Am. J. Hum. Genet. 46:407-414, 1990
IWMAR 1 3 199G
Opinion: The Human Genome Project-Some Implications of Extensive "Reverse Genetic" Medicine Theodore Friedmann Department of Pediatrics, Center for Molecular Genetics, University of California at San Diego School of Medicine, La Jolla, CA
During the past several years, methods have become available that will be used to define completely the organization and even the nucleotide sequence of the entire human genome. The application of these methods to large-scale characterization of the human genome by physical mapping and nucleotide sequencing methods has already been undertaken. Interestingly, the concerted pursuit of an extensive characterization of the human genome has become somewhat contentious, partly because it involves issues of reallocation of scarce funding resources and the relative roles of government agencies and the academic and business communities in the pursuit of science and the delivery of health care. One of the major rationalizations often given for this enormous "human genome project" has been the presumption that a thorough physical characterization of the organization and, ultimately, the nucleotide sequence of the human genome will result in the eventual isolation and understanding of many loci already linked to human disease or sought in disease associations, and that a characterization of their structure and function will lead to improved methods of detecting, preventing, and treating the diseases. Despite the obvious certainty that a great deal of important basic scientific information will flow from undertaking a thorough characterization of the entire human genome and other genomes, the likelihood and timing of the anticipated medical benefits is somewhat less clear. It is the purpose of this report to examine the potential impact of a large-scale physical and genetic characterization of the human genome on the detection, diagnosis, and possibly the treatment of human genetic disease.
Received April 14, 1989; final revision received October 25, 1989. Address for correspondence and reprints: Theodore Friedmann, M.D., Department of Pediatrics, Center for Molecular Genetics, M-034, U.C.S.D. School of Medicine, La Jolla, CA 92093. o 1990 by The American Society of Human Genetics. All rights reserved. 0002-9297/90/4603-0001$02.00
Genetics in Human Disease Since the first description, in 1903, of "inborn errors of metabolism" by Sir Archibald Garrod (Childs 1970), a large number of human genetic diseases have been found to be associated with major genetic components. There are now approximately 4,000 recognized genetic loci or disorders (McKusick 1988), including those responsible for some of mankind's most prevalent and burdensome diseases - cancer, forms of neuropsy-
chiatric, degenerative, and cardiovascular diseases, developmental defects, and many others. Recently, and now with increasing pace, these disorders have become susceptible to efficient diagnosis at the genetic level, either through characterization of mutant genes identified and isolated through aberrant metabolic pathways ("forward genetics") or through linkage with other mapped loci ("reverse genetics"). Traditional Forward Genetics Until recently, most methods for cloning specific
disease-related target genes have required prior characterization ofthe aberrant biochemical expression characteristic of the disease followed by purification of the relevant gene product. With a gene product available in pure form, it becomes possible to devise cloning strategies based on the use of synthetic oligonucleotides or antibodies to screen several different kinds of gene libraries-both genomic and cDNA libraries. Because these cloning strategies are therefore useful only for disorders in which the mutant gene product can be identified, purified, and at least partially characterized, it has proven to be especially useful for studying and developing therapies for the traditional inborn errors of metabolism. Unfortunately, for most human genetic disorders, the responsible gene product or products have not yet been identified, and it is therefore not possible to characterize and understand the genes involved in most human genetic diseases or to devise detection or novel new treatment methods by forward genetics. While technical advances will undoubtedly in407
408
crease the efficiency of detection and characterization of mutant gene products, the identification of the defective gene product or products in many genetic diseases, especially the polygenic ones, will continue to be a major obstacle to the forward genetic approach. The Linkage Map of the Human Genome and Reverse Genetics
There is obviously a great need to be able to identify and characterize genes associated with the vast majority of human genetic disorders with no identified biochemical defect and no known target gene products-that is, diseases that are therefore not amenable to the forward genetic approach. The approach to this problem is conceptually the reverse of the traditional scheme and has come to be called reverse genetics. This approach uses as its starting point nucleotide sequences known to map near the disease locus followed by isolation of the entire relevant nucleotide sequence by gene mapping and chromosomal "walking" or jumping procedures. Until several years ago, the reverse genetic approach might have seemed unlikely and cumbersome, but it has now become not only feasible but in fact one of the most promising keys to an understanding of the bulk of human genetic disease. This scheme depends most crucially on the development of a thorough genetic, or linkage, map of the human genome. Human gene mapping was virtually nonexistent until less than 20 years ago. In 1971, only 15 human markers had been localized to specific chromosomes (McKusick 1971), 12 of them to the X chromosome because of the ease of identifying X-linked disorders by family studies. There were only 11 recognized sets of linked loci, of which only three contained three loci and the remaining sets contained only two linked loci each. By 1979, there were 123 confirmed autosomal and 230 confirmed X-linked markers available for linkage analysis. With the development of the polymorphic RFLP and variable-number-of-tandem-repeat markers, the rate of acquisition of new genetic mapping data has increased markedly, and a recent compilation has identified something like 1,800 expressed genes and more than 2,500 additional RFLP and other markers that have been mapped on the human genome (Human Gene Mapping 9.5 1988; V. A. McKusick, personal communication). Based on these mapped markers, a complete linkage map of the human genome at the 3-5centimorgan level is now very close to reality, in principle permitting linkage analysis of most genetic diseases. The recent striking successes with previously inacces-
Friedmann
sible diseases such as Huntington disease, cystic fibrosis, Duchenne muscular dystrophy, and chronic granulomatous disease (CGD) and the beginnings of similar studies for colonic polyposis, neurofibromatosis, Alzheimer disease, schizophrenia, bipolar disease, and others (Gusella et al. 1983; Knowlton et al. 1985; Tsui et al. 1985; Wainwright et al. 1985; White et al. 1985; RoyerPokora et al. 1986; Barker et al. 1987; Bodmer et al. 1987; Egeland et al. 1987; Koenig et al. 1987; Seizinger et al. 1987; Solomon et al. 1987; Sherrington et al. 1988) indicate the usefulness of a complete human genetic linkage map for the development of diagnostic tools for other human genetic diseases, even the genetically heterogeneous and multigenic ones. Obviously, one of the major promises of the mapping approach to the characterization of human disease is the eventual isolation of a disease locus itself by walking or jumping to it from a nearby linked marker. Until recently, there has been no example of the isolation of any such disease-related gene by this purely reverse genetic approach. The successes with CGD, retinoblastoma, Duchenne muscular dystrophy, and others have come largely through the prior identification of a translocation breakpoint, a deletion, or some other cytological landmark. However, the recent isolation and characterization of the gene responsible for cystic fibrosis (Riordan et al. 1989) has been accomplished entirely by reverse genetics, thereby providing important support for this approach to the identification and understanding of human genetic disease loci. While the methods of reverse genetics are impressive, they remain difficult, laborious and time-consuming. Indicative of the potential problems surely to be encountered frequently in reverse genetic projects have been the difficulties associated with the isolation of the Huntington disease locus. For instance, it has been more than 6 years since the first published report of linkage of the RFLP marker G8 to Huntington disease, and even with the extensive, highly collaborative efforts of a number of large groups consuming probably many hundreds of person-years of work, the HD gene has remained elusive and has not been isolated. Cystic fibrosis has presented similar problems, but seems closer to solution. Of course, the availability of a more thoroughly saturated linkage map will speed progress in the future. Since a complete linkage map of the human genome with a resolution of no more than 3-5 centimorgans is nearing completion, the rate at which genetic diseases will be mapped will now increase more rapidly than ever before. Sequences revealed by linked RFLPs also provide a
Opinion on the Human Genome Project
starting point for the isolation of relevant disease loci themselves by a variety of rapidly developing and improving physical methods, including genomic walking and jumping and physical mapping with pulsed-field mapping techniques to isolate the sequences of the target gene itself. This step in reverse genetics-that is, the isolation of the specific disease-related gene rather than distant but linked markers-would permit the identification of the relevant gene product and, in some cases, an identification of its function through comparative and evolutionary studies (Kunkel et al. 1985; Royer-Pokora et al. 1986). While this approach is conceptually relatively straightforward, it is still filled with technical problems, and it is not clear that one will always recognize the disease locus when one has reached it or even possibly walked right through it. Approaches to a Physical Map of the Human Genome
A detailed genetic linkage map of the human genome will define the relative positions of all known genetic loci, allow the identification of RFLPs linked to most and perhaps to all human genetic diseases, and provide a starting position for the isolation of the relevant disease loci themselves. On the other hand, a linkage map will not rapidly provide nearly as much useful information on intergenic and regulatory regions of the genome that may affect the development of human diseases, and it will not eliminate the difficulties of walking even the several-hundred- or thousand-kilobase distances that will separate most markers on even the relatively densely packed genetic map. It is these problems that require a different solution and for which the physical characterization of the human genome and the development of a fine-structure physical map holds so much promise. A physical map will be colinear with, but not identical to, the genetic linkage map. Genetic distances defined by recombinational events do not always reflect the physical distance between markers because of recombinational "hot spots" and "cold spots" and variations in recombination frequency between males and females and with position of the chromosome (Drayna et al. 1984; Hartley et al. 1984). For some disorders, such as the fragile X syndrome (Nussbaum and Ledbetter 1986), the disease locus lies close to, and may indeed include, a recombinational hot spot, and the task of designing a walk from closely linked markers is made difficult by the problem of not knowing how large a distance the walk would require.
409
The ultimate goal of the development of a detailed physical map of the human genome is to obtain the ultimate map possible -that is, the complete sequence of the human genome. The path toward this goal might be envisioned to take one of three directions, each with different implications for disease detection, characterization and treatment. 1. Total Sequence Determination of the Genome
Many consider that this approach will not be at all feasible unless and until a thorough human genetic map is available and even until ordered sets of overlapping clones, either of cosmids or of yeast artificial chromosomes (YACs) becomes available. The argument is that only when such a resource is available can reasonable decisions be made on the nature and extent of the sequencing effort. Furthermore, many critics of this approach contend that most of the human genome consists of introns, intergenic sequences, or vast and uninteresting noncoding sequences of one sort or another, and that an untargeted sequencing effort would rapidly become bogged down in irrelevant and uninteresting portions of the genome. One potential solution to this problem would be to target a broad sequencing approach entirely at libraries in which coding regions have been selected with cDNA libraries or through the identification of clones containing "gene"-characteristic sequences such as HTF islands (Bird 1986), promoters, or other similar hallmarks of coding sequence. In any event, of all potential approaches to the human genome, immediate sequencing is the least likely to have major, immediate effects on the management of human disease, but it is certain to become increasingly feasible and assume a high priority in the longer term as automated instruments become more readily available and efficient data-handling methods are developed. This approach represents the ultimate in reverse genetics -to define the function of vast stretches of nucleotide sequence from sequence alone and to map and understand all genetic functions. Predicting the exact structure and regulated expression of any gene, the tertiary and quaternary structures of its products, their interactions with other molecules and finally, their exact functions, is difficult and inexact and constitutes a problem of the highest priority in molecular genetics. Extensive interpretations of sequence information will require a great deal of sequence comparison, not only between individual humans to distinguish between polymorphisms and disease-related mutations, but also between species to identify the significance of conserved sequences. A sequencing project per se would - slowly
410
at first and then with increasing efficiency- isolate, map and characterize genes and other sequences useful as direct or linked probes for disease detection and possibly disease treatment. This approach would be aided immensely by an "order of magnitude" improvement in sequencing technology, including both manual and automated methods (Prober et al. 1987). A major aspect of this potential approach would be the need for a judicious choice of the portions of the genome to be targeted for sequencing. 2. The Development of a Fine-Structure Physical Map of Ordered and Overlapping Fragments to Span the Entire Genome, Chromosome by Chromosome
Most versions of this approach have envisioned a combination of restriction mapping together with the construction of a "contig" map, a collection of overlapping and contiguous cosmid and/or YAC clones to span very large portions of the genome. An important scheme designed to augment this combination of approaches, to increase the efficiency of data accumulation from many diverse sources, and to eliminate the need for central clone storage and distribution resources has recently been proposed by Olson et al. (1989). This plan takes advantage of universally available sequence information from the termini of all large cloned human fragments that constitute "sequence-tagged sites" (STSs), which would include primer sequences from which the intervening fragment sequences could be generated by all interested investigators by polymerase chain reaction. If the plan is implemented and is well organized, the goal of producing a complete STS map of the human genome within 5 years seems not to be out of reach. The most important feature of this approach is the fact that it would not only provide map information, but would also make available a set of overlapping, cloned, and ordered human sequences spanning an entire chromosome or portion of a chromosome. Since the relationship of any clone of the set to any and all disease loci on that chromosome would be known, it would, in principle, not be necessary to do difficult and tedious genomic walks or leaps to move from a linked marker to a disease locus. The walk or jump would not be down the genome but rather to the freezer to retrieve a clone. This approach has been used in model studies with the genomes of E. coli and Caenorhabditis elegans. The entire E. coli genome has been ordered and aligned as a contiguous set of clones (Kohara et al. 1987), and a similar study with the genome of C. elegans is well under way (Coulson et al. 1986). The C. elegans genome is approximately 80 x 106 nucleotide base pairs in length, approximately the same size as
Friedmann a small human chromosome such as 21 or a fragment of a larger chromosome, indicating that a similar approach may already be technically feasible for large regions of the human genome. The recent development (Burke et al. 1987) of yeast cloning vectors of higher cloning capacity than the previously used cosmid vectors and the use of pulsed-field electrophoretic mapping methods makes this approach even more attractive. As overlapping contiguous sets of clones become available for large regions of the genome, disease loci will become physically available quickly, since the most difficult and rate-limiting step - genomic walking- can be eliminated. If ordered libraries had been available when the RFLP-linked probes were identified for cystic fibrosis, Duchenne muscular dystrophy, fragile-X syndrome, Huntington disease, Alzheimer disease, bipolar disease, and others, those genes would probably already be isolated and available for studies of pathogenesis and for diagnosis, detection, and effective genetic counseling. The same is obviously true of the many RFLP-linked disease loci that will be identified in the coming years. 3. The Development of a Physical Map, as Above, Combined with Nucleotide Sequence Determination of Selected "Important" or "Interesting" Regions Containing Disease Loci of Particular Interest
Not only will the mapping and ordered-library approach generate specific target genes and other regions requiring sequence determination, but, very soon after it is started, sizable contiguous and overlapping sets of cosmid or other clones (contigs) will be generated. The experience with the mapping projects of both E. coli and C. elegans has indicated that the generation of many contigs of several hundred kilobases in length is rapid, and sequences of this length now represent quite reasonable targets for sequencing, certainly in collaborative settings. The goal of isolating and characterizing sequences that will illuminate the mechanisms of human genetic disease and provide diagnostic reagents is likely to be achieved most efficiently by large-scale genetic mapping accompanied by the preparation of libraries of overlapping, contiguous clones. Through such an approach, disease loci will be identified, isolated, and characterized at the nucleotide-sequence level and used for the preparation of diagnostic and therapeutic materials. Implications for Clinical Medicine For most human diseases, the description of biochemical aberrations and the isolation and characterization
Opinion on the Human Genome Project
of genes has usually had a far greater impact on detec-
tion, diagnosis, screening, prevention, and counseling than on therapy. Despite the temptation of scientists and medical investigators to overemphasize the speed with which detailed structural knowledge, including complete nucleotide sequence information, will become translated into efficient disease treatment, there remains a serious gap between disease characterization and treatment. There are few if any genes whose structure and function are understood as well as those of the globin proteins and, in fact, few diseases whose pathogenesis is as well understood as that of the hemoglobinopathies. And yet, progress toward truly effective rational therapy for these diseases has not moved with nearly the same pace as has biochemical and molecular understanding. Knowledge gained from a detailed understanding of the structure and sequence of the human genome will therefore certainly have its most rapid and immediate effect on detection, diagnosis, screening, and genetic counseling, with only longer-range effects on therapy, whether it be more efficient forms of traditional therapy or conceptually new forms of gene therapy (Friedmann 1989). The anticipated flood of structural and sequence information concerning the human genome that will result from the genome project will be made much more useful and intelligible through the identification and characterization of families in which diseases are expressed. In that regard, the genome project underscores the importance of more effective collection of families with all sorts of genetic disorders -the single gene defects and, of greater importance, the more complex, multifactorial diseases. In the absence of thoroughly studied pedigrees, it will not be possible to make disease associations for the many anonymous pieces of the genome that will become available and whose function and relationship with human disease phenotypes will be sought. The nature and the role of somatic mutations in the development of human cancer, aspects of aging, and many degenerative disorders all seem to be areas in which the application of sequence, mapping, and genomic structural information will very likely be of great usefulness. A particularly important application will be in establishing an effective avenue toward the understanding of multifactorial diseases. Target diseases will not be limited to those determined by simple, singlegene Mendelian loci. The presumption that many human genetic diseases are multigenic and heavily influenced by the interaction of quantitative traits has made genetic characterization seem remote, but recent studies have demonstrated that a sizable number of such
411
traits could be resolved into discrete Mendelian factors by the use of RFLP mapping (Paterson et al. 1988), and it is obvious that a similar approach will be taken to explicate the complicated interactions between the genes in human multigenic disorders as well. For some common, high-burden diseases such as coronary artery disease, hypercholesterolemia, cystic fibrosis, fragile-X syndrome, bipolar disease, Alzheimer disease and others, large-scale population screening would seem to be important, but could probably not be carried out until mutation-specific probes became available to replace RFLP analysis, since very extensive screening based on RFLP analysis is not feasible with present methods. On the other hand, of course, instituting broad genetically-based screening programs is not necessarily to be taken as an obvious good, and one does not even have to point to the most egregious abuses of genetic or "pseudogenetic" information to recognize the potential for misadventure or downright evil. We should remember that it was the "scientists" in prewar Germany- the psychiatrists, geneticists and anthropologists-just as much as the politicians who established the concepts and the principles that came to represent the racial underpinnings of the Nazi extermination programs (Muller-Hill 1988). There have been a number of admittedly less sinister instances in which apparently reliable scientific information has been used to arrive at inappropriate social policy, as in the problems that arose in this country following the institution of screening programs for sickle-cell anemia, in which many sickle-cell carriers experienced problems with employment and insurability. Even more commonly, problems have already arisen and will continue to arise to confound the issue of large-scale screening for disorders for which there are no effective preventive or therapeuric measures. Disorders such as Huntington disease, :chizophrenia, and bipolar disease illustrate the point that genetic screening programs aimed at some diseases, especially untreatable ones, might not be universally desired or accepted. However, it does seem very likely that populationwide programs will eventually come for a variety of disorders. There will be an opportunity to add many newly diagnosable diseases to those for which prenatal diagnosis is feasible. While the identification of genetic components of human diseases has principally and naturally been the province of genetic epidemiology, there is little question that the availability of some molecular genetic tools can facilitate the unambiguous identification of such genetic components, as has already occurred in some forms of bipolar disease, schizophrenia, and other disorders.
412
Even then, the costs and logistical problems associated with such widespread programs will be enormous. As with most other medical and scientific advances, the medical application of new genetic knowledge, techniques, and reagents derived from an extensive characterization of the human genome will carry costs - financial, scientific, and societal. Most of the problems raised by the advent of so much new knowledge of genetic diseases and of so many new screening and detection programs are the same problems now faced by modern medicine in this era of apparently limited resources, will, imagination, and commitment, but they are magnified manyfold. Choices will have to be made on the diseases for which large-scale detection and screening programs should be established. No doubt, if and when effective detection methods become available for major, common, high-burden diseases such as most forms of cancer, neuropsychiatric disorders, coronary artery disease, diabetes mellitus and others, they will and should represent some of the suitable target diseases. However, many severe but rare diseases will become susceptible to detection and control but will become "orphaned" or ignored because of their low societal impact. LeschNyhan disease and ADA-deficiency immune defects would be just two examples of such disorders. This too is not a new phenomenon in medicine. It is far from certain that our medical, public-policy, and insurance institutions, both public and private, nonprofit, and for-profit, are adequately trained and structured to handle the anticipated heavy load of new screening and diagnostic procedures that may eventually involve a very large proportion, even the majority, of the population. There are uncertainties whether our societies can predict, afford, or contain the financial costs of screening programs. With the likelihood that genetic diagnosis will be feasible for the most common and severe diseases of our society, the overall costs of large-scale screening programs can become very great very quickly. For instance, in many cases, insurance companies and other third-party carriers do not uniformly reimburse patients fully for heterozygote detection procedures or for genetic counseling, activities that will certainly increase markedly with the development of many new genetic diagnostic programs. On the other hand, extensive screening programs might be justified by the knowledge that disease prevention can and will save immense amounts of money that otherwise would be spent on treatment. Large-scale development of new genetic information will pose few if any new ethical dilemmas for medicine but will certainly greatly exacerbate those with which
Friedmann we already struggle. There will be an increase in the number of diseases, such as Huntington disease, susceptible to early detection but for which there is no effective therapy, greatly complicating the personal burdens of threatening medical information. Since treatment methods will continue to lag behind detection efficiency, the obvious application to prenatal diagnosis of genetic disease predicts increasing use of medically indicated abortion. The already abysmal quality of our national discourse on this issue is not likely to be improved by a markedly enhanced capability for genetic characterization before birth. With the need to study entire families to make RFLP-based diagnoses, it will become increasingly difficult to ensure the confidentiality of vastly expanded genetic diagnostic information, to identify exactly who the patient is and to determine who should have access to genetic information. With long-range health prognostication possible through genetic analysis, tensions between the interests of individual interests and those of employers and insurers will become increasingly severe. The role of routine and possibly even mandatory screening and monitoring programs, undertaken in the interests of ensuring safe workplaces and productive workers or in the pursuit of other public health or public policy considerations, will come into conflict with the desires and the autonomy of individuals. One potential approach to solving such problems and to ease the resulting ethical tensions is to restrict the development of the genetic knowledge itself- a step in fact proposed by a number of political and social bodies who primarily see, not completely without cause, overwhelming social danger in the acquisition and use of genetic information. However, a more appropriate activity would seem to be to examine and restructure our sociopolitical institutions to ensure that they do not have the opportunity to use genetic information as a tool for injustice and inequity. In such a rapidly developing and changing field, services will not be distributed uniformly, and inadequately trained or prepared health-care providers will be faced with a flood of new techniques and procedures. It is inevitable that questions of breached confidentiality, genetic damage, wrongful life, and medical negligence will be raised. Again, there is little that is conceptually new in any of these issues except for their magnitude. Summary and Conclusions Impressive progress has been made during the past several decades in understanding the pathogenesis of
Opinion on the Human Genome Project
human genetic disease. The tools of molecular biology have allowed the isolation of many disease-related genes by forward and a few by reverse genetics, and the imminent completion of a complete human genetic linkage map will accelerate the genetic characterization of many more genetic diseases. The major impacts of the molecular characterization of human genetic diseases will be 1. To increase markedly the number of human diseases that we recognize to have major genetic components. We already understand that genetic diseases are not rare medical curiosities with negligible societal impact, but rather constitute a wide spectrum of both rare and extremely common diseases responsible for an immense amount of suffering in all human societies. The characterization of the human genome will lead to the identification of genetic factors in many more human diseases, even those that now seem too multifactorial or polygenic for ready understanding. 2. To allow the development of powerful new approaches to diagnosis, detection, screening and even therapy of these disorders aimed directly at the mutant genes rather than at the gene products. This should eventually allow much more accurate and specific management of human genetic disease and the genetic factors in many human maladies. The preparation of a fine-structure physical map of the entire human genome together with an overlapping contiguous set of clones spanning entire chromosomes or large portions of chromosomes is rapidly becoming feasible, and the information that will flow from this effort promises eventually to affect the management of many important genetic diseases. This notion is an extension of the concept of reverse genetics to what might even be called reverse genetic medicine, in which the isolation and characterization of sequences from the genome allow not only improved understanding of pathogenesis but also improved treatment of human genetic diseases. Large-scale, non-locus-specific nucleotide sequence determination requires still further improvements in methodology before it can be expected significantly to affect the diagnosis, detection, or understanding of human disease or to illuminate mechanisms of human genome expression. However, within the coming decade, extensive nucleotide sequence analysis of the human genome will also begin to provide clinically useful reagents. The wide-scale availability and use of these genetic techniques for a large number of human disorders may not raise any truly novel problems for the delivery of clinical care to patients, but
413
will certainly exacerbate current problems in the present health-care system, including issues of just and equitable identification of target populations, costresource allocation, the choice of target diseases for major emphasis, the establishment of delivery systems for massive new screening and detection programs, the training and the regulation of health-care providers knowledgeable in genetics, and the preparation and education of the system of jurisprudence for the inevitable challenges to it.
References Human Gene Mapping 9.5 (1988) Update to the Ninth International Workshop on Human Gene Mapping. Cytogenet Cell Genet 46:1-25 (1988). Cytogenet Cell Genet 49: 1-260 Barker D, Wright E, Nguyen K, Cannon L, Fain P, Goldgar D, Bishop DT, et al. (1987) Gene for von Recklinghausen neurofibromatosis is in the pericentric region of chromosome 17. Science 236:1100-1102 Bird AP (1986) CpG-rich islands and the function of DNA methylation. Nature 321:209-213 Bodmer WF, Bailey CJ, Bodmer J, Bussey HJR, Ellis A, Gorman P, Lucibello FC, et al (1987) Localization of the gene for familial adenomatous polyposis on chromosome 5. Nature 328:614-616 Burke DT, Carle G, Olson MV (1987) Cloning of large segments of exogenous DNA into yeast by means of artificial chromosome vectors. Science 236:806-812 Childs B (1970) Sir Archibald Garrod's conception of chemical individuality: a modern appreciation. N Engl J Med 282:71-77 Coulson A, SulstonJ, Brenner S, KarnJ (1986) Toward a physical map of the genome of the nematode Caenorhabditis elegans. Proc Natl Acad Sci USA 83:7821-7825 Drayna D, Davies KE, Hartley DA, Mandel J-L, Camerino G, Williamson R, White RL (1984) Genetic mapping of the human X chromosome by using restriction polymorphisms. Proc Natl Acad Sci USA 82:2836-2840 Egeland JA, Gerhard DS, Pauls DL, Sussex JN, Kidd KK, Allen CR, Hostetter AM, et al (1987) Bipolar affective disorders linked to DNA markers on chromosome 11. Nature 325:783-787 Friedmann T (1989) Progress toward human gene therapy. Science 244:1275-1281 Gusella JF, Wexler NS, Conneally PM, Naylor SL, Anderson MA, Tanzi RE, Watkins PC, et al (1983) A polymorphic DNA marker genetically linked to Huntington's disease. Nature 306:234-238 Hartley DA, Davies KE, Drayna D, White RL, Williamson R (1984) A cytological map of the human X chromosome: evidence for non-random recombination. Nucleic Acids Res 12:5277-5285
414 Knowlton RG, Cohen-Haguenauer 0, Cong NV, Frezal J, Brown VA, Barker D, Braman JC, et al (1985) A polymorphic DNA marker linked to cystic fibrosis is located on chromosome 7. Nature 318:380-382 Koenig M, Hoffman EP, Bertelson CJ, Monaco AP, Feener C, Kunkel LM (1987) Complete cloning of the Duchenne muscular dystrophy (DMD) cDNA and preliminary genomic organization of the DMD gene in normal and affected individuals. Cell 50:509-517 Kohara Y, Akiyama K, Osono K (1987) The physical map of the whole E. coli chromosome: application of a new strategy for rapid analysis and sorting of a large genomic library. Cell 50:495-508 Kunkel LM, Monaco AP, Middlesworth W, Ochs HD, Latt SA (1985) Specific cloning of DNA fragments absent from the DNA of a male patient with an X chromosome deletion. Proc Natl Acad Sci USA 82:4778-4782 Levinson B, Janco R, Phillips J, Gitschier J (1987) A novel missense mutation in the factor VIII gene identified by analysis of amplified hemophilia DNA sequences. Nucleic Acids Res 15:9797-9805 McKusick VA (1971) Mendelian inheritance in man, 3d ed. Johns Hopkins University Press, Baltimore . (1988) Mendelian inheritance in man, 8th ed. Johns Hopkins University Press, Baltimore Muller-Hill B (1988) Murderous science. Oxford University Press, Oxford Nussbaum R, Ledbetter D (1986) Fragile X syndrome: a unique mutation in man. Annu Rev Genet 20:109-145 Olson M, Hood L, Cantor C, Botstein D (1989) A common language for physical mapping of the human genome. Science 245:1434-1435 Paterson AH, Lander ES, Hewitt JD, Peterson S, Lincoln SE, Tanksley SD (1988) Resolution of quantitative traits into Mendelian factors by using a complete linkage map of restriction fragment length polymorphisms. Nature 335: 721-726
Friedmann ProberJM, Trainor GL, Dam RJ, Hobbs FW, Robertson CW, Zagursky RJ, Cocuzza AJ, et al (1987) A system for rapid DNA sequencing with fluorescent chain-terminating dideoxynucleotides. Science 238:336-341 Riordan JR, Rommens JM, Kerem B-S, Alon N, Rozmahel R, Grzelczak Z, Zielenski J, et al (1989) Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science 245:1066-1073 Royer-Pokora B, Kunkel LM, Monaco AP, Goff SC, Newburger PE, Baehner RL, Cole FS, et al (1986) Cloning the gene for an inherited human disorder-chronic granulomatous disease-on the basis of its chromosomal location. Nature 322:32-38 Seizinger BR, Rouleau GA, Ozelius LJ, Lane AH, Farniarz AG, Chao MV, Huson S, et al (1987) Genetic linkage of von Recklinghausen neurofibromatosis to the nerve growth factor receptor gene. Cell 49:589-594 Sherrington R, Brynjolfsson J, Petursson H, Potter M, Dudleston K, Barraclough B, WasmuthJ, et al (1988) Localization of a susceptibility locus for schizophrenia on chromosome 5. Nature 336:164-167 Solomon E, Voss R, Hall V, Bodmer WF, Jass JR, Jeffreys AJ, Lucibello FC, et al (1987) Chromosome 5 allele loss in human colorectal cancer. Nature 328:616-619 Tsui L-C, Buchwald M, Barker D, Braman JC, Knowlton R, Schumm JW, Eiberg H, et al (1985) Cystic fibrosis locus defined by a genetically linked polymorphic DNA marker. Science 230:1054-1057 Wainwright BJ, Scambler PJ, Schmidtke J, Watson EA, Law H-Y, Farrall M, Cooke HJ, et al (1985) Localization of cystic fibrosis locus to human chromosome 7cen-q22. Nature 318:384-385 White R, Woodward S, Leppert M, O'Connell P, Hoff M, HerbstJ, Lalouel J-M, et al (1985) A closely linked genetic marker for cystic fibrosis. Nature 318:382-384