across the whole genome and, by genotyping only this panel, to achieve an excellent first approximation of the common variants in an individual’s entire, 3-billion–base pair genome. HapMap-based SNPs already have enabled research that was impractical, if not unimaginable, only a few years ago. An example is the research published this year that identified the complement factor H gene as commonly involved in age-related macular degeneration.^{13 - 15}

Another current initiative is the Encyclopedia of DNA Elements project,¹⁶ which seeks to completely identify the biological mechanisms active in a carefully chosen representative selection (about 1%) of the human genome. This effort has become increasingly important as evidence has mounted that not only the small portion of human DNA that codes for proteins has biological importance. In fact, evolutionary comparisons between multiple mammalian genome sequences indicate that about two thirds of the most strongly conserved elements of the human genome fall outside of these protein-coding regions.¹⁷ The “junk” in “junk DNA” describes the current level of understanding of this part of the genome, not its biological significance.

Another ongoing effort that will help realize the promise of genomics is in the area of “chemical genomics.” The goal of this application of genomics is to expand the universe of small molecules that can be used as probes for understanding biological pathways. While the human genome appears to contain more than 20 000 genes, the entire current pharmacopoeia targets only about 500 genes and their products.¹⁸ In the past decade, advances in combinatorial chemistry have allowed the generation of libraries that contain hundreds of thousands of compounds, representing a broad diversity of chemical shapes. Improved methods of high-throughput screening have made it possible to screen these libraries with assays for particular targets, generating agonists and antagonists for particular proteins, pathways, or cellular phenotypes. However, most of this activity has occurred in the private sector, where only a minority of targets have been pursued as being “druggable.” As part of the Roadmap process,¹⁹ the National Institutes of Health has recently established a network of chemical genomics centers available to all researchers, as well as a new database, PubChem,²⁰ that makes much of the resultant data freely accessible. Availability of these tools will enable academic researchers to broaden the diversity of targets for chemical genomics. While most of the compounds identified in this way will be limited to use as research probes, a small percentage may even go forward into early-stage drug discovery.

Another critical area of current emphasis is to understand the complex interactions of genetic and environmental factors in health and disease. Case-control studies have proved valuable in adding to such understanding. In the past, however, such studies have generally been forced to limit their search to candidate genes. Given that the chosen candidate genes were generally based on hunches with a high risk of error, such studies have frequently lacked power or have uncovered only weak associations that held up poorly in replication studies. That is all about to change. The use of new tools, particularly the HapMap, should dramatically increase the usefulness and power of case-control studies by making practical a SNP-based methodology that effectively samples the entire human genome. Based on this approach, it is likely that many of the major gene variants that contribute to diabetes, heart disease, Alzheimer disease, common cancers, mental illness, hypertension, asthma, and a host of other common disorders will be discovered in the next few years.

Despite the usefulness of case-control studies for discovering gene variants associated with increased disease risk, they have limits. They often contain significant biases due to case ascertainment methodology, provide little information about predictive biomarkers, and are flawed with respect to determination of environmental risk factors because of recall bias. For these and other methodological and biological reasons, rigorous quantitative understanding of the role of genes and environment in health and disease can come only from large, population-representative, prospective cohort studies.²¹ Several such studies have recently begun or entered the planning stage in a number of countries. These will be important resources for understanding the relationships among genetic and environmental factors and health and disease. But, no matter how helpful studies in other countries are, they alone cannot meet the needs of the United States. Only a US-based study could adequately sample important US minority populations, gather data about environmental risk factors characteristic of the United States, and provide US researchers with full access to data and biological specimens. Failing to mount a US-based cohort study in the near future will compromise US scientists as well as the ability to understand and improve health for US citizens.

New technologies, and new applications of those already established, will enable and shape much of the future of genomics. An obvious example is genome sequencing. Over the past decade, technological advances have lowered the cost of sequencing at a fairly constant rate, halving it approximately every 22 months. Currently, sequencing 1000 high-quality bases of DNA costs less than $1 (J. Peterson, PhD, and K. Wetterstrand, MS, written communication, August 1, 2005). It appears that refinements of the current gel-based sequencing methods will allow this trend to continue for another few years. Moreover, within a few years new “disruptive” technologies, many based on single-molecule sequencing on solid supports, should accelerate this rapid decrease in sequencing costs. If that proves true, the promise of the $1000 genome may be less than a decade away.

Whether within 10 or 12 (or 8) years, such inexpensive sequencing will change both research and clinical care, and progress does not need to wait even that long. The National Human Genome Research Institute (NHGRI) plans to focus a significant portion of the sequencing capacity that it supports on medical sequencing. For instance, the NHGRI and the National Cancer Institute are actively considering a Human Cancer Genome Project,²² which would use DNA sequencing and a host of other genome technologies to gather information about the mutations and functional abnormalities found in multiple samples from many major types of cancer. Medical sequencing should also provide important insight into many other diseases. For example, sequencing all exons in X-linked mental retardation syndromes may reveal much about their etiology. Sequencing candidate genes in the extremes of the distribution of quantitative traits should also reveal much of importance about common diseases, such as coronary atherosclerosis.²³ With further technological advances, other previously unimaginable research approaches will become real.

Similarly, clinical care will change dramatically. In a relatively few years, when the role of specific genetic factors in disease is more fully understood and a human genome can be sequenced for less than the cost of a colonoscopy (for example), an individual’s sequence will likely become part of the standard medical record, especially since, unlike the colon, an individual’s genome sequence is relatively static. Thus, unlike colonoscopy, sequencing will not require frequent repetition. Similarly, it will become the standard of care to sequence cancer patients’ tumors and to use that information to refine prognosis and guide therapy.

New technologies must also integrate with older ones. For instance, it is incorrect to assume that genetic testing will soon relegate that old standby of genetics, the family history, to the dustbin. Instead, optimal patient care will use a combination of the old and new, as exemplified by interpretation of BRCA1 testing, in which the risk that a woman who tests positive for a BRCA1 mutation will actually develop breast cancer is most precisely calculated based on knowledge of her family history.²⁴ Indeed, an important step in preparing patients and health professionals for the onslaught of new genomics-based clinical tools is to make the use of the family history more efficient, effective, and ubiquitous in health care.²⁵

The broad impact of genomics on health in the near future should come from its applications to diagnostics, pharmacogenomics, and drug development. Some genomic medicine has already started to move from promise to reality. An example in diagnostics is the use of multigene analysis to predict the need for chemotherapy in certain breast cancers.²⁶ In pharmacogenomics, knowledge of a patient’s genetic makeup can sometimes help avoid adverse drug reactions, as a study earlier this year showed for warfarin.²⁷ In drug development, genomics can already salvage a drug that might otherwise be abandoned, as demonstrated by gefitinib, which has little benefit for the majority of those with lung cancer but may prove lifesaving in the sizable minority with a specific genotype.^{28 - 29}

Successfully integrating genomics into health care will require more than simply an understanding of how genetic factors influence health and drug response. Health professionals will need to become genomically literate. New curricula and educational models must be developed to achieve that goal, an effort that the National Coalition for Health Professional Education in Genetics is spearheading.^{30 - 31} Behavioral science research will need to establish how best to use genomic information to affect health behaviors, and outcomes research will need to validate which applications of genomics to health care are cost-effective.

Another challenge will be to apply the benefits of genomic medicine to the developing world.³² It would be incorrect to assume that genomic approaches are irrelevant to the health needs of resource-poor nations. For example, the genome sequences of the Anopheles gambiae mosquito³³ and the malaria parasite³⁴ are both in hand, and knowledge of genetic factors that influence host response to malaria and its treatments is increasing.³⁵ With such advances, it is easy to foresee genomics contributing to more effective treatment and prevention for malaria, as well as many other leading causes of morbidity and mortality in the resource-poor world.

Genomics will have a paradigm-altering impact, not only on current understanding of biology and on health, but also on other aspects of society. The ability to prepare for and manage that societal impact will determine whether genomics lives up to its promise, as much as will the ability to use genomics effectively in biology and health. For instance, genomics will help unravel not only those complex gene-environment interactions that lead to health and disease but also those that lead to many human traits; eg, genomics tools will identify inherited factors that play a role in atherosclerotic disease and diabetes but will also identify inherited factors that play a role in traits such as height, nicotine dependence, and intelligence.

The wisdom of the traditional watchwords primum non nocere applies to the societal implications of genomics at least as much as to its health implications. Recognition of this is one reason for the financial and intellectual investment that the field of genomics has historically made in consideration of ethical, legal, and social implications (ELSI) issues³⁶ —an investment unparalleled in biomedical research. It is also a reason that education of a broad array of society about genomics remains a vital challenge. Society will deal effectively with the complex issues unveiled by genomics only if a broad array of its members are sufficiently knowledgeable and comfortable about the science involved to engage actively in societal discussion and decision making. The societal impact of genomics also makes it imperative that genomics researchers reflect the society from which they come; otherwise, society’s interests in the research questions asked and how they are answered will be only incompletely represented.

One societal concern raised by new genetics technologies, ie, potential discrimination in health insurance or employment based on results of genetic testing, has been apparent for several years and requires a national legislative solution. Data show that, because of this concern, much of the US public hesitates, or even refuses, to participate in genetics research or to avail themselves of potentially helpful clinical genetic testing.³⁷ Fortunately, this year the US Senate unanimously passed legislation, the Genetic Information Nondiscrimination Act of 2005,³⁸ that protects against such discrimination, and the president has indicated that he would sign this bill. The bill has been introduced in the House of Representatives and is assigned to 3 committees, but no hearings have yet been scheduled.

It is also evident that genomics will alter views of race and ethnicity. Much has been made about the facts made clear by genomics that any 2 individuals share about 99.9% of their DNA, that variation in the DNA sequence is greater within population groups than between them, and that few, if any, common DNA sequence variants occur only within a single population group. Nonetheless, it is also clear that, largely for geographic and historical reasons, some DNA variants differ significantly in frequency in different groups. Among such variants will be those that help create the surface physical features that many people use to identify, and some to define, racial and ethnic groups. How will identification of those variants affect concepts of race and ethnicity? Will it reify race as a biological concept or emphasize the meaninglessness of using features influenced (and not even fully determined) by only a small number of genes to define populations? Panels of variants, because of their population frequencies, are already being used to link individuals to certain ancestral populations.³⁹ But the precision of such estimates is likely overstated by many commercial suppliers. What happens when these “scientific” population assignments differ from those previously applied to individuals—either by themselves or by others? Will this affect how health professionals approach individual patients and, for public health purposes, populations? Genomics will force facing these sorts of questions.

While the claims for the immediate impact of genomics have sometimes been overstated, the ultimate consequences of the integration of genomics into medical research and medical practice are likely to be revolutionary. By providing insights into the networks and pathways of biology, genomics has already begun to alter the fundamental understanding of health and disease, even for diseases that few would describe as “genetic.” By providing more sophisticated knowledge of biology at the individual level and of disease typology, genomics has already begun to personalize health care. By widening the number of potential drug targets and better identifying those people a specific drug is likely to benefit and those it is likely to harm, genomics has already begun to expand the pharmacopoeia. By offering novel approaches to diseases that haunt the developing world, genomics has already begun to affect global health. By changing societal discussion of race and ethnicity, genomics has already begun to change society.

Two years into the genome era, we have already begun to realize the promise of genomics—but the best is yet to come.