The Foundations of Genomic and Personalized Medicine

Geoffrey S. Ginsburg and Huntington F. Willard

1

CHAPTER

Essentials of Genomic and Personalized Medicineby Ginsburg & Willard

GENOMIC AND PERSONALIZED MEDICINEThe Human Genome Project, completed in 2003, has provided scientists and clinicians with a diverse set of novel molecular tools that can be used to understand health and manage disease. Variation in the human genome has long been the cornerstone of the fi eld of human genetics (see Box 1.1), and its study led to the establishment of the medical specialty of medical genetics (Nussbaum et al., 2007). Now genome sequencing, copy number variation, transcriptional readouts, and comprehensive measure-ments of micro RNA, protein, and metabolite levels provide a “systems approach” to probe and predict human health and dis-ease states that has greatly broadened the impact of principles of genetics and genomics on clinical medicine. These advances have provided both a conceptual and technological underpin-ning for the development of the fi eld of genomic medicine as a driver of personalized health care. For the fi rst time in the his-tory of medicine, health care providers as well as patients can use predictive tools to develop a new model for health care based on health planning that is proactive and preventive, as opposed to the current model in health care that is reactive, episodic, and geared toward acute crisis intervention once disease is already manifest and largely irreversible.

Copyright © 2010, Elsevier Inc.All rights reserved.

The growing transformation of clinical practice in the era of genomic and personalized medicine is perhaps best exempli-fi ed today in the fi eld of cancer care, as illustrated in some detail in subsequent chapters of this volume. Oncologists now prac-tice with a suite of genomic testing opportunities that include BRAC1/BRAC2 testing in familial syndromes of breast and ovarian cancers. In colorectal cancer, Hereditary Non-polypo-sis Colon Cancer (HNPCC) or Lynch syndrome and familial adenomatous polyposis (FAP) coli are conditions for which there is testing for mismatch repair gene mutations (for HNPCC) or APC mutations (FAP) that has been widely adopted. The para-digm for oncology is largely based on the principle that accurate prognosis and proper therapy can be matched to the molecular characteristics of the individual patient’s tumor. Thus, at the time of diagnosis, whole-genome expression data are now being used routinely to identify subtypes of cancer not previously recognized by traditional methods of analysis (Bullinger and Valk, 2005; Dave et al., 2006; Potti et al., 2006a; Staudt, 2003; Valk et al., 2004).

There is now compelling evidence of clinical adop-tion of genomic testing by oncologists: in 2008, for example, RNA expression signatures were used for risk stratifi cation and prognosis in breast cancer for more than 39,000 “treat” versus “no-treat” decisions (Securities and Exchange Commission, 2009). However, there continues to be a need to develop the

1

2 CHAPTER 1 ■■ The Foundations of Genomic and Personalized Medicine

BOX 1.1 Genetics and Genomics

Throughout this and the many other chapters in this Essentials volume, the terms “genetics” and “genomics” are used repeatedly, both as nouns and in their adjectival forms. While these terms seem similar, they in fact describe quite distinct (though frequently overlapping) approaches in biology and in medicine. Having said that, there are inconsistencies in the way the terms are used, even by those who work in the field. To some, genetics is a subfield of genomics; to others, genomics is a subfield of genetics. Arguably, depending on the perspective one has in mind, both may be right!

Here, we provide operational definitions to distinguish the vari-ous terms and the subfields of medicine to which they contribute.

The field of genetics is the scientific study of heredity and of the genes that provide the physical, biological, and conceptual bases for heredity and inheritance. To say that something – a trait, a disease, a code, or information – is genetic refers to its basis in genes and in DNA.

Heredity refers to the familial phenomenon whereby traits (including clinical traits) are transmitted from generation to gen-eration, due to the transmission of genes from parent to child. A disease that is said to be inherited or hereditary is certainly genetic; however, not all genetic diseases are hereditary (witness cancer, which is always a genetic disease, but is only occasionally an inherited disease).

Genomics is the scientific study of a genome or genomes. A genome is the complete DNA sequence, containing the entire genetic information of a gamete, an individual, a population, or a species. As such, it is a subfield of genetics when describing an approach taken to study genes. The word “genome” originated as an analogy with the earlier term “chromosome,” referring to the physical entities (visible under the microscope) that carry genes from one cell to its daughter cells or from one generation to the next. Genomics gave birth to a series of other “-omics” that refer to the comprehensive study of the full complement of genome prod-ucts – for example, proteins (hence, proteomics), transcripts (transcrip-tomics), or metabolites (metabolomics). The essential feature of the “omes” is that they refer to the complete collection of genes or their derivative proteins, transcripts, or metabolites, not just to the study of individual entities. While formally the field of genomics refers to the study of genomes (and hence, DNA) only, it some-times takes on the broader meaning of referring to any large-scale approach; the less specific term “genome sciences” is also sometimes used to refer to all of the -omics to connote global and compre-hensive approaches to the study of biology and medicine.

By analogy with genetics and genomics, epigenetics and epigenom-ics refer to the study of factors that affect gene (or, more globally, genome) function, but without an accompanying change in genes or the genome. Some typical epigenetic factors involve changes in DNA methylation or modifications to chromatin that change genome structure and hence influence gene expression even in the absence of changes in the DNA sequence. The epigenome is the comprehensive set of epigenetic changes in a given individual, tissue, tumor, or population. It is the paired combination of the

genome and the epigenome that best characterize and determine one’s phenotype.

Medical genetics is the application of genetics to medicine with a particular emphasis on inherited disease. Medical genetics is a broad and varied field, encompassing many different subfields, includ-ing clinical genetics, biochemical genetics, cytogenetics, molecular genetics, the genetics of common diseases, and genetic counseling. Medical genetics is one of 24 medical specialties recognized by The American Board of Medical Specialties, the preeminent medical organization overseeing physician certification in the United States. As of 2007, there were approximately 2300 board-certified medical geneticists in the United States.

Genetic medicine is a term sometimes used to refer to the appli-cation of genetic principles to the practice of medicine and thus overlaps medical genetics. However, genetic medicine is somewhat broader, as it is not limited to the specialty of Medical Genetics but is relevant to health professionals in many, if not all, specialties and subspecialties. Both medical genetics and genetic medicine approach clinical care largely through consideration of individual genes and their effects on patients and their families.

By contrast, genomic medicine refers to the use of large-scale genomic information and to the consideration of the full extent of an individual’s genome, proteome, transcriptome, metabolome, and/or epigenome in the practice of medicine and medical decision-making. The principles and approaches of genomic medicine are relevant well beyond the traditional purview of medical genetics and include, as examples, gene expression profiling to characterize tumors or to define prognosis in cancer, genotyping variants in the set of genes involved in drug metabolism or action to determine an individual’s correct therapeutic dosage, scanning the entire genome for millions of variants that influence one’s susceptibility to disease, or analyzing multiple protein biomarkers to monitor therapy and to provide predictive information in presymptomatic individuals.

Finally, personalized medicine refers to a rapidly advancing field of health care that is informed by each person’s unique clinical, genetic, genomic, and environmental information. The goals of personalized medicine are to take advantage of a molecular understanding of dis-ease to optimize preventive health care strategies and drug therapies while people are still well or at the earliest stages of disease. Because these factors are different for every person, the nature of disease, its onset, its course, and how it might respond to drug or other inter-ventions are as individual as the people who have them. In order for personalized medicine to be used by health care providers and their patients, these findings must be translated into precision diag-nostic tests and targeted therapies. Since the overarching goal is to optimize medical care and outcomes for each individual, treatments, medication types and dosages, and/or prevention strategies may dif-fer from person to person – resulting in unprecedented customiza-tion of patient care.

The principles underlying genomic and personalized medicine and their applications to the practice of clinical medicine are presented throughout the chapters that comprise this volume.

evidence that these signatures will enhance outcomes across a broad population of women with breast cancer. A prospective cooperative group clinical trial has been initiated in Europe (MINDACT) that aims to measure the effectiveness of a gene expression predictor of breast cancer prognosis to guiding

adjuvant chemotherapy when compared to predictions based solely on the traditional clinical parameters for prognoses (Bogaerts et al., 2006). A study sponsored by the US National Cancer Institute (TAILORx) aims to use an RNA signature to identify low risk breast cancer patients unlikely to benefi t from

Genomic and Personalized Medicine ■■ 3

chemotherapy (Sparano, 2006). Similar studies are getting started for prognostic signatures in lung cancer (Potti et al., 2006a). These are exemplars of the approach of genomic and person-alized medicine and serve as clear examples where a genome-based approach has resulted in the opportunity to redefi ne disease phenotypes and at the same time redefi ne therapeutic strategies.

Of equal if not greater importance in achieving the goal of personalized treatment is an ability to predict response to specifi ed therapies, particularly for those regimens that are part of routine clinical practice today. The selection of therapy for many cancer patients is still largely empiric and guided by large randomized clinical trials on populations of patients. Estimates of benefi ts from this approach for individuals are extrapolations from the effects seen in these large trials and do not necessarily apply to individual patients.

Genomic signatures that predict response and resistance to a spectrum of cytotoxic chemotherapies may now allow assign-ment of patients to effective treatment regimens best suited to the unique characteristics of their tumor (Potti et al., 2006b). Genomic predictors of chemotherapy response thus provide an opportunity to determine which drug would be optimal for an individual patient in clinical scenarios for which past stud-ies have not shown a clear superiority for any of the currently available drugs.

Beyond cancer, other fi elds of medicine are also benefi ting from whole-genome approaches that are defi ning both suscepti-bility to complex disease through genome-wide association studies (GWAS) or genome resequencing (see below), as well as signa-tures that defi ne disease states and predictive outcomes based on analyses from both disease tissues and from blood (Altshuler et al.,

2008; Aziz et al., 2007). Blood-based gene expression profi ling is particularly important as it presents the opportunity to report on disease processes from remote and often inaccessible sites for direct analyses. Instead of analyzing single genes, global gene expression provides a “molecular signature” that may distinguish between one disease state and another. In addition to identifying signatures or patterns of gene expression that represent a disease state, analyses can be constructed to identify representative path-way genes that might point to novel pathophysiology relevant to the underlying disease state. Peripheral blood gene expres-sion signatures have now been reported in a variety of condi-tions described in subsequent chapters, including rheumatoid arthritis (Lequerré et al., 2006; Shou et al., 2006), systemic lupus erythematosis (Rus et al., 2004), multiple sclerosis (Bomprezzi et al., 2003; Singh et al., 2007), asthma (Brutsche et al., 2002), malig -nancies (Alizadeh et al., 2000; Golub et al., 1999), solid organ transplantation (Baron et al., 2007; Horwitz et al., 2004), as well as environmental exposures (Dressman et al., 2007; Lodovici et al., 2007; Meadows et al., 2008; Wu et al., 2003). Many of these con-ditions have an infl ammatory component and thus affect immune cells in the vascular compartment. It is hypothesized that these cellular changes are the basis for the differences in gene expres-sion that is observed in RNA extracted from whole blood speci-mens or from specifi c circulating cell types. The greatest potential of this approach would be the enhancement in accurately clas-sifying patients by the type and severity of their disease and to individualize the therapy based on the biology of the disease in an individual patient. Signifi cant power can be anticipated from approaches that combine both assessment of an individual’s con-stitutional genome and evaluation of gene expression signatures of one’s health status (Figure 1.1) (Hardy and Singleton, 2009).

Genome sequence

Transcriptome

Proteome

Phenotype

• Diagnosis• Natural history• Treatment options• Prognosis

• Predisposition todisease

Individual BIndividual A

Somatic mutation

Epigenetic changes

Personal environment

Figure 1.1 The promise of personal genomics in the era of genomic and personalized medicine.

4 CHAPTER 1 ■■ The Foundations of Genomic and Personalized Medicine

GENES, GENOMES, AND DISEASEIn the context of genomic and personalized medicine, a key question is to what extent genetic variation infl uences the like-lihood of disease onset; determines or signals the natural history of disease; and/or provides clues relevant to the management of disease. Variation in one’s constitutional genome can have a number of different direct or indirect effects on gene expres-sion, thus contributing to the likelihood of disease (Frazer et al., 2009). It is not, however, just the human genome whose vari-ation is relevant to an individual’s state of health, but there are thousands of microorganisms, both symbiotic and pathogenic, whose genomes are also relevant to human phenotypes, and sequence determination of their genomes is providing new insights and approaches for the diagnosis, study, and treatment of infectious disease (see Box 1.2).

Contemporary frameworks for considering the impact of variation in the human genome on disease build on decades of success in establishing the role of individual, typically rare, mutations as causal determinants of now more than 2000 sim-ple Mendelian diseases (Online Mendelian Inheritance in Man, 2009). As a result of that success, much of which paralleled the development of technologies in the early stages of the Human Genome Project (Altshuler et al., 2008; Peltonen and McKusick, 2001), attention has now turned to the genes presumed to underlie susceptibility to common complex diseases, which are the subject of many of the subsequent chapters in this book.

There are two distinct, but nonexclusive models for think-ing about human genetic variation and disease (Altshuler and Clark, 2005; Fearnhead et al., 2005; Florez et al., 2003; Pritchard,

2001). One – the “common allele, common disease” hypothesis – posits that variation common in the population accounts for the relatively higher or lower risk that some individuals (and their families) have for a particular condition. Under this model, the collection of 10–15 million common variants in the genome (see Chapter 2) underlies the range of susceptibility that one fi nds in the general population, modulated by the particular and often variable environmental inputs and factors that are present in that population and that may, in fact, shift or even obscure the relative impact of inherited factors. An alternative model – the rare variant hypothesis – argues that genetic susceptibility to dis-ease is due to the accumulated risk conferred by multiple rare variants in an individual’s genome, variants therefore not likely to be captured by study of the common variants identifi ed by studies to date (Goldstein, 2009).

These two hypotheses suggest different approaches that will likely be informative for delineating the genetic contribution to disease, both for designing research studies and for eventual clin-ical surveillance. It is worth emphasizing that these two hypoth-eses are not mutually exclusive and are each likely to be correct in some cases; indeed, there is already evidence supporting each for different diseases (Frazer et al., 2009).

Genome-Wide Association StudiesThe common allele, common disease hypothesis has been explored with notable success in a number of conditions, uti-lizing large cohorts of well-phenotyped patients and high-throughput methods to genotype up to 500,000 or a million variants (so-called single nucleotide polymorphisms [SNPs], see Chapter 2) in the genome (Manolio et al., 2008). These

BOX 1.2 The Genomes Within

The human genome is not the only genome relevant to the practice of medicine.

Both in states of health and disease, our own genome is vastly outnumbered by the genomes of a host of microorganisms, many living peacefully and continuously on various body surfaces, espe-cially throughout the gastrointestinal tract, others wreaking havoc as adventitious viral, bacterial, or fungal pathogens.

The genomes of thousands of microorganisms have been deter-mined and are being utilized to provide rapid diagnostic tests in clinical settings, to predict antibiotic or antifungal efficacy, to iden-tify the source of airborne, water, or soil contaminants, to moni-tor hospital or community environments, and to better understand the contribution of microbial ecosystems and various environmental exposures to diverse human phenotypes.

The human colon contains more than 400 bacterial species com-prising some 1013–1014 microorganisms. Each adult’s gut provides a unique environment – the microbiome – whose origins and impact on human disease are just being explored. The microbiotic gene set is significantly different from that of the human genome and thus has the capacity to alter the metabolic profile of different individuals or different populations, with clinically meaningful effects on drug metabolism, toxicity, and efficacy (Gill et al., 2006; Li et al., 2008; Palmer et al., 2007). The applications to microbiomes of approaches

in genomics (as well as proteomics and transcriptomics) are revolu-tionizing clinical diagnostics, for example, to identify unknown viral infections (Delwart, 2007; Long et al., 2004; Wang et al., 2003) or to diagnose antibiotic-resistance infections such as methicillin-resistant Staphylococcus aureus (MRSA) (Francois et al., 2007).

The field of metagenomics explores this heterogeneous ecosystem by comprehensive sequence analysis of the collected genomes from biological specimens (such as stool, urine, sputum, water sources, and air), followed by both taxonomic and bioinformatic analysis to deconvolute the many genomes contained in such specimens and to define the different organisms, their genes, and genome vari-ants. This approach is particularly informative for characterizing organisms that cannot be cultured in standard microbiology labs. A number of diseases have been associated with large-scale imbalances in the gut microbiome, including Crohn’s disease, ulcerative colitis, antibiotic-resistant diarrhea, and obesity (Frank and Pace, 2008; Ley et al., 2006; Turbaugh et al., 2006).

Undoubtedly, the states of health and disease are determined in part by the balance of genomes both within us and external to us. The full complement of genomic information from both of these sources of genomes will provide insights into defining the states of health and disease and the basis for unsurpassed precision in both the prevention of disease and its treatment.

From the Genome to Personalized Medicine ■■ 5

GWAS report the statistical association of one or more vari-ants in a narrow genomic region (which may or may not contain an annotated gene) with the presence or absence of the clinical condition. The reported SNPs defi ne immedi-ately accessible risk factors for that condition, at least in the population(s) under study and can provide novel insights into the biology of the disease. It should be stressed, however, that in most instances, causality of the reported SNP(s) and the increased risk has not been proved; it may be that the actual causal variant is not the SNP itself but is a currently undetec-ted variant that lies in linkage disequilibrium with the SNP (Frazer et al., 2009).

In the most favorable cases, the associated SNP may be a nonsynonymous variant, leading to a pathological amino acid change in the relevant gene (Thorleifsson et al., 2007), or may be a variant in an RNA splice site, leading to a clini-cally meaningful change in the production of the gene’s transcript(s) (Heinzen et al., 2007). But in most instances, the functional impact of the associated SNP is obscure, notwith-standing very clear genetic evidence of a role of genome vari-ation in susceptibility to the particular condition. SNPs are but one of several types of genome variation that can infl u-ence gene expression and/or disease (see Chapter 2). Copy number variants have also been associated with some disorders (reviewed in Estivill and Armengol, 2007), and it will require integrated genomic, genetic and functional studies to eluci-date the precise basis for the role(s) of genome variation in different diseases.

Medical Resequencing in Search of Rare VariantsWhile GWAS certainly establish that common alleles do indeed increase susceptibility to common disease in some instances, they do not allow one to conclude that this is always the case. Indeed, in most (but not all) cases to date, the com-mon SNPs found to be associated with disease only explain a small fraction of the total genetic variation, implicating an as yet undiscovered (and presumably rarer) basis for most genetic variation underlying a given condition. While such asso-ciation is suffi cient to conclude that a particular variant does indeed contribute to disease in the population under study, such a fi nding is insuffi cient to say anything declarative about the cause of our likelihood of disease in a particular case. This conclusion is, of course, highly relevant to the prospects of genomic and personalized medicine.

An alternative or complementary approach to genome-wide genotyping of common variants is to resequence specifi c genes in a cohort of affected individuals, in an effort to uncover rare variants responsible for (or at least statistically associated with) the disease in question. To date, most efforts have focused on one or several genes that were believed to be strong candi-dates for the phenotype understudy; the notable exception thus far is exemplifi ed by the whole-genome resequencing of the HuRef genome and correlation of novel variants detected in that study with his family and personal medical histories (Levy et al., 2007). Rare variants, including nonsynonymous variants,

in relevant candidate genes have been detected at a statistically signifi cant higher frequency in the genomes of patients with colorectal adenomas (Fearnhead et al., 2004), with low-plasma high-density lipoprotein cholesterol (HDL-C) (Cohen et al., 2004), with triglyceride levels in the lowest quartile (Romeo et al., 2007), and with familial cases of X-linked mental retardation (Tarpey et al., 2009).

These successes point to a strategy of resequencing relevant genes in individuals at the extremes of the population distri-bution for measurable traits (Topol and Frazer, 2007) in which the only limiting parameters are the cost of sequencing and the quality of the phenotypic or quantitative data. In a proof-of-principle for this case-control approach, a recent study sequenced coding exons and splice junctions of 58 genes in nearly 400 obese and lean individuals, at the �95th or �10th percentile of body mass index (Ahituv et al., 2007). Of the �1000 variants detected, most were rare variants, including over 270 nonsynonymous mutations; many were found only in the obese cohort and thus become strong functional candidates for a role in obesity.

Searching for Somatic MutationsWhile GWAS are restricted to inherited variation, medical rese-quencing studies can target either inherited or somatic variants. In cancer especially, it is of interest to use medical resequenc-ing to search for somatic mutations in tumor tissue in order to identify genes potentially relevant to cancer development and/or progression (Stratton et al., 2009). Two important points emerge from such studies (Greenman et al., 2007; Jones et al., 2009; Sjoblom et al., 2006). First, the genes implicated by virtue of discovering rare somatic variants in multiple cases of a par-ticular cancer tend to be different from those identifi ed in previ-ous genetic studies as inherited risk factors. This provides novel insights into the biology of human cancer and suggests candidates for further exploring mechanisms of tumorigenesis or metastasis or for developing therapeutic approaches. Second, however, the large number of mutations uncovered by this approach intro-duces the need for caution, as many will be “bystander” or “pas-senger” mutations only associated with cancer, not genes involved directly in cancer.

FROM THE GENOME TO PERSONALIZED MEDICINEOf all the promises of the current scientifi c and social revolution stemming from advances in our understanding of the human genome and its variation, genomic and personalized medicine has been the most eagerly awaited. The prospect of examining an individual’s entire genome (or at least a signifi cant fraction of it) in order to make individualized risk predictions and treat-ment decisions is an attractive, albeit challenging, one (Bentley, 2004; Kraft and Hunter, 2009; Willard et al., 2005).

Having access to the reference human genome sequence has been transformational for the fi elds of human genetics and

6 CHAPTER 1 ■■ The Foundations of Genomic and Personalized Medicine

genome biology but by itself is an insuffi cient prerequisite for genomic medicine. As alluded to above and discussed in subse-quent chapters, equally important are the various complemen-tary technologies to reliably capture information on individual genomes, their epigenetic modifi cation, and their derivatives in the transcriptome, proteome, and metabolome for health and disease status (Figure 1.1). Each of these technologies provides information that, in combination with clinical data and evalua-tion of environmental triggers, will in time contribute to assess-ment of individual risks and guide clinical management and decision-making. Critical enablers of this new approach have been innovations in laboratory technology (to address bio-logically and medically relevant questions on a scale and with a throughput hardly imaginable just a few decades ago), paired with equally transforming developments in informatics and information systems to handle the onslaught of genomic data (West et al., 2006).

Personal GenomicsAt the heart of the genomic approach to personalized medicine will be information from individual genomes, a fast-moving area of technological development that is spawning a social and information revolution among consumers. Dramatic improve-ments in sequencing technology (Bentley, 2006; Mardis, 2008) have reduced the cost and time of resequencing projects to a level that invites conjecture about the long awaited “$1000 genome” (Dalton, 2006; Service, 2006; Wolinsky, 2007). The much-ballyhooed release of the genome sequences of Craig Venter (“HuRef”) (Levy et al., 2007) and James Watson (Wheeler et al., 2008) has stirred up concerns about “celebrity genomics” (Check, 2007); however, while the Venter and Watson sequences may have been the fi rst, they have been rapidly fol-lowed by several others (see Chapter 2), and numerous addi-tional genomes are already in various sequencing pipelines.

What remains unsettled for now is what degree of genome surveillance will be most useful, either for research or for clinical practice, a topic that is discusssed frequently in subsequent chap-ters. Some have argued that, while whole-genome sequencing is increasingly possible, it is unlikely to provide more information about established disease associations than would high-density, genome-wide genotyping to detect both SNP and copy number variation. Targeted resequencing of, for example, exons and known regulatory regions would allow detection of rare variants in portions of the genome most relevant to disease at a fraction of the cost of whole-genome sequencing (Hodges et al., 2007).

The availability of associated clinical data is variable among the studies announced to date, but there exists, at least among some participants, a strong sense of “health-information altru-ism” to contribute to the much needed large-scale correla-tion of genotype and phenotype (Kohane and Altman, 2005). Notwithstanding individual’s willingness to make genome sequence data (much less medical information) available more-or-less publicly, substantial concerns have been raised about pri-vacy, since a surprisingly (to some) small number of SNPs or other genome variants are suffi cient to allow identifi cation of

individuals (Lin et al., 2004; McGuire and Gibbs, 2006; McGuire et al., 2008). Absolute privacy and anonymity may be an impos-sible standard.

A Consumer RevolutionWhile the genome revolution has without doubt been driven by technological improvements and by an explosion in the avail-ability of genome data, the push for incorporation of genome information into clinical practice may come as much or more from consumers as from professionals. A half dozen or more companies are already offering genome-wide SNP profi les to the public, some with associated risk estimates for relevant clini-cal conditions. At least a few companies will also sequence indi-vidual genomes for a cost that, while high currently, is not out of reach for some individuals. There is a clear and important research agenda that needs to be developed in concert with these technological breakthroughs that allows health providers and the public to understand the information and more so to believe that is accurate, informative, and actionable (Feero et al., 2008; Hunter et al., 2008; McGuire et al., 2007, 2008). With the vast amount of information contained in the human genome sequence, the stakes are high for patients, physicians, and the public to ensure proper reading, interpretation, and communi cation of the infor-mation are carried out (Janssens et al., 2008). This may be a “dis-ruptive technology” in health care delivery with the provision of health and disease risk information to consumers without physician intervention and guidance. It will not be long before a patient will bring a report of a whole genome to a physician’s offi ce and ask for guidance. What will the physicians of today, armed with a paucity of genomic training, tell them?

CHALLENGES IN THE TRANSLATION OF GENOMICS TO HUMAN HEALTHDespite the clear advances in technology to bring genomic information closer to physicians, patients, and the public, loom-ing ever closer are issues that are outside of the sphere of the scientists that have been involved in the discovery and early translational activities. In the United States, the Institute of Medicine has convened a Roundtable on Translating Genomics to Health (Institute of Medicine, 2007), the Centers for Disease Control has independently developed a pilot project on the Evaluation of Genomic Applications in Practice and Prevention (EGAPP, 2004), and the Genomic and Personalized Medicine Act of 2007 (Personalized Medicine, 2007) was introduced in Congress to also address similar issues.

Recently, Scheuner et al. (2008) carried out an extensive meta-analysis of studies using genomics toward clinical appli-cation in chronic disease. This study aimed to understand the current state of translation focusing on the following questions: “What are the outcomes of genomic medicine? What is the current level of consumer understanding about genomic medi-cine and what information do consumers need before they seek services? How is genomic medicine best delivered? What are the

The Future of Personalized Medical Care ■ 7

challenges and barriers to integrating genomic medicine into clinical practice?”

Using a total of 68 articles in their analysis, the authors syn-thesized information on the delivery of genomic medicine for common adult-onset conditions (Scheuner et al., 2008). The major fi ndings of the study were not surprising in terms of the genome policy issues that have been previously summarized (Haga and Willard, 2006).

● Education – The primary care workforce feels woefully unprepared to integrate genomics into regular practice. Consumers are also unclear about genetics and genetic test-ing for common diseases.

● Privacy – Consumers are worried about the possible adverse consequences of genetic testing, particularly the privacy issues and discrimination against receiving employ-ment and health insurance.

● Evidence – There needs to be outcome data for genetic testing and chronic disease to assess whether patients who receive the test do better clinically.

● Cost – Cost uncertainty (both in terms of delivery and reimbursement) is an important issue to many of the stake-holders of genomic issues.

Health Professional and Public EducationEducation of health professionals and the public will be essential to advancing the use of genomics into health care (Frueh et al., 2005). With all of the rapid advancements in genomics research and technologies, it will be challenging to keep health profes-sionals informed about the benefi ts, risks, and limitations of new tools as they become available. In addition, the public and health care workforce will need to understand the appropriate applica-tions of genomic tools, including their benefi ts, risks, and limi-tations; how they may improve clinical management; inherited versus acquired genomic variations (e.g., implications for family members); and privacy and confi dentiality. Although several sur-veys have documented the below average physician knowledge of genetics (Metcalfe et al., 2002), none has assessed knowledge of the newer fi eld of genomics. But, several papers have been published recognizing the importance of pharmacogenetics (Frueh and Gurwitz, 2004), and steps are underway to develop more educational materials in this area.

Privacy FearsThere has been ongoing debate about the uniqueness of genetic information and whether it warrants special protections beyond those in place for standard medical information (Haga and Willard, 2006). In the United States, fear of discrimination by employers and health insurers is the main concern, whereas in the United Kingdom, use of genetic information by life insurers is the major concern (Apse et al., 2004; Hall et al., 2005). Despite the outcome of these debates, the attention paid to genetics by the popular press and public has raised concern about genetic information. The potential for genetic discrimination has been a major concern for researchers, health professionals, patients, and

the public. In order for genomic biomarkers to be integrated into routine clinical practice, associated fears with this type of testing must be put to rest. In 2008, the United States enacted legislation to protect against genetic discrimination by employ-ers and health insurers. This is an important step in allaying pub-lic fears that may otherwise hinder a comprehensive genomic and personalized approach to medicine.

Building the Evidence for Clinical UtilityPerhaps the most important factor hindering the appropriate integration of genomics into clinical practice is the lack of evi-dence for its clinical utility (i.e., evidence that use of a genomic technology actually improves health outcomes). Evidence gener-ation needs to be more practicable and practical. There is also a need for greater collaboration among stakeholder groups and for innovation in both study design and analysis methods. Clinical outcome studies are needed that demonstrate the clinical utility of genomic interventions that are linked to specifi c, actionable clinical recommendations, or practice guidelines. Public–private partnerships are likely to be required to generate the evidence base for genomic medicine. These collaborations are desirable because no signal stakeholder group is likely to have suffi cient resources or expertize to conduct the necessary studies.

Cost IssuesAs with any new innovation, genomic testing must be demon-strated to be clinically useful and cost effective and of value. But, because genomic technologies inherently involve diagnostic or prognostic testing, in addition to the complexities of incomplete gene penetrance and multiple gene � environment interactions, their assessment can be more challenging. In addition, perhaps more than in any other area of medicine, questions have arisen in regard to the economic incentives to develop these technologies. Formal health economics frameworks can be used to gain insights into these issues and provide guidance for research and develop-ment (Carlson et al., 2005; Flowers and Veenstra, 2004; Stallings et al., 2006). It is important to examine the drivers of cost effec-tiveness of genomic technologies and to consider approaches that include value-based reimbursement for genomic testing tech-nologies. A particularly challenging area is pharmacogenomics, in which the economic incentives for developing diagnostics linked to therapeutics in the pharmaceutical industry are unclear. An inte-grated business model is needed that will be favorable for the effec-tive delivery of genomic information to patients and clinicians.

THE FUTURE OF PERSONALIZED MEDICAL CAREWhile the human genome sequence is now available, it is impor-tant to acknowledge that our knowledge of the genome and its biological complexity is nowhere near complete, and the use of genomic protocols in standard clinical care faces many challenges. There are a host of clinical, economic, insurance, privacy, and

8 CHAPTER 1 ■■ The Foundations of Genomic and Personalized Medicine

commercialization concerns that will need to be addressed and that vary substantially among different countries. As a fi eld, we can confront those with the certainty that the science behind genomic medicine is sound and the practice of medicine that it informs

is evidence based. These issues are being dealt with systematically, and the prospects of using genomic information to offer patients health care that is truly personalized in nature – as will clearly see in the subsequent chapters – is fi nally within our reach.

REFERENCES

Ahituv, N., Kavaslar, N., Schackwitz, W., Ustaszewska, A., Martin, J., Hebert, S. Doelle, H., Ersoy, B., Kryukov, G., Schmidt, S. et al. (2007). Medical sequencing at the extremes of human body mass. Am J Hum Genet 80, 779–791.

Altshuler, D. and Clark, A. (2005). Harvesting medical information from the human family tree. Science 307, 1052–1053.

Altshuler, D., Daly, M. and Lander, E. (2008). Genetic mapping in human disease. Science 322, 881–888.

Alizadeh, A., Eisen, M., Davis, R., Ma, C., Lossos, I., Rosenwald, A., Boldrick, J., Sabet, H., Tran, T., Yu, X. et al. (2000). Distinct types of diffuse large B-cell lymphoma identifi ed by gene expression pro-fi ling. Nature 403, 503–511.

Apse, K., Biesecker, B., Giardiello, F., Fuller, B. and Bernhardt, B. (2004). Perceptions of genetic discrimination among at-risk relatives of colorectal cancer patients. Genet Med 6, 510–516.

Aziz, H., Zaas, A. and Ginsburg, G.S. (2007). Peripheral blood gene expression profi ling for cardiovascular disease assessment. Genomic Med 1, 105–112.

Baron, C., Somogyi, R., Greller, L., Rineau, V., Wilkinson, P., Cho, C., Cameron, M., Kelvin, D., Chagnon, R., Roy, D. et al. (2007). Prediction of graft-versus-host disease in humans by donor gene-expression profi ling. PLoS Med 4(1), e23.

Bentley, D.R. (2004). Genomes for medicine. Nature 429, 440–445.Bentley, D.R. (2006). Whole-genome resequencing. Curr Opin Genet

Devel 16, 545–552.Brutsche, M., Joos, L., Carlen, B., Bissinger, R., Tamm, M., Custovic, A.

and Woodcock, A. (2002). Array-based diagnostic gene-expression score for atopy and asthma. J Allergy Clin Immunol 109, 271–273.

Bomprezzi, R., Ringner, M., Kim, S., Bittner, M., Khan, J., Chan, Y., Elkahloun, A., Yu, A., Bielekova, B., Meltzer, P. et al. (2003). Gene expression profi le in multiple sclerosis patients and healthy con-trols: Identifying pathways relevant to disease. Hum Mol Genet 12, 2191–2199.

Bogaerts, J., Cardoso, F., Buyse, M., Braga, S., Loi, S., Harrison, J., Bines, J., Mook, S., Decker, N., Ravdin, P. et al. (2006). TRANSBIG con-sortium. Gene signature evaluation as a prognostic tool: challenges in the design of the MINDACT trial. Nat Clin Pract Oncol 10, 540–551.

Bullinger, L. and Valk, P.J. (2005). Gene expression profi ling in acute myeloid leukemia. J Clin Oncol 23, 6296–6305.

Carlson, J., Henrikson, N., Veenstra, D. and Ramsey, S. (2005). Economic analyses of human genetics services: A systematic review. Genet Med 7, 519–523.

Check, E. (2007). Celebrity genomes alarm researchers. Nature 447, 358–359.

Cohen, J., Kiss, R., Pertsemlidis, A., Marcel, Y., McPherson, R. and Hobbs, H. (2004). Multiple rare alleles contribute to low plasma levels of HDL cholesterol. Science 305, 869–871.

Dalton, R. (2006). Sequences step up to the speed challenge. Nature 443, 258–259.

Dave, S.S., Fu, K., Wright, G.W., Lam, L., Kluin, P., Boerma, E., Greiner, T., Weisenburger, D., Rosenwald, A., Ott, G. et al. (2006). Lymphoma/Leukemia Molecular Profi ling Project. Molecular diagnosis of Burkitt’s lymphoma. N Engl J Med 354, 2431–2442.

Delwart, E.L. (2007). Viral Metagenomics. Rev Med Virol 17, 115–131.Dressman, H., Muramoto, G., Chao, N., Meadows, S., Marshall, D.,

Ginsburg, G., Nevins, J. and Chute, J. (2007). Gene expression sig-natures that predict radiation exposure in mice and humans. PLoS Med 4, e106.

EGAPP. (2004). http://www.egappreviews.org/about.htmEstivill, X. and Armengol, L. (2007). Copy number variants and com-

mon disorders: Filling the gaps and exploring complexity in genome-wide association studies. PLoS Genet 3, e190.

Fearnhead, N.S., Wilding, J., Winney, B., Tonks, S., Bartlett, S., Bicknell, D., Tomlinson, I., Montensen, N.J. and Bodmer, W.F. (2004). Multiple rare variants in different genes account for multifactorial inherited susceptibility to colorectal adenomas. Proc Natl Acad Sci USA 101, 15992–15997.

Fearnhead, N.S., Winney, B. and Bodmer, W.F. (2005). Rare variant hypothesis for multifactorial inheritance: Susceptibility to colorec-tal adenomas as a model. Cell Cycle 4, 521–525.

Feero, W., Guttmacher, A. and Collins, F. (2008). The genome gets per-sonal – Almost. J Am Med Assoc 299, 1351–1352.

Florez, J., Hirschhorn, J. and Altshuler, D. (2003). The inherited basis of diabetes mellitus: Implications for the genetic analysis of complex traits. Annu Rev Genomics Hum. Genet 4, 257–291.

Flowers, C.R. and Veenstra, D. (2004). The role of cost-effectiveness anal-ysis in the era of pharmacogenomics. Pharmacoeconomics 22, 481–493.

Francois, P., Scherl, A., Hochstrasser, D. and Schrenzel, J. (2007). Proteomic approach to investigate MRSA. Methods Mol Biol 391, 179–199.

Frank, D.N. and Pace, N.R. (2008). Gastrointestinal microbiology enters the metagenomics era. Curr Opin Gastroenterol 24, 4–10.

Frazer, K., Murray, S., Schork, N. and Topol, E. (2009). Human genetic variation and its contribution to complex traits. Nat Rev Genet 10, 241–251.

Frueh, F.W. and Gurwitz, D. (2004). From pharmacogenetics to per-sonalized medicine: A vital need for educating health professionals and the community. Pharmacogenomics 5, 571–579.

Frueh, F.W., Goodsaid, F., Rudman, A., Huang, S.M. and Lesko, L.J. (2005). The need for education in pharmacogenomics: A regula-tory perspective. Pharmacogenomics J 5, 218–220.

Gill, S.R., Pop, M., DeBoy, R., Eckburg, P.B., Turnbaugh, P.J., Samuel, B.S., Gordon, J.I., Relman, D.A., Fraser-Liggett, C.M. and Nelson K.E. (2006). Metagenomic analysis of the human distal gut micro-biome. Science 312, 1355–1359.

Goldstein, D.B. (2009). Common genetic variation and human traits. N Engl J Med 360, 1696–1698.

Golub, T., Slonim, D., Tamayo, P., Huard, C., Gaasenbeek, M., Mesirov, J., Coller, H., Loh, M., Downing, J., Caliguri, M. et al. (1999). Molecular

References ■■ 9

classifi cation of cancer: Class discovery and class prediction by gene expression monitoring. Science 286, 531–537.

Greenman, C., Stephens, P., Smith, R., Dalgliesh, G., Hunter, C., Bignell, G., Davies, H., Teague, J., Stevens, C., Edkins, S. et al. (2007). Patterns of somatic mutation in human cancer genomes. Nature 446, 153–158.

Haga, S. and Willard, H. (2006). Defi ning the spectrum of genome pol-icy. Nat Rev Genet 7, 966–972.

Hall, M., McEwen, J., Barton, J., Walker, A., Howe, E., Reis, J., Power, T., Ellis, S., Tucker, D., Harrison, B. et al. (2005). Concerns in a primary care population about genetic discrimination by insurers. Genet Med 7, 311–316.

Hardy, J. and Singleton, A. (2009). Genomewide association studies and human disease. N Engl J Med 360, 1759–1768.

Heinzen, E., Yoon, W., Tate, S., Sen, A., Wood, N., Sisodiya, S. and Goldstein, D. (2007). Nova2 interacts with a cis-acting polymor-phism to infl uence the proportions of drug-responsive slice vari-ants of SCN1A. Am J Hum Genet 80, 876–883.

Hodges, E., Xuan, Z., Balija, V., Kramer, M., Molla, M., Smith, S., Middle, C., Rosesch, M., Albert, T., Hannon, G. et al. (2007). Genome-wide in situ exon capture for selective resequencing. Nat Genet 39, 1522–1527.

Horwitz, P., Tsai, E., Putt, M., Gilmore, J., Lepore, J., Parmacek, M., Kao, A., Desai, S., Goldberg, L., Brozena, S. et al. (2004). Detection of cardiac allograft rejection and response to immunosuppressive therapy with peripheral blood gene expression. Circulation 110, 3815.

Hunter, D., Khoury, M. and Drazen, J. (2008). Letting the genome out of the bottle – Will we get our wish? N Engl J Med 358, 105–107.

Institute of Medicine. (2007). http://www.iom.edu/CMS/3740/44443.aspx

Janssens, A., Gwinn, M., Bradley, L., Oostra, B., van Duijn, C. and Khoury, M. (2008). A critical appraisal of the scientifi c basis of commercial genomic profi les used to assess health risks and per-sonalize health interventions. Am J Hum Genet 82, 593–599.

Jones, S., Hruban, R.H., Kamiyama, M., Borges, M., Zhang, X., Parsons, D., Lin, J., Palmisano, E., Brune, K., Jaffee, E. et al. (2009). Exonic sequencing identifi es PALB2 as a pancreatic cancer susceptibility gene. Science 324, 217.

Kohane, I. and Altman, R. (2005). Health-information altruists – A potentially critical resource. N Engl J Med 353, 2074–2077.

Kraft, P. and Hunter, D. (2009). Genetic risk prédiction – Are we there yet? N Engl J Med 360, 1701–1703.

Lequerré, T., Gauthier-Jauneau, A.-C., Bansard, C., Derambure, C., Hiron, M., Vittecoq, O., Daveau, M., Mejjad, O., Daragon, A., Tron, F. et al. (2006). Gene profi ling in white blood cells predicts inf-liximab responsiveness in rheumatoid arthritis. Arthritis Res Ther 8, R105.

Levy, S., Sutton, G., Ng, P., Feuk, L., Halpern, A., Walenz, B., Axelrod, N., Huang, J., Kirkness, E., Denisov, G. et al. (2007). The diploid genome sequence of an individual human. PLoS Biol 5, e254.

Ley, R.E., Turnbaugh, P.J., Klein, S. and Gordon, J.I. (2006). Human gut microbes associated with obesity. Nature 444, 1022–1023.

Li, M., Wang, B., Zhang, M., Rantalainen, M., Wang, S., Zhou, H., Zhang, Y., Shen, J., Pang, X., Zhang, M. et al. (2008). Symbiotic gut microbes modulate human metabolic phenotypes. Proc Natl Acad Sci USA 105, 2117–2122.

Lin, Z., Owen, A. and Altman, R. (2004). Genomic research and human subject privacy. Science 305, 183.

Lodovici, M., Luceri, C., Filippo, C., Romualdi, C., Bambi, F. and Dolara, P. (2007). Smokers and passive smokers gene expression profi les: Correlation with DNA oxidative damage. Free Rad Biol Med 43, 415–422.

Long, W.H., Ziao, H., Gu, X., Zhang, Q., Yang, H., Zhao, G. and Liu, J.H. (2004). A universal microarray for detection of SARS corona-virus. J Virol Meth 121, 57–63.

Manolio, T.A., Brooks, L.D. and Collins, F.S. (2008). A HapMap harvest of insights into the genetics of common disease. J Clin Invest 118, 1590–1605.

Mardis, E.R. (2008). Next-generation DNA sequencing methods. Ann Rev Genomics Hum Genet 9, 387–402.

McGuire, A. and Gibbs, R. (2006). No longer de-identifi ed. Science 312, 370–371.

McGuire, A., Cho, M., McGuire, S. and Caulfi eld, T. (2007). The future of personal genomics. Science 317, 1687.

McGuire, A., Caulfi eld, T. and Cho, M.K. (2008). Research ethics and the challenge of whole-genome sequencing. Nat Rev Genet 9, 152–156.

Meadows, S.K., Dressman, H.K., Muramoto, G.G., Himburg, H., Salter, A., Wei, Z., Ginsburg, G.S., Chao, N.J., Nevins, J.R. and Chute, J.P. (2008). Gene expression signatures of radiation response are specifi c, durable and accurate in mice and humans. PLoS ONE 3, e1912.

Metcalfe, S., Hurworth, R., Newstead, J. and Robins, R. (2002). Needs assessment study of genetics education for general practitioners in Australia. Genet Med 4, 71–77.

Nussbaum, R., McInnes, R. and Willard, H.F. (2007). Genetics in Medicine, 7th edition. WB Saunders Co., Philadelphia.

Online Mendelian Inheritance in Man. (2009). http://www.ncbi.nlm.nih.gov/sites/entrez?db�OMIM&itool�toolbar

Palmer, C., Bik, E.M., DiGiulio, D., Rleman, D.A., and Brown, P.O. (2007). Development of the human infant intestinal microbiota. PLoS Biol 5, e177.

Peltonen, L. and McKusick, V. (2001). Dissecting human disease in the postgenomic era. Science 291, 1224–1229.

Potti, A., Mukherjee, S., Petersen, R., Dressman, H., Bild, A., Koontz, J., Kratzke, R., Watson, M., Kelley, M., Ginsburg, G. et al. (2006a). A genomic strategy to refi ne prognosis in early-stage non-small-cell lung cancer. N Engl J Med 355, 570–580.

Potti, A., Dressman, H.K., Bild, A., Riedel, R.F., Chan, G., Sayer, R., Cragun, J., Cottrill, H., Kelley, M.J., Petersen, R. et al. (2006b). Genomic signatures to guide the use of chemotherapeutics. Nat Med 12, 1294–1300.

Personalized Medicine. (2007). http://olpa.od.nih.gov/tracking/110/senate_bills/session1/s-976.asp

Pritchard, J. (2001). Are rare variants responsible for susceptibility to complex diseases? Am J Hum Genet 69, 124–137.

Romeo, S., Pennacchio, L., Fu, Y., Boerwinkle, E., Tybjaerg-Hansen, A., Hobbs, H. and Cohen, J. (2007). Population-based resequencing of ANGPTL4 uncovers variations that reduce triglycerides and increase HDL. Nat Genet 39, 513–516.

Rus, V., Chenb, H., Zernetkinaa, V., Magderb, L.S., Mathaia, S., Hochberg, M. and Via, C. (2004). Gene expression profi ling in peripheral blood mononuclear cells from lupus patients with active and inactive disease. Clin Immunol 112, 231–234.

Securities and Exchange Commission. (2009). http://www.sec.gov/Archives/edgar/data/1131324/000095013409005338/f51197e10vk.htm

Scheuner, M., Sieverding, P. and Shekelle. P. (2008). Delivery of genomic medicine for common chronic adult diseases: a systematic review. J Am Med Assoc 299, 1320–1334.

10 CHAPTER 1 ■■ The Foundations of Genomic and Personalized Medicine

Service, R. (2006). The race for the $1000 genome. Science 311, 1544–1546.

Shou, J., Bull, C., Li, L., Qian, H., Wei, T., Luo, S., Perkins, D., Solenberg, P., Tan, S., Chen, X. et al. (2006). Identifi cation of blood biomarkers of rheumatoid arthritis by transcript profi ling of peripheral blood mononuclear cells from the rat collagen-induced arthritis model. Arthritis Res Ther 8, R28.

Singh, M.K., Scott, T.F., LaFramboise, W.A., Hu, F.Z., Post, J.C. and Ehrlich, G. (2007). Gene expression changes in peripheral blood mononuclear cells from multiple sclerosis patients undergoing α-interferon therapy. J Neurological Sciences 258, 52–59.

Sjoblom, T., Jones, S., Wood, L., Parsons, D., Lin, J., Barber, T., Mandelker, D., Leary, R., Ptak, J., Silliman, N. et al. (2006). The consensus cod-ing sequences of human breast and colorectal cancers. Science 314, 268–274.

Sparano, J.A. (2006). TAILORx: Trial assigning individualized options for treatment (Rx). Clin Breast Cancer 4, 347–350.

Stallings, S., Witt, W., Finkelstein, S., Hiller, A., Sinskey, A. and Ginsburg, G. (2006). A framework to evaluate the economic impact of pharma-cogenomics. Pharmacogenomics 7, 853–862.

Staudt, L.M. (2003). Molecular diagnosis of the hematologic cancers. N Engl J Med 348, 1777–1785.

Stratton, M.R., Campbell, P. and Futreal, P.A. (2009). The cancer genome. Nature 458, 719–724.

Tarpey, P., Smith, R., Pleasance, E., Whibley, A., Edkins, S., Hardy, C., O’Meara, S., Latimer, C., Dicks, E., Menzies, A. et al. (2009). A sys-tematic, large-scale resequencing screen of X chromosome coding exons in mental retardation. Nat genet 41, 535–543.

Thorleifsson, G., Magnusson, K., Sulem, P., Bragi Walters, G., Gudbjartsson, D., Stefansson, H., Jonsson, T., Jonasdottir, A., Jonasdottir, A. and Stefansdottir, G. (2007). Common sequence

variants in the LOXL1 gene confer susceptibility to exfoliation glaucoma. Science 317, 1397–1400.

Topol, E. and Frazer, K. (2007). The resequencing imperative. Nature Genet 39, 439–440.

Turnbaugh, P.J., Ley, R.E., Mahowald, M.A, Magrini, V., Mardis, E. and Gordon J.I. (2006). An obesity-associated gut microbiome with increased capacity for energy harvest. Nature 444, 1027–1031.

Valk, P., Verhaak, R., Beijen, M., Erpelinck, C., Barjesteh vWvD-K.S., Boer, J., Beverloo, H., Moorhouse, M., van der Spek, P., Lowenberg, B. et al. (2004). Prognostically useful gene-expression profi les in acute myeloid leukemia. N Engl J Med 350, 1617–1628.

Wang, D., Urisman, A., Liu, Y.T., Springer, M., Ksiazek, T.G., Erdman, D., Mardis, E., Hickenbotham, M., Magrini, V. and Eldred, J. (2003). Viral discovery and sequence recovery using DNA microarrays. PLoS Biol 1, e2.

West, M., Ginsburg, G., Huang, A. and Nevins, J. (2006). Embracing the complexity of genomic data for personalized medicine. Genome Res 16, 559–566.

Wheeler, D., Srinivasan, M., Egholm, M., Shen, Y., Chen, L., McGuire, A., He, W., Chen, Y., Makhijani, V., Roth, G. et al. (2008). The complete genome of an individual by massively parallel DNA sequencing. Nature 452, 872–876.

Willard, H.F., Angrist, M. and Ginsburg, G. (2005). Genomic medicine: Genetic variation and its impact on the future of health care. Philos Trans R Soc Lond B Biol Sci 360, 1543–1550.

Wolinsky, H. (2007). The thousand-dollar genome: Genetic brinksman-ship or personalized medicine? EMBO Reports 8, 900–903.

Wu, M., Chiou, H., Ho, I., Chen, C. and Lee, T. (2003). Gene expres-sion of infl ammatory molecules in circulating lymphocytes from arsenic-exposed human subjects. Environ Health Perspect 111, 1429–1438.