 Thanks, Kathy. Like prior speakers, I would begin by pointing out the amount of hubris taken to really execute a talk like this as someone who is much more accustomed to worrying about what I'm going to be doing this afternoon. Trying to project myself 20 years into the future is particularly difficult. In the setting, I thought at the outset that I might go back to prior sages and see what they had thought about these kinds of issues. And one that came to mind was this, making predictions is risky, particularly about the future, which Dan Quayle said a number of years ago. And I think this really is particularly apt, but nonetheless trying to move forward. If you think about the future of genetics in medicine, I think there will be a number of unifying principles over the coming years, and these are fairly easy to state. The identification of genes and pathways that predispose to virtually every human disease, the ability to rapidly identify inherited predisposition prior to the development of disease, the production of medicines targeting specific causes of disease, and finally the ability to tailor treatment to the specific abnormality in individual patients. The notion that genes do indeed contribute to all of these traits is nicely illustrated by this apocryphal nuclear family. I think in the absence of genetic variation, we can see that there would be astonishingly little variation in gross phenotypic features such as height and weight, body habitus, down to very fine detailed features of facial structure, and behavioral traits as well such as taste and clothing. So clearly genes are contributing to just about everything we might be interested in, and we now suspect that these are all going to ultimately be attributable to variations in DNA sequence, and the task before us is to how to relate these variations that we're recognizing to phenotypes that are relevant to health and disease. And this really is where we've been in this enterprise in medicine up to the last several years. These lonely riders on a dry, dusty trail hoping to find something interesting but not quite sure how we're going to get there, and as many have pointed out, what we're headed to in medicine is something that looks much more like this, a much more crystalline form in which we know all of the genes, all of the common variants, and it really becomes a matter of putting the puzzle together of which variants contribute to particular diseases and how does this impact our diagnosis and therapeutic approaches. So the key for the next 20 years is, I think, going to be the roots to target identification. At the end of the day, the public expects us to be coming up with things that are going to improve human health, and these are going to be diagnostic and new treatment approaches. The roots to target identification, looking back now from the perspective of 20 years, from the year 2020 back to today, in retrospect it was apparent that Mendelian models of both human and model organisms, multifactorial trait genetics in human and animal models proved to be enormously important in identifying targets for new therapeutic intervention. Equally important in retrospect proved to be altered gene and protein expression in disease states. Also proved critical to be able to bootstrap along pathways to reach drug-able targets. That is to say that many of the genes that were turned out to be causative for disease were not themselves amenable to therapeutic intervention, so it proved important to be able to work our way along pathways until we could reach a target that we could intervene with. And finally, the development of antagonists and agonists against molecules and classes of fertile drug targets that already exist. This is something that the pharmaceutical industry, again in retrospect, recognized that simply they could use genomic mining to identify classes of drug-able target, and without much knowledge about what their actual functions were, were able to develop small molecule antagonists and agonists, and then assess the phenotypic consequences of these molecules, and having these genes in hand made an enormous difference. One illustration of, in retrospect, how genetics and genomics made a big impact on human disease is exemplified by this slide, which is the brain of a patient with Alzheimer's disease. Alzheimer's disease in the year 2001 affected about 4 million individuals in the United States alone, and in the year 2020 this disease is largely controlled, although not eradicated. So this shows the plaques of a patient with Alzheimer's disease from the brain. The biochemical nature of the molecules in this plaque was identified by biochemical studies. Molecular biology demonstrated that one of the proteins, the beta amyloid, was derived from the amyloid precursor protein, a classic, straightforward molecular biology project. Genetic studies had a major impact in identifying that mutations in the amyloid precursor protein, indeed, caused Alzheimer's disease as a rare Mendelian trait. Other rare forms of Alzheimer's disease were shown to be due to mutations in the presynolins, which turned out to be the gamma secretases that were responsible for one of the cleavages that was on the route to production of beta amyloid. And this led to the recognition as well that the beta secretase was involved in this pathway, and although no mutations were identified in the beta secretase, this also was cloned by expression cloning. And this led to recognition of these as amenable targets for drug discovery, and this proved to be very important for advances in prevention of Alzheimer's disease. Multifactorial genetics also demonstrated that there was a common allele that was a major cause for Alzheimer's disease. ApoE4 is a single allele in apolipoprotein E. There was a major risk factor for development of Alzheimer's disease, and was the first example of a common gene allele contributing to a common disease in the new era. Several other diseases have turned out to be, had marked improvement in their treatment as a consequence of genetic studies. Many of these places have been diseases in which we've been able to slow down disease progression, and these have had a substantial impact. One of the major areas has been kidney failure. They're currently in the year 2001. There are 300,000 Americans on dialysis for renal failure. Mortality of these patients was comparable to most forms of cancer. One of the first advances in this area was understanding the molecular basis of polycystic kidney disease, which is a Mendelian form of end-stage renal failure, and just simply slowing, agents that simply slowed down the rate of cyst progression had a major impact on the development and progression of this disease. Aging is another area that has had the substantial impact. Understanding rare Mendelian forms of aging has led to new approaches to treatment. And in the year 2020, I'm pleased to say that Sherm Weissman, one of the fathers of the entire field of genomics, now 95 years old, wrote me recently to tell me about the birth of his 21st child. So, looking back at the genetics of Mendelian traits, in the year 2020, it's evident that expectations of its rapid decline were greatly exaggerated because in the year 2001, only a quarter of the known Mendelian traits had been solved, and only the most obvious half of the viable Mendelian traits had in fact been recognized. So, simply in the period from September to November of 2001, three important new pathways for human physiology have been recognized. A new kinase pathway involved in blood pressure homeostasis, protease involved in maintenance of vascular homeostasis, and mutation in a new gene that has identified an entirely unrecognized pathway in the maintenance of normal bone density, whose loss of function leads to osteoporosis. Over the ensuing 20 years, there were many more of these examples. These contribute to fast and convincing validation of targets for pharmaceutical development that really had a major impact on pharmaceutical development. Just in the year 2001, the development of Gleevec had a big impact on the pharmaceutical industry, who finally realized the opportunities to use targets identified from genetic studies to really direct new pharmaceutical development. Sources of new Mendelian traits became important. As I indicated, we only knew about half of them in the year 2001. Large-scale resequencing of genes and pathways of interest, followed by clinical and biochemical and cell biology, identified a host of new Mendelian traits that we previously had not recognized. In addition, a number of new recessive traits were identified by transcultural studies in the Middle East. What I mean by that is physicians going to the Middle East looking for diseases that had much higher prevalence than found in the West identified a host of diseases that trend out to be recessive traits in these populations with high levels of consanguinity. These had dramatic impact in a number of fields, particularly the neuropsychiatric disorders, which have been desperate for identification of some simple Mendelian paradigms to get a foot in the door for pathophysiology. Another area that had been under explored but proved to be very fruitful was looking for genes with large effect but very low penetrance. Studying very extended kindreds with a small number of individuals who were affected by a disease who were quite distantly related to one another, turned out to be productive for identifying single genes with large effect. And finally, the use of animal models, particularly forward genetics in models such as the mouse and zebrafish, continued to provide a lot of insight into pathways underlying basic human physiology that proved relevant to disease biology. There are a number of obstacles that needed to be overcome in the genetics of human studies. The rate limiting step for identification of Mendelian traits became appropriate patient collections. And from the year 2000 to the year 2008, we had a real problem in clinical medicine. Regulatory agencies became the rate limiting step to genetic studies. There were overstated privacy concerns that made recruitment of families almost impossible. Paternalistic regulatory attitudes to the study of participants denied their full participation. It was impossible to go back to patients after we had identified mutations to do meaningful physiology on them. There was a regulatory focus on paperwork rather than on protection of patients and human studies. And there was insistence of identical IRB procedures in groups with drastically different cultural norms that inhibited the international studies that would ultimately prove so fruitful to understanding human disease. There are a few improvements that made a big difference. There are major advances made when legislation that prohibited genetic discrimination was passed and real world sensibilities gave way to tangible benefits of research in addition to theoretical harms. There was another important issue that has been alluded to earlier, and that is the finding and training of physician scientists. I would simply begin by addressing the question of, is it important to have physician scientists in this mix? And I would answer it with two words, Barbara McClintock. So if you think back to Barbara McClintock's remarkable career, she had insights that were 30, 40, 50 years ahead of her time, primarily I think because she had such profound understanding of the biology of corn, the organism that she was so interested in. Well, the simple answer to who has that same understanding of human biology, the simple answer is well-trained physicians. And these will be without substitute if we're going to take the full advantage of the opportunities in biomedical science. And things that I think have made a difference are investing in what works. It's turned out to be very important to identify and engage people early in their training, trying to bring somebody after four years of medical school, one or two years of clinical fellowship, and then say, okay, you're going to turn your life around and start to become a scientist. It was a difficult road for many to go. But we know one thing that works is the MD-PhD program. 75% of MD-PhDs funded by the medical scientist training program at NIH go on to academic faculties and NIH funding. And we ought to invest in this until these returns start declining. We need to develop ways for career development to evolve that allows for larger academic teams. This is something that industry long ago figured out, physicists long ago figured out, and we need to do the same thing in biology to allow individuals to make different types of contributions that don't all require that their PI on every grant and our first or last author on every manuscript in order to be advanced. And finally, one of the great opportunities that we have is to take advantage of this knowledge for renaissance in human physiology. GCRC's clinical research centers are lying relatively dormant, but these genes are going to provide enormous opportunity for understanding human physiology, and we have to make sure that we still have physicians around who know how to take advantage of these opportunities. So Mendelian diseases have identified a host of pathways that have proved important in a number of disorders, this one example in blood pressure. Being able to go back to patients and do detailed clinical physiology has proved key. Once you identify these genes, this shows one example where a gene that we had identified causing that changes blood pressure. Once we knew what the gene was and we're able to tease out the patients with these specific mutations and do physiology, we found that they had an unexpected effect on calcium homeostasis and found that the mutation in this gene actually had a large effect on bone density, a classic example of genetic pleotropy, and we never would have been able to do this if we had not had the molecular genetic background and the opportunity to go forward with clinical studies once genes have been identified. So the outcomes of Mendelian genetics over the last 20 years have been identification of new pathways underlying disease, identification of drug-able targets and new small molecule antagonists that were identified, that turned out to be applicable to millions of individuals who did not have mutations in the pathway. And as Jeff Dewick alluded to, this paradigm was well demonstrated by the effect of statins on cholesterol homeostasis. We now are treating tens of millions of individuals with drugs that were developed largely on the basis of the understanding of a disease that affects one in a million in the population. Another area that turned out to be surprisingly productive is pharmacogenomics, the ability to predict adverse drug response, the ability to predict beneficial drug response, and the big surprise was that this actually turned out to be much more productive than I think I would have anticipated, and it turned out in retrospect that drugs are brand new environmental factors and selection had not had a chance to select against undesirable genotypes, so over the last 20 years we found surprisingly many genes with large effect on drug response and risks of complication were rapidly identified. One of the daunting challenges in addressing this problem was routine collection of samples in clinical trials was very slow to develop, and promoting this turned out to be of substantial importance in unraveling this area. Genetics of multi-effectorial traits turned out to be one of the granddaddies over the last 20 years, and it nearly lost all credibility before the year 2003. The expectations were high, and the false positive rates were much higher. So, among the outcomes, there were some standout successes like Alzheimer's disease, where there turned out to be common alleles with large effect that were previously unrecognized pathways that were discovered, and these had an enormous impact on public health. There also were a number of other modest successes where there turned out to be common alleles with very small additive effects to disease risk, and those that defined new pathways turned out to be a very big help, and in the neuropsychiatric field this proved to be particularly important. However, there were also variants that were identified in previously known pathways that had very small incremental effect, and these turned out to contribute very little to either diagnosis or treatment. There were also diseases that turned out not to have common alleles underlying common diseases, and there were many independent disease alleles, and obviously one of the major problems looking back was that in the year 2001, we had almost no ability to discriminate a priori which of these was going to hold. All we were able to do was recognize which genes, diseases were Mendelian, which were not, and trying to distinguish whether there were two genes, three genes or 50 genes contributing to a trait was something that was well beyond our grasp, and we spent a long time trying to take vets and take risks with these approaches. Some worked, some didn't. There were two surprises in multifactorial trait genetics. One was the transcriptional variation turned out to be central to many multifactorial traits, and genomic identification of transcriptional regulatory sites and circuits turned out to be fundamentally important in understanding these traits. And the second part that surprised me in particular is that understanding genetic contributions to disease turned out to be key to identifying environmental contributions to disease as well. Once we knew pathways from genetic factors, identifying how environmental factors played into these turned out to be fundamentally important, and much to my surprise, the epidemiology once again became an important field. Things that made a difference in the genetics of multifactorial traits, basic blocking and tackling. NIH made an investment in large collections of well-phenotyped subjects. There was better consideration of power to find genes with small effect by emphasizing disease severity and emphasizing sample size. We gave up on the notion of collecting 100 sub-pairs and trying to find a gene, or getting 100 cases and 100 controls and trying to find genes that way. This turned out to be the major factor. Systematic testing for population stratification also became routine very rapidly, and VANA I'd like to buy a decent control sample version. Other things that made a difference, wide access to collected populations was made possible once we knew how to amplify genomes so that we could no longer make the argument that I can't afford to send you a sample because I don't have enough DNA to send it to everybody. Once we were able to do faithful genome-wide amplification, that argument went away, and this permitted us to distribute all of these collected samples to anyone who wanted them. On an open access basic, a genome project model that NHGRI deserves enormous credit for, and this permitted much better exploration of data space than we ever would have achieved if we had samples with confined access to a small number of individuals. We also finally solved the question of how to provide wide access to genotype and phenotype data for the same reason. Having a small number of individuals sitting on all of the data and analyzing traits whenever they had the time and opportunity was insufficient, and eventually we were able to give this wide access and allow everybody who was interested to have a go at it. Dealing with the ethical issues here turned out to be particularly thorny, but ultimately was doable. Tool and technology development really drove a lot of this. The decision to do the linkage disequilibrium map of the genome proved to be what permitted the genome-wide assessment of association studies in linkage disequilibrium. Cheap high throughput SNP genotyping turned out to be critical. We look back now at sequinom and just are astonished that we spent $450,000 on every one of those machines. Cheap high throughput resequencing of large chunks of genomic DNA also turned out to be something that really made an enormous difference. We had many small LOD score peaks that we were able to take five, 10-centimorgan intervals and simply resequence 100 cases and 100 controls, and then obviously had enormous computational issues to address in handling the data. And finally, the ability to recognize marks of selection on genomic segments in humans turned out to be surprisingly useful in recognizing when we were actually in a chromosome segment that impacted a trait. Finally, giving up the naive expectation that case control studies would commonly solve disease pathogenesis and instead generate starting hypothesis. This turned out to be important. The outcome of many of these multifactorial trait studies gave us interesting hypotheses but unconvincing final data to go forward, and we realized that geneticists have to be biologists too and use all of the tools available. Human and animal genetics, cell biology and biochemistry, obviously now in retrospect were essential to wrestle these problems to the ground. The ability to monitor gene expression and protein expression obviously has had an enormous impact. Early examples of how we can stratify patients on the basis of altered gene expression, shown in this slide of large cell lymphoma, was just the starting point for being able to reclassify diseases in useful ways based on altered patterns of gene expression. Critically, the systematic characterization of pathways from genetic and genomic studies of model systems let us define biochemical pathways, and then the ability to analyze human disease tissues in both humans and animals. This gave us a major advance in understanding disease biology. The generation of these large data sets under different experimental conditions permitted the definition of functional biochemical pathways. Expression studies identified pathways underlying disease process, and understanding these pathways was fundamental in letting us make sense of microarray data. Finding 10 genes in the same pathway, each with two-fold altered expression, was a lot better than going point by point. So identifying drugable targets from non-drugable leads. Most genes identified are not tractable targets for drug development in the near term. The use of model organism genetics to flesh out pathways from the initial gene turned out to be very important, and as suggested earlier today, genome-wide interaction maps also proved fundamental for bootstrapping our way to tractable targets for drug development. This simply shows one example of this from Tian Xu's work demonstrating that mutations in the tuberous sclerosis complex genes or tyrosine phosphatases had surprising impact when put into flies on cell and organ size, and the ability to go through these kinds of pathways in model organisms was critically important. Well, the impact of genetics and genomics on drug development has been quite profound. The drugs directed to validated targets identified by genetic and genomic studies have become routine. Drugs directed to known drugable classes identified from genomic sequence, this turned out to be very fruitful for the pharmaceutical industry as well. Advances in structure determination and modeling greatly aided drug design and discovery, and greater diversity chemical libraries played an important role. I would also point out the important role that genomic and proteomic markers of toxicity in vitro played for the pharmaceutical industry. Pharmaceutical industry, one of the major barriers was having phase one, phase two, or even phase three failures due to unanticipated toxicity, and the ability to use genomic and proteomic approaches to predict this toxicity and basically kill projects before they got into the clinic had a very important impact on the pharmaceutical industry and has greatly speeded the number of compounds that the pharmaceutical industry has been able to triage and get into the pipeline. Finally, by 2020, new classes of transcriptional modifiers have begun to have a major impact. Minor group binding transcriptional activators, SIRNA or RNAI-mediated suppression of gene expression has started to find its way into the clinic, and gene therapy, even in 2020, I'm still agnostic about its prospects. Implementation in medical practice. So by 2020, the precise molecular basis of most human diseases has been understood. Remaining questions include what are the boundaries of medicine, behavior, and cosmetic traits? Is this part of the medical armamentarium or not? And does it make a difference? Well, there's a big surprise. Physicians and patients are both surprisingly pragmatic groups. Given tools that have meaningful impact on diagnosis or treatment, they are rapidly assimilated into standard practice. Patients demand these. Physicians want to apply them, and these rapidly overran many of the other barriers that we imagined. Obviously, we require much greater reliance on expert systems for ordering and interpreting tests and selecting therapies, but the results of these still feed into comprehensible disease categories that physicians have been dealing with for the last 100 years, and this provided much less of a barrier than was imagined 20 years previously. We have many more preclinical markers of disease risk and provide opportunities for therapeutic, for disease prevention. And ultimately, one of the major questions that we're still dealing with in the year 2020 is the question of whether all of this knowledge has been much to the good of our development of new health care, but on individual patients has it created more neurosis, rather than having imagined targets that you can worry about on your health. You now have defined targets that you can really focus on. And that has become one of the issues going forward. So I think I'll stop there. Thank you very much for your attention. Stanford. That was nice, Rick. I enjoyed that. Do you really think that the reason that we have the physicians, that it was really whole genome amplification that allowed us to be able to do this, or that it was the fact that the laws finally changed and made people actually put their samples in so that other people could get access to them? It was a sociological problem. So I think it's a bit of both. I think you can help just by making it a non-issue to be able to say, we have infinite supplies of DNA, so we can't use as a reason the fact that it's a finite supply. And I think that will end up having a significant effect. Obviously, going forward, we can require that these samples be made widely available. Well, I just remember in 2005 I was making cell lines, and I had infinite supplies of DNA, and people were doing that, and they still weren't available because of restrictions. I think that's really probably as much a problem or maybe more of a problem than that. I know it's an excuse, but it's the culture of wanting to hold on to the samples. It's hard in almost every study that I've ever seen or been involved in. I absolutely agree with that, and the larger the studies become as we recognize that we need really big studies to do these, I think we will have to insist that samples be generally available or else we will clearly not get the most out of them. I think as taxpayers we ought to demand. Robert Klitzman from Columbia University. Concerning your comment about the overstatement of privacy concerns that impeded things way back in the early years of the century. If you remember also in the 2005 or so, the well-publicized cases of breaches of confidentiality that occurred where people lost jobs and some people weren't able to get married because of course everyone was able to look up everyone else's genetic information on the net. What do you think should have been the privacy regulations that were in place then? Yeah, I think this is clearly a very important issue. So obviously it is very important to maintain privacy and to take every step necessary to ensure that patient confidentiality within the research setting is maintained. There have been I think precious few examples of where breaches of that has actually caused harm in practice. I also think that some of the things that are actually practically a foot now to basically make it impossible to inquire about family history in the setting of research to contact patients as part of family studies without going having your pro-band contact them and have them write you. These are things that make doing clinical research of this sort impossible. And having collected thousands of families and patients by this route, I'm very concerned that we put too much emphasis on theoretical risk to patients over I think quite tangible benefits from the research that we're doing. Obviously trying to, I think in 2005 of course the legislation to protect against genetic discrimination has already been passed and put into law. So some of those issues are less pressing now. But clearly we're always going to have to respect patient protection, patient privacy, but we do need to have a balance between tangible benefits from research and theoretical harms. And obviously Dr. Young's just going to address this. Art. Art Podette of Baylor in Houston. I just wanted to second Rick's comment about there's already infinite supplies of DNA available. It's perfectly normalized in lymphoblasts and I think that that's something that, the technology's there and there's something about the culture that's not ready. I wanted to make another comment though. I really enjoyed your presentation and I thought it was very stimulating. For another perspective too though, that somewhere between now and 2020 it was realized that epigenetic variation had a huge impact on gene expression and disease. That epigenetics was very malleable. You couldn't study the changing DNA methylation that was going on in individual cells in the brain by studying lymphoblasts. And we found out that this malleability of the chromatin and epigenetic factors opened great avenues for therapy as well. I apologize for forgetting about that part. David. Rick, I couldn't agree more that identification of genetic risk factors may stimulate and guide epidemiologic studies to identify the environmental variables. But I wonder if we shouldn't be thinking about things that we could do rather than waiting for the epidemiologists to sort of come along with this if we shouldn't be thinking about proactive strategies to make the epidemiologic community or to help make it come to the realization that these studies really need to get going to have sort of the equivalent of an environmental genome project. Yeah, I think that's a very interesting idea. I think the studies that combine environmental factors and genetic factors obviously are going to have a lot of opportunity for advancing understanding of both areas. And my impression, and you can chime in as well, is that the epidemiology world is very enthusiastic about these issues. And the major issue that we still, I think, deal with periodically is the epidemiologists' desire to both find the genes and simultaneously determine their population impact. And I think this is a recurrent issue of should we be doing population-based studies or should we be doing disease-based ascertainment and investigation. For my money, I think the biggest bang for the buck is going to continue to be doing population-based studies after you have genes in hand and have some idea about what you're dealing with rather than trying to solve both equations simultaneously. But I think there will continue to be debate on that point. I'll just share from the University of Wisconsin. Actually, you have a question, but the previous comment does make me want to just add the following first. And that's the first thank you for the 2005 anti-discrimination legislation because that was the biggest ticket to women's equality that we've ever had since it's the biggest form of genetic discrimination that's prevalent in the United States. But even more, I'd like to thank the scientists here because apparently it was in 2008, according to your estimates, scientists finally looked up from their benches and their state of victimhood and frustration and organized and coordinated to endorse any one of the several perfectly rational efforts to standardize and streamline and make uniform the system for working with large-scale tissue banks, medical records, and patient contacts so that there'd simply be one system that everybody could agree upon. And it was that organizational effort by the scientists interacting with the political and regulatory community that finally achieved that success of 2008 that you mentioned. But the question... I couldn't agree with you more. The question that I had for you seriously was this. Most of the data that you presented was in the context of what seems to me to be individual doctor-patient interactions and the use in that kind of setting. And I found myself wondering how much of an event would be useful for understanding on a population level the patterns of disease and resistance to existing therapies, the demography of it in a sense, age of onset, geographic distribution, and the degree to which you think this might actually get used not so much in the clinical setting, but in the public health setting for infrastructure planning and for public health intervention planning. Yeah, I think that's a very apt point that I did not go into. It certainly will give us a much better definition of what we're up against from a public health perspective in both the near term and the long term. And I think that is very important. Understanding the prevalence of mutant alleles in particular populations has the potential to really help shape public policy and I think there will clearly be a lot of work necessary to define that. I think we'll still be starting to sort those out in the year 2020, I suspect. And I think there will continue to be practical barriers to doing that because of continuing concerns about stigmatization of particular populations. And I think those will be important issues to wrestle with. Jennifer Puck from NHGRI and I may be the only person here who's actually the head of an IRB and I just wanted to point out that many of the frustrations we all experience in wishing we could go back and revisit patients and lots of the troubles with current IRB regulations can be addressed by prospective changes in consent processes and communication with the subject. So in so many cases the subjects really would love to have their samples used in the way that scientists want. It's just a matter of putting these things together with the proper prospective consent to make it happen. I absolutely agree with you. However, I would not back down as I would be interested in your thoughts. Certainly the IRB at Yale is being almost run out of town by the amount of paperwork that they're now being required to do. And of course the impact on investigators. Many physician, particularly young physician scientists are looking at what's happening and deciding that mouse looks like a very good organism to be studying. And I think as a practical issue again getting back to the prior comment about physicians and scientists needing to be better organized to play a role in the decision making process is very important in particularly with regard to genuinely protecting patients who participate in clinical investigation rather than documenting what's been done that doesn't add to protection. Maybe if the rest of the people here served on IRBs more, we could have more light and less heat and less paperwork and I agree completely. Rick, all of us that is North Carolina I think you were wrong in one of your predictions. Regarding the usefulness of identifying common variants in existing pathways versus new variants in new pathways I think you're quite correct in saying that that will help in the derivation of obtaining new drugs but you forgot the large advances in prevention that occurred as a result of understanding genomic variation so that individuals who had variants in a large number of genes that were already known in pathways you could prevent the onset of the disease by prevention rather than treatment and that's always more economical in the long run. I couldn't agree more and I certainly agree by 2020 we're still maybe not quite there but in terms of knowing for the truly polygenic diseases whether we've identified all the variants in order to make very good predictive medicine with genes that have very small effect on the trait but I think your point is well taken at the end of the day if you can identify people who have 25 of those 50 alleles who are much more likely to go on that that might be helpful for prevention. My K-Back, San Diego, this is somewhat of a Rip Van Winkle type question I feel as if I was asleep and woke up and heard your talk in the year 2020. That beats the alternative. Yeah, that's true. It beats any alternative. But I guess the question that comes to my Rip Van Winkle mind is where did the regulation and or legislation come to deal with another parameter in this whole scenario and that is the private sector. Much of the technology that we've heard about this morning and the development and innovation that occurred after 2001 had to do with diagnostics and had to do with predictive tests and it had to do with pre-symptomatic tests which should be distinguished from pre-symptomatic tests certainly in terms of public application and it was the private sector at least at that time that kind of had control of that regulation and the potential harm from the perspective of the public was perhaps greater from that perspective than any other. So I wake up and I'd love to hear what happened. Yeah, so I think that's a very good point. I think one of the things that particularly the recent experience may teach us and Jeff Dewick might want to chime in is that private enterprise is very sensitive to public opinion and if private enterprise does things that end up having adverse consequences that will have immediate consequences for the fate of their companies. So in the near term I don't know what is going to happen in terms of legislation but one would hope that companies are in part kept in line by paying attention to what expectations for their bottom line is going to be. Jeff you want to comment on that? I mean the thing to realize is that to do patient collections whether they're for experiments or for clinical trials even in the private sector requires IRB approval and once you enter into the FDA process for the development of a drug it is completely an audible process and you're subject to all the necessary federal regulations and so I think that and if anything the private sector will not make use of any of the genetic data until the necessary commercial and more importantly consumer protections out there that will enable it to actually be commercialized so I think there's a tremendous trickle down and I think at some levels you overstate the issues in the private sector because they're subject to the same kind of oversight and it can pace deeper penalties. If Dr. Lifton you know violates some IRB maybe he loses his job maybe he gets slapped on the wrist if I do that I'm accountable at a financial level that you cannot imagine so I think I think you have to realize there is a tremendous amount of concern about these things. This really goes back to neuroses my name is Karen Rothenberg University of Maryland your very last point about how we deal with information and I think you're being extremely optimistic in the year 2020 that when we get this genetic information as consumers or even as scientists how are we going to figure out relatively what we're going to change in our behavior to make a difference so just for example let's just think about one thing lung cancer and all of a sudden you can give us a prediction about whether we're going to be more or less likely to have an increased predisposition to getting lung cancer are we going to be more or less likely to then smoke to start smoking to stop smoking I mean we already know what causes lung cancer we can't get the public to stop so when you then add to it sort of having to compare whether we care more about dying from cancer or from heart disease and those could be sort of at odds in terms of how we can deal with changing our behavior or taking certain drugs we see this happening now currently going on and they're not based on genetic information they're just based on very well proven common information already out there so tell us how are we going to get the public to be quote rational based on the perspective of scientists in changing their behavior and the way they live their lives and not be paralyzed by the information but rather to do things and there are people in this audience oh it's clapping your hands but I think that's really critical if we're going to make predictions into the future to know how we're going to manage that information I think these are things that physicians and patients deal with every day now as you point out we already know that if you're a smoker you are at markedly increased risk of lung cancer and of heart disease adding an additional piece of information that you have a particular mutation that makes it even more likely to develop lung cancer than someone in the general population is one additional reason for you to think about changing your behavior but we certainly can't prescribe behavior as physicians and tell our patients that they absolutely have to do this all we can hope to do is provide the most useful information that will have practical implications for individual patients as to how they might best try to maintain their health well-being and try to provide them with the information and rationale and persuasiveness that there's someone who pays attention to their well-being and cares about it but at the end of the day these are going to be decisions that individual patients are going to make and I don't think all of the genetic information in the world is not going to change that fundamental point I'm Ellen Clayton and since Jeff has made this comment about accountability twice I feel compelled to respond to it since I did it the first time the accountability of the market is not the same as the accountability of the political setting or the accountability of the moral setting the market is not concerned with justice it is not concerned about the underserved and I think that we have to recognize that the market does care acutely about public opinion it really cares about certain types of public opinion so as we talk about accountability it is not enough to say that the market will take care of it and that the invisible hand will make all ships rise together we've been down that road and it isn't what's happening and I would say that one of the things that's great about this country and great about our morality is a sense of commitment to all in a sense of commitment to trying to ensure that all get the benefits now I'm not arguing for a communist or a completely level playing field but I am insisting that we at least question the ability of the market to do the ultimate the ultimate distribution without any other sort of regulation the reason we regulate in this market is because the market is not the optimal ultimate optimal distributor at least according to some people so let's just I mean we have to get that on the table but the market is not the only answer okay so I'm Bonnie Pagan University of Washington Seattle one thing I'd like to do is just emphasize one of the comments that you made and maybe rephrase it a little and that is I think we recognize that there's a big pool of basic science information over here and a need to apply that in a clinical setting and right now that pipeline is very skinny and the concerns that a lot of us have is that LC and the LC issues that are being addressed are narrowing that pipeline further and I think you said it very nicely that the balance between the good and the possible detriments need to be weighed and I think it's time for the good to fit into that equation and help widen that pipeline rather than narrow it absolutely thank you very much Rick