 Good morning everyone. My name is Carol Kovac. I'm from IBM and I'm going to be your moderator this morning. I guess there's the old saying that says today is the first day of the rest of your life. And so after yesterday's announcement, I'm glad to see that you're all back here ready to start sort of the first day of the rest of our lives after we now have kind of declared the genome complete. I don't know about all of you, but yesterday I found it, you know, fascinating, fascinating retrospective looking back over the last 50 years of discovery, referring back to my own times in graduate school and marveling at how far we have come in just the last few decades. But by the end of yesterday, I think, you know, one of the things that I was ready for was what we started to talk about yesterday was the look ahead. Where are we going? What comes next? How are we going to handle the new data and how are we going to bring information out of that data? From my perspective, in the information technology industry, we are fascinated by the questions of how will the convergence of progress in information technology, which continues to proceed for another 10 years at least along that, you know, exponential growth curve we call Moore's Law, you know, how will the convergence of that with the equivalent exponential scaling laws in biology, which we have just begun to see, they are just at the very beginning of their sort of Moore's Law growth. How will that fundamentally and remarkably change the world over the next decade? And so I think what we're going to talk about all through today is very much more about what comes next. This morning I'm very pleased to moderate a panel, which will begin by talking about the implication of genomics for human disease. And we have with us today four speakers. Unfortunately, Fernando Martinez is unable to be with us this morning. So we will have four speakers. You will have a little bit more leisurely discussion from each of our four. We will have questions and answers, unlike yesterday. So I hope that you will prepare some lively questions and answers for this group of people. Our first speaker this morning is Janet Rowley. Janet is from the University of Chicago, and her work is in the area of, she is a distinguished service professor of medicine, bloom rice, distinguished service professor of medicine at University of Chicago. She has worked on chronic myeloid leukemia and is the recipient of the last award for medical research in 1998. This morning she will speak to us from the genome to successful treatment in leukemia. Well, thank you, Carolyn, for introducing this session. It is a great pleasure for me to be here amongst the speakers on this very auspicious occasion as we look to the future, having now most of the human genome accessible to us. And as I thought about what I would like to talk about this morning, I decided that I would go very, very quickly through the portion of the title that talks about successful treatment and really emphasize the kinds of things that we can now do with the genome in hand. And I will speak in part from the standpoint of leukemia, which is a very small segment of cancer. But I think the analysis of leukemia and research in leukemia has led the way for many of the studies and many of the paradigms developed in cancer. This is a quick overview of that from chromosome abnormalities, the discovery of Peter Knoll and David Hungerford in 1960, that there was a unique abnormality in chronic myeloid leukemia called the Philadelphia chromosome to the discovery that it was really a translocation, cloning of the translocation breakpoints and identifying the genes as BCR-Able. And finally here to successful treatment, and you notice that this is a period of 40 years. This shows you the karyotype of a patient with CML. Here's the Philadelphia chromosome, normal chromosome 22, and you see how much smaller it is. And I discovered that the piece wasn't missing, but rather was moved over here to the end of chromosome nine. And then with heister camp and grofin it was shown that the Abel gene and the BCR gene were involved. And in fact this is the critical new gene that's formed by the translocation, namely the fusion of five prime BCR and three prime Abel. And this leads to the activation of the Abelson gene, which David Baltimore had shown was a tyrosine chymase. This shows you a very schematic drawing of the BCR Abel fusion gene. And Abelson has a pocket that binds ATP and is involved in the enzymatic transfer of a phosphate to tyrosine on a substrate. And this is the beginning of the pathway that leads to activation of all of the target genes resulting in chronic myeloid leukemia. The miraculous thing of STI 571 is it binds to this same pocket, preventing, unfortunately you can't see it over here, but this is ATP here, prevents the binding of ATP and therefore this activation of the pathway does not occur. This shows you the data upon which the approval of Gleevec or Imantanib is now called. This was based where 532 patients who had failed interferon, which was the best therapy available at the time of these patients. And there's anyway, I think this margin of all the slides is clearly going to be cut off. This is 88% of the patients showed a complete hematologic response and almost half of them excited genetic response. And so this was actually approved in 72 days, which is obviously a record for the Food and Drug Administration. So this is the good news, but I should point out that this success occurred with relatively little input from DNA. Clearly DNA was used to clone the breakpoint and identify the genes and identify the fusion gene, but it was sort of a peripheral accomplishment, if you will, or had a peripheral input. Now when you look at the future, here we have a number of karyotypes of both hematologic diseases and solid tumors. The number of recurring rearrangements is greater than 300 in the hematologic diseases and greater than 100 in the solid tumor. We've cloned many of the genes here at the translocation breakpoints, but except for the treatment of acute promyelocytic leukemia with a drug that targets retinoic acid, we have no genotypic specific therapy. So the challenge for the future is to get a GLEVAC for each one of these different rearrangements. And how we're going to go about that is obviously a matter of using a number of strategies. One that you heard about extensively yesterday is to look at microarrays. This is an example from acute lymphoblastic leukemia published from St. Jude's Hospital. Here, as you will recall from yesterday, patient samples are on the vertical, genes are on the horizontal, and they're segregated by particular chromosome abnormalities. And as was emphasized yesterday, there is a unique fingerprint for each of these translocations and rearrangements in patients that have similar rearrangements. The concern is that the genes or the oligonucleotides on the chip are more only those that we know about. So it doesn't tell you anything about what you do not know. And it's difficult to quantitate and compare from one experiment to another. So, Simon Wang in the lab decided that he would try a different strategy that's using SAGE. This was developed by Ken Kinsler and Bert Vogelstein at Hopkins, and we select only poly A plus RNA, select the 3 prime end, use restriction enzyme to cut at the last C-A-T-G site, and then it cuts 10 base pairs downstream so that in our selection we're getting the last C-A-T-G site before the poly A tail from all these different transcripts. And of course you're not selecting for the transcripts except they have to be polyadenylated. Here you have the individual SAGE tags, you ligate them and sequence them. Then based on sequence information of each tag, you can go back and match it to the genome. And again, this is where a public database is so extremely important. And you can sequence multiple SAGE tags in a single lane so that we do 30 to 40 as compared with doing CDNAs one at a time. This is a summary of the work that has been done in the laboratory. These are normal human cells, both hematopoietic stem cells, myeloid progenitor cells as well as lymphoid cells. And then we're in the process of looking at leukemia, the most common chromosome translocations in acute myelogenous leukemia. And here we're looking at five patients each one selected for CD15 positive. You'll notice that we've done more than 100,000 SAGE tags in most of these libraries, that is individual SAGE tags. When you look at the origin of the SAGE tags, you can see for the normal 34 and 15 positive cells that 12 and 14 percent are known genes, roughly 40 percent are ESTs, but almost half are novel transcripts, that is they're not present in any database, either GenBank or ESTs. And therefore, of course, none of these will be on a microarray. Just in summary as to where we are with the leukemias, we have compared the level of expression of the various SAGE tags with the normal CD15 positive libraries. And you can see that for the different translocations, a number are present or up-regulated, absent or down-regulated. And one of the things that's astonishing to me is that the 1517, which is the translocation associated with acute promyelocytic leukemia is distinctly different in its pattern of transcripts than the other two types of leukemia. And clearly we will be involved in analysis of this in more detail when we finish the study, which includes the 821 translocation. There are also some transcripts that are common in all of them. Five of them are up-regulated and 82 are either absent or down-regulated so that these would appear to be transcripts that may be more directly related to the leukemic condition and less related to the specific subtype of leukemia. I want to come back to the novel SAGE tags. And here you can see that there are 50,000 SAGE tags. I didn't point, a novel SAGE tags, I didn't point out in the earlier slide, but this is out of more than a million and a half total SAGE tags that have been identified. These are novel SAGE tags that are unique. And of interest we have been studying Drosophila as well. And here you can see that though the Drosophila genome is well studied there are more than 22,000 unique novel SAGE tags in Drosophila from a total analysis of 350,000. So this is an image of a nucleus stained with DAPI. So the DNA in this nucleus is what is stained blue on the screen. And as you heard yesterday, 1 to 2% of the DNA, roughly 30,000 genes or so, though this is a number that is in some dispute, but 1 to 2% codes for protein. So the question is, what does the other 98% do? And I've always been fascinated with astronomy, cosmology and the wonderful images that they can show and the terms they use, black holes and supernova and of course dark matter or dark energy. And I submit to you that we've got our own dark matter. What's the 98% doing here? And I thought the comment of the quote yesterday from Sidney Brenner, there is a distinction between junk and garbage. And so I assume that this 98% is junk, not garbage. And the challenge to us is to figure out its function. And clearly the genome is going to be, and the accessibility of the DNA sequence is going to be an extraordinarily powerful tool to help us with that challenge. So why are there so many novel sage tags? They can be novel genes not identified using current algorithms, and that's certainly going to be the case for some. Transcripts derived from alternative splicing, and we heard an example yesterday of one with 38,000 possible splice variants. Transcripts of non-coding RNA, both antisense, intergenic transcripts involved in imprinting. And I want to emphasize this, the transcripts are present often at very low levels, and they won't be detected by many of the present techniques that we are using to identify transcripts. So the transcripts may be novel genes. And this is an example from the laboratory. So Jin Jin Chen and individuals in the laboratory did a great deal of sequencing, and here was a particular sequence they found, which they could then go and interrogate the database using BLAT, again emphasizing why it is so important for the scientific community to have free access to all of the genomic information. And the sequence here you see in the black bars clearly looks like exons with entronic sequence. It codes for an open reading frame of 79 amino acids, and the direction of transcription, as you can see from the arrows, is from left to right. You also see using all of the programs and the known ESTs that there are no matches, so it does look to be a novel gene. I'll skip over the alternative splicing, though that's a fascinating area, and move on to transcripts of non-coding RNA, anti-sense RNA. So again, taking a sage tag and through strategies that we have of extending them so we get far more sequence information. Here is another example of sequence mapped back using BLAT. This one is on the long arm of chromosome 15, and you'll notice that the arrows in the direction of transcription are from right to left. You'll also notice that it matches a known gene precisely. This is the beta 2 microglobulin gene. So this is an example of expression of anti-sense to beta 2 microglobulin. It's of interest that as we have done our studies in Drosophila, 38% of all of the novel transcripts in Drosophila are anti-sense, so that I think this is a whole area of biology which we have not been aware before. So what might anti-sense be doing? I thought this was a marvelous example from a paper from Kramer in Nature this year showing the reciprocal expression of sense in anti-sense RNA in Neurospora. This shows you part of the map of the frequency gene, it's a circadian rhythm gene in Drosophila, the sense and the anti-sense strand. And this shows you the level of expression of sense transcripts and anti-sense transcripts over time. Anti-sense transcripts are induced when Neurospora is in light. And so that they're very high in light and as the Neurospora are put in dark, they reduce. And the authors propose that the anti-sense strands are annealing to the sense RNA and regulating the transcription of that gene. So this is endogenous RNAI functioning, and this is one of the first examples which I'm aware of showing that anti-sense is a very critical component of gene regulation, and one that I think we haven't gotten into. So I think we've reached a new frontier. And that new frontier is a change in the paradigm. Now this is an example from a paper from the classical paradigm, or as we heard yesterday, classical dogma with DNA to RNA to protein and protein regulating DNA and RNA. And we now have to add in the role that RNA is playing within the nucleus, regulating not only itself, but also regulating DNA through imprinting and other of the regulatory processes. So I think in terms of the relevance to cancer, the question is are we focusing on the right targets? We've been focusing on protein coding genes. Are there other things we should be paying attention to? And if there are other things, are we using the right strategies to find them? Because ultimately our goal is really to come back to the adequate management of cancer patients in the year 2000, whatever. And I'm not going to say I have a crystal ball that knows how long we're going to get there, or at least for every patient. But we're certainly going to be genotyping the involved tissues and being much more precise in our assessment, even the breast cancer dividing it into multiple things and looking at gains and losses and translocations. We'll also then institute genotypic specific therapy, such as GLIVAC. And the hope is that this will lead to less toxicity and improve survival. And I've already commented on the colleagues in the laboratory, particularly San Ming Lang, who were very much involved in the SAGE project, long-term colleagues in the laboratory. I want to acknowledge that this, in version 16, analysis and functional assessment is being carried out with Pupol Yu here at NIH with support from Jeff Trent, Eric Green, and Gerard Buffard. And I acknowledge the support both of NIH as well as a number of foundations and the University of Chicago. Thank you very much. Thank you very much, Janet. Our next speaker is Richard Lifton from Howard Hughes Medical Institute, and Richard is chair of the Department of Genetics at Yale University. And he will be speaking to us today. His research is in the area of cardiovascular disease, hypertension, stroke, osteoporosis. And he will speak to us today about genetic and genomic insights into cardiovascular disease. Thank you, Carolyn. It's a real pleasure to be here today, and I'd like to add my congratulations to NHGRI and the sequencers for a remarkable set of achievements. Having served on the External Advisory Board for the project for the last five years, it's really been a remarkable opportunity to see this unfold. And given all of the opportunities for failure that presented themselves along the way, I think it's really testimony to all involved to have accomplished this goal in such a remarkable fashion. So my task this morning is to talk about cardiovascular disease. And if I could have the first slide on, please. Cardiovascular disease remains the leading cause of deaths in the industrialized world, including the United States. About a third of all people in this room will ultimately die of heart disease or stroke, indicating that we have a lot of work to do. The pathophysiology of heart disease has been defined in rather gross terms by epidemiologic studies, which have succeeded in identifying a number of risk factors for the development of atherosclerosis. These risk factors include hypertension, hyperlipidemia, smoking, and diabetes. These result in atherosclerosis, leading to narrowing of the arteries that supply blood to the heart, resulting in their closure in myocardial infarction, which then results in heart failure, cardiac arrhythmias, and ultimately death. Well, the specific causes that result in these epidemiologic risk factors, we now have the opportunity to begin to identify by genetic approaches. So Brown and Goldstein have pointed out that even the most ardent proponent of the healthy lifestyle proposal would have a hard time explaining the paradox of Jim Ficks and Winston Churchill. Jim Ficks, 5'10", 150-pound marathon runner, healthy lifestyle promoter on the one hand, Winston Churchill on the other, 5'8", 270, slothful, legendary for his gluttony, a heavy smoker. Nonetheless, Ficks died at age 52 while running, and Churchill lived to the ripe old age of 90. And of course, one possible contribution to this explanation for this paradox is that Ficks had a remarkable family history of early coronary artery disease. And we expect that the risk factors for cardiovascular disease will in large part have genetic underlying components. And the opportunity then that presents itself is to use genetics to try to identify these underlying causal factors. We specifically can identify causal relationships between genes and pathways and health outcomes, can identify new targets and pathways for early diagnosis and therapeutic intervention, and ultimately may be able to identify individuals with increased likelihood of benefit and minimize risk of adverse effects from the treatment with specific agents. So that broadly speaking, the genetic contributions to disease causation, Mendelian traits, typically rare, single genes with large effect on the one hand, on the other hand, multifactorial traits where there are multiple inputs from genes and environmental factors. This is where the bulk of disease burden lies. Going forward, the optimal strategies for tackling these genes and identifying them is not well understood because we don't know how many genes influence the trait nor the magnitude of the effect imparted by any single locus. And we also don't know whether they're common or independent alleles. Most of the progress to date has been made in the understanding of these rare Mendelian traits, but I would make the specific pitch that understanding these rare diseases may be of critical importance because it may point us to key gene products and pathways that can be manipulated for health benefit. And importantly, the treatments that arise from these may be highly applicable to the general population and need not be confined to the rare patients with these diseases. So trying to put a molecular face on these epidemiologic risk factors is one of the major enterprises of the last decade, and there have been a number of striking successes. You're all familiar with the work of Brown and Goldstein who identified mutations in the LDL receptor underlying homozygous familial hypercholesterolemia. This is a disease that affects one in a million in the general population, but it established in a formal way the causal relationship between elevated cholesterol and early coronary artery disease and validated the approach of attempting to antagonize the cholesterol biosynthetic pathway by identifying specific antagonists of the rate limiting step in cholesterol biosynthesis, leading ultimately to the development and use of HMG CoA reductase inhibitors. These are drugs that have been demonstrated by randomized controlled trials to have a dramatic impact on reduction of cardiovascular risk, and we're now treating tens of millions of individuals worldwide with drugs that were validated from the understanding of a disease that affects one in a million in the general population. In the last several years there have been dramatic advances in understanding other aspects of the cholesterol homeostasis pathway. Again, from the understanding of very rare diseases. For example, Michael Hayden and colleagues identified mutations underlying tangier disease as being in a specific member of the ABC transporter pathway that is involved in the export of cholesterol out of cells, the so-called the reverse cholesterol transport pathway, getting cholesterol out of cells and into HDL, understanding of a new pathway underlying hypercholesterolemia. Similarly, Helen Hobbs and colleagues have identified mutations underlying another rare disease, testosteroneemia, as members of a related gene family of ABC transporters that are involved in secreting plant sterols out of the liver into the bile, and in the absence of these plant sterols which are abundant in our diet are abnormally absorbed and can't be secreted, resulting again in another form of hypercholesterole of high sterol levels and early coronary artery disease. These are nice examples of rare diseases that are leading us to new understanding of basic human physiology. Our own work has been involved in understanding the pathogenesis of hypertension. This is the most common disease of the industrialized world. It affects more than 20% of the adult population and is a major risk factor for death from heart attack, stroke, end stage renal disease and congestive heart failure. Its pathogenesis has been unknown and its treatment has been inadequate. Part of the reason for the difficulty in understanding its fundamental causation is the complexity of the regulation of blood pressure. This is Arthur Guyton's model of the regulation of blood pressure in humans, and as you can see this is a rather complicated schematic and it's difficult to reduce to more simple terms because unlike cancer this is a phenotype you can only measure in an intact living organism. So as a consequence, hypertension has variously been proposed to be a disease of the brain or of the heart or of the kidney or of the adrenal gland or of the vasculature, and our approach has been to try to settle this using genetic approaches, looking for rare families segregating single genes with very large effect. This has succeeded in identifying mutations in a number of genes that affect blood pressure in humans, and the intriguing aspect of this work is that all of these genes act in a single, final, common pathway that regulates how the kidney handles salt. Every day the kidney filters one and a half kilos of salt on a typical Western diet. We have to reabsorb all but about a half to one percent of the filtered load in order to maintain salt balance. The kidney does this by a highly coordinated pathway of regulated co-transporters, exchangers, and channels. And this whole process is regulated by the renin angiotensin system, whose output angiotensin 2 causes increased secretion of a steroid hormone aldosterone, which interalic corticoid receptor, whose activity regulates the activity of this epithelial sodium channel. To date, all of the mutations that affect blood pressure in humans act in this final common pathway, mutations that raise blood pressure act by increasing net salt reabsorption in the kidney, mutations that raise blood pressure do the opposite. And the pathophysiologic sequence is increased salt reabsorption, results in increased plasma volume due to water following the sodium to maintain isogenicity of blood. This expands plasma volume, increases venous blood return to the heart. The heart obligingly pumps out more blood, and by Ohm's law this raises blood pressure. So all of the mutations act in this fashion. And I'll give you an example of a few of these following this pathway down, the mineralic corticoid receptor and the epithelial sodium channel. So by studying one-of-a-kind family, we identified mutations in the mineralic corticoid receptor in which substitution of a single amino acid in helix 5 of the receptor substituting leucine for serine results in severe hypertension. We've been able to demonstrate at the atomic level that the mechanism of this is that this leucine substitution creates a new Vanderwall's interaction that eliminates the requirement for steroids having 21 hydroxyl group to interact at this portion of helix 3. And so as a consequence this single amino acid substitution changes the specificity of the receptor such that steroids that normally would bind but fail to activate the receptor are now potent agonists. Similarly, mutations just downstream of this receptor, the epithelial sodium channel itself, mutations in this channel cause an early severe form of human hypertension. And the mechanism of this is that these mutations that either modify or chop off this cytoplasmic tail of subunits of the epithelial sodium channel, eliminate the ability of the channel to be cleared from the cell surface by endocytosis via clathrin coated pits. And as a consequence, these channels accumulate at high levels on the cell surface, mediate increased sodium reabsorption and result in hypertension in humans. Most recently we've identified by positional cloning an entirely new signaling pathway not previously recognized by physiologic study. This turns out to be a negative regulator of sodium transport and it's a serine threonine kinase and there are two of them that are involved in the regulation of salt reabsorption. They normally antagonize the reabsorption of salt by a specific co-transporter, the sodium chloride co-transporter and mutations that eliminate just the function that regulates the sodium chloride co-transporter result in hypertension in humans as well. And this has identified a new signaling pathway that we're still trying to understand what its general role is in human homeostasis. Having identified a bunch of these genes, we can begin to test the proposition that this matters in terms of clinical applications. Some of the genes we have identified are autosomal dominant in action and we can identify extended pedigrees based simply on the identification of an index case and genotyping through the extended families. These patients are typically diagnosed with severe hypertension that is refractory to treatment but knowing the specific underlying diagnosis we can begin to treat these patients with selective agents that are tailored to their underlying diagnosis. A nice example of tailored therapy. I'd like to turn now to some of the outcomes of cardiovascular disease. One that I mentioned is congestive heart failure. This shows a massively enlarged left ventricle caused by a specific mutation in sarcomere genes and this is the work of John and Cricket Seidman who have taken a comparable approach to what we have done in hypertension to understanding left ventricular hypertrophy. By collection of rare families from around the world, they have demonstrated that this disease is largely the result of mutations in genes involved in force production in the myocardial cells. As a result of these mutations, force generation is impaired and this is their proposal that the ultimate mechanism by which this results in cardiac hypertrophy is altered regulation of intracellular calcium. These are beginning to flesh out the specific pathway by which the heart adapts to both normal and abnormal force generation and how this common phenotype of cardiac hypertrophy which contributes both to heart failure as well as cardiac arrhythmia arises. Similarly on the outcome side, cardiac arrhythmia is one of the major causes of death in the setting of heart attack. The reasons for this are that the electrical impulses that normally flow through the heart in a coordinated fashion to lead to contraction in a coordinated way become abnormal and the normal depolarization and repolarization goes awry resulting in ventricular fibrillation and that is ultimately leading to inability to propel blood through the body and death. Mark Keating and his colleagues again have taken this approach of looking for rare families with abnormal heart rhythms and have identified a number of genes that underlie these traits and intriguingly all of these genes fall into the pathway of being the specific mediators of components of the cardiac action potential. So mutations that impair, that prolong the influx of sodium at the beginning of cardiac excitation cause cardiac arrhythmia and similarly mutations in a series of potassium channels that normally mediate the specific repolarization of myocardial cells also cause cardiac arrhythmias. Importantly these have already begun to have impact on our understanding of pharmaceutical complications. Long QT syndrome induced by drugs is a common mode of failure of drugs in clinical trials and the Keating lab has identified one of the specific potassium channels that appears to be the target through which this acts and virtually all drugs are now being tested for their ability to antagonize the specific potassium channel. So thinking about the future of genetics and cardiovascular disease, we have come a very short way but we have done it I think quite rapidly. We have begun to identify new genes and pathways and I think going forward we need to continue to pursue all of the known Mendelian traits recognizing that these might provide very key insights that we hadn't in understanding human physiology that might identify key targets that can be manipulated for health benefit. It goes without saying that there are likely to be many of these that we haven't yet recognized because we haven't looked carefully enough and we need to have eyes wide open. Similarly there's another area, genes with large effect but reduced penetrance which has up to now been systematically underexplored. I think there are opportunities to find these genes, both autosomal dominant and recessive forms and then the granddaddy problems, the multifactorial susceptibility genes where as we heard about yesterday from David Bentley we will have the opportunity hopefully in the future, in the not too distant future to be able to do genome wide linkage disequilibrium studies that may help us identify these multifactorial genes that are common in the population but have small incremental effects on traits and in addition, as we heard about from Pat Brown, genomic approaches to identifying of the identification of these pathways will likely also be of critical importance. And then at the end of course there's a critical need to translate these validated targets and pathways into new types of gene that are most likely to be developed by the gene of the Peter Goodfellow. Among the challenges that we face going forward is that there's a paucity of clinical investigators and we need to improve this pipeline. We need larger and better phenotype cohorts, the tools that are being developed by the genome project are going to be most readily applicable to very large well phenotype cohorts. We're woefully unprepared for this task in most investigations. It would be tragic if the fruits of the labors of the genome project are thwarted in their application because we are unable to overcome some of the regulatory barriers that we now face and ultimately we need a better handoff from academia to the pharmaceutical industry. However, based upon the optimism that carried the genome project over the last 15 years to where we are today, I confess that even a hardened pessimist would be very rapid and stunning pace. Thanks very much for your attention. Our next speaker is Nancy Cox. Nancy is a quantitative human geneticist and professor of human genetics and medicine at the University of Chicago. She'll speak to us today about her research in diabetes and the title of her talk is Paradigms for the Genetic Component of Diabetes. Thank you and it's really exciting to be here as part of this terrific celebration. I'm particularly excited to be here to talk about what I always call a poster child for complex disorder, complex disorders, diabetes. Diabetes is not really a single disorder, but many heterogeneous disorders, all of which are characterized by high blood glucose levels. Diabetes is common as are many complex disorders affecting large numbers of individuals in the United States, more than 15 million, and worldwide more than 150 million. Diabetes arises collectively as the consequence of the actions and interactions of many different genetic and non-genetic risk factors. Some forms of diabetes result from mutations in single genes, so there are monogenic forms of diabetes, similar to those for other complex phenotypes as Rick Lifton discussed. But the majority of diabetes is multifactorial polygenic in origin. And as with most complex disorders, we all hope that the identification of genetic risk factors for diabetes will lead to improved diagnosis, treatment and prevention strategies that address the underlying molecular defects rather than just the broad symptoms of the disorder. And there are already indications that this will in fact be the case for diabetes. Preliminary studies indicate that some of the rare monogenic forms of diabetes, maturity onset diabetes, the young, has at least if not more effective treatment strategies, at least some that are more easily tolerated by patients. So genetic identification of particular disorders is already paying off in terms of treatment strategies for individuals with diabetes. Diabetes is increasing epidemically worldwide now. What is shown in the top box is the number of individuals in the millions affected with diabetes in the year 2000. In the second box, the projections for the year 2010 and the last box, the percentage increase. And I would draw your attention to several aspects of this epidemic. One is that ironically, as developing nations adopt the modern western diet and lifestyle, their per capita rates of diabetes actually exceed those of the western societies. And therefore the percentage increases that are projected for many parts of the world will have really catastrophic implications for public health care in areas that can least afford to deal with these problems. It was also horrifying to me to learn at a recent meeting that these projections were made largely without reference to epidemic increases in obesity that are also projected. So as bad as these increases look for the increase in prevalence of diabetes over the next ten years, they are probably underestimates based on the increasing prevalence of obesity. What I want to try to go over today are some of the paradigms that have been used for gene identification in diabetes, some of the traditional paradigms such as candidate gene studies and linkage mapping and positional cloning, but also some of the novel paradigms that are already being used in studies on diabetes, some of which were brought up yesterday in presentations by Eric Lander and others. And I also want to focus on how what we are learning about diabetes susceptibility loci are changing our perceptions, shaping our perceptions about what we ought to be looking for and how that is modifying the paradigms for susceptibility locus identification going forward. So candidate gene studies have been successful in diabetes. This is a very abbreviated list of the genes that have been identified through candidate gene studies and designed not to be exhaustive, but just to illustrate some key points. And unfortunately the first point I would make is that based on my own unscientific survey, the success rate of candidate gene studies thus far has been about one in a thousand. That's true not just for diabetes, but for complex disorders in general. And I think a key point to remember going forward is that we will get smarter about what the right candidates are. So while candidate gene studies have not been tremendously successful so far, the success of the Human Genome Initiative and the biology that will follow that will improve our ability to do intelligent candidate gene studies. And those will probably be the most efficient ways of identifying susceptibility genes in the not too distant future. Points that I'd like to make from the candidate gene studies in the maturity onset diabetes of the young, which is a rare monogenic form of type 2 diabetes, is the observation that many of these candidate genes were identified as a consequence of original linkage mapping studies. So for example, the first transcription factor identified for maturity onset diabetes of the young was a result of a regular, fairly brute force positional cloning study. But once the first transcription factor was identified, many additional transcription factors have been shown to give rise to maturity onset diabetes of the young. And so we have a much more effective candidate gene strategy there. The candidate genes for identified in type 1 diabetes, HLA, was identified in the 1970s in classical case control studies and really helped to clarify the nature of the genetic heterogeneity in diabetes by distinguishing type 1 diabetes from type 2 diabetes. So most, most diabetics have a later onset form of diabetes, termed type 2 diabetes, where there's a relative insufficiency of insulin leading to elevated blood glucose levels. About 7 to 10 percent of diabetics have this form of autoimmune diabetes, type 1 diabetes, and only about 3 percent of individuals with diabetes have monogenic forms such as modi or mutations in the insulin or insulin receptor locus. So HLA was identified in classic case control studies and is a major risk factor for type 1 diabetes. The insulin gene region, the insulin gene region VNTR was also first identified in case control candidate gene studies, but was the first and perhaps best success of the family-based methods for association studies. The TDT was developed and applied first to the insulin gene region in type 1 diabetes. Both of these loci can be detected in linkage studies. So the insulin gene region takes about, takes more than 700 families to detect significant evidence for linkage. The HLA locus in type 1 diabetes takes about 50 families to detect significant evidence for linkage. So you can see that there's a real disparity in terms of the magnitude of the familiarity contributed by these two loci. But in fact, in large-scale linkage studies where the LOD score at HLA and more than 700 families was 65, the next, the locus with the next highest LOD score was in fact the insulin region which had a LOD score just over four. So that gives you a sense of the frustration that we have in type 1 diabetes because of the linkage mapping of the susceptibility loci after collecting many, many families, the linkage mapping identified two loci, two regions that we already knew about from candidate gene studies. Again, for type 2 diabetes, these particular candidate genes that I'm focusing on are not meant to encompass all of those that have been identified in candidate gene studies for type 2 diabetes, but rather to illustrate the evolution in our understanding of the underlying pathways. So in the early days, when we understood much less about the pathways involved in glucose homeostasis, the insulin gene, insulin receptor gene were the obvious candidates. Now as we understand better the pathways of glucose homeostasis, loci like P par gamma come to the forefront, again illustrating that as we understand the pathways better, we will get smarter about the candidate genes. So what have we learned from these candidate genes? Do these studies inform our understanding of diabetes? How do they help us to generate new hypotheses about what to look for, utilizing the data emerging from the human genome initiative? I think a key point is thinking about what we have learned from the genes that can give rise to this complex metabolic phenotype through single gene mutations. And in particular the paradigms that emerge from the study of the identification of transcription factors as genetic susceptibility loci for maturity onset diabetes of the young offer some interesting and important hypotheses for the kinds of loci that may generate linkage signals in diabetes in general. So if we think about these loci, they're higher level regulatory genes. They regulate the expression of many other genes important in glucose homeostasis. With the new data that will be emerging from the human genome initiative with the description, fuller description of pathways involved in metabolism and glucose homeostasis, it would be possible to test the hypothesis that higher level regulatory genes, whether we mean regulatory at the level of at the level of the DNA or regulatory at the level of proteins or as Dr. Raleigh referred, RNA, the hypothesis that the higher level regulatory genes will be the ones that generate linkage signal while those genes and proteins that they regulate will be the polygenic background is one that we'll be able to test in a systematic way going forward. Regulation at a different level, I think is illustrated by the emerging understanding of how the insulin gene region variation affects susceptibility to type 1 diabetes. So in contrast to the very rare mutations that give rise to susceptibility to monogenic forms of type of diabetes including maturity onset diabetes of the young, the allele increasing the risk to type 1 diabetes at the insulin, the NTR has a frequency of 0.75 in populations of European descent. That is a very common risk factor. This variation leads to different, differential expression of insulin in different tissues in the body. So in the beta cells of the pancreas the protective class 3 VNTR alleles are associated with lower insulin mRNA levels than the diabetes predisposing class 1 alleles. Whereas in the thymus the protective class 3 VNTR alleles are associated with higher insulin mRNA levels. These results are consistent with the model in which high level expression of insulin in the thymus facilitates immune tolerance induction resulting in dominant protection for type 1 diabetes. The association of this variation at the insulin locus with type 1 diabetes has been known for more than 15 years and has been essentially universally believed certainly since the studies using the TDT. It took a very long time to generate these observations and this hypothesis. And I think even with the biology that will come out of the human genome initiative we will need to be patient in understanding how variation at susceptibility loci for complex disorders actually generates the differences in risk that we see. And I think another point that we maybe need to keep in mind going forward is that we won't always need to understand or even identify with certainty the exact variation that affects susceptibility to disease. If we can provide enough evidence that the loci are actually the key susceptibility loci and disease that alone will help us understand pathways better and open up new understanding of the disease processes. Another traditional paradigm for identifying genetic variation for monogenic and complex disorders is the linkage mapping followed by positional cloning. The strategy has been spectacularly successful for the monogenic subtypes of diabetes as it has for hypertension and cardiovascular disease and other forms of cancer. The monogenic subtypes of complex diseases were very amenable to this strategy. And in many cases the linkage mapping was followed by immediate gene identification because of the location of a candidate gene directly under the linkage signal. So this was true for example for glucokinase in the linkage mapping for one of the MODI genes. In other cases linkage mapping requires more brute force positional cloning as with the first transcription factor identified for MODI. In complex disorders there's a lot of additional complications to using this general approach. In the first place the linkage mapping has been far less universally successful. In general our signals are much more modest even in very large data sets and the resolution of linkage mapping is very broad. So we are looking for genes in a very large physical region for the complex multifactorial polygenic components to type 1 and type 2 diabetes. One of the responses to these difficulties has been the establishment of consortia for linkage mapping in both type 1 and type 2 diabetes. These consortia involve virtually all of the world's investigations for linkage mapping and these disorders encompassing investigators from academia and industry and I think it allows us to consider not just the very large scale pooling of data to facilitate linkage mapping but a key strategy will be to use more phenotypically homogeneous subsets of families for example defined on the basis of BMI or aged onset to increase both the power and the resolution of linkage mapping. The advantages of this sort of strategy the linkage mapping followed by the positional cloning is the ability to identify genes and pathways outside of our current understanding of the pathophysiology of disease. This will also be true for the newer approaches that will be based on genome-wide association mapping or linkage disequilibrium mapping that we all foresee as an outcome of the studies from the Human Genome Initiative. So the advantage of being able to identify new genes will be new pathways will be really important. This was one of the outcomes of our own studies in type 2 diabetes identifying variation at CalPain 10 as being associated with type 2 diabetes. CalPains are processing proteases that cleave proteins usually at a single site to either inactivate or activate the protein. And so in some ways the general take-home message from CalPain 10 as a diabetes susceptibility locus is analogous to what was observed for the transcription factors in Modi. This is regulation of proteins rather than regulation of genes at the level of the DNA but it's still about regulation at a relatively high level. The models that we characterized for how CalPain 10 variation affects susceptibility to type 2 diabetes are now while they were fairly radical departures from expectations when we first proposed them they're now pretty mainstream and characterized in positional cloning studies for other complex disorders. So we identified primarily non-coating sequence variation as necessary to characterize risk and multiple polymorphisms need to be considered simultaneously. These sorts of models have now been similarly identified in genes positionally cloned for schizophrenia for asthma for stroke and I think this is raises the importance of understanding functional elements in the DNA that we have not previously recognized as functional. So non-coating sequence variation I think will continue to be very important in understanding the risk for complex disorders certainly type 1 and type 2 diabetes for complex disorders as well. I'd like to talk now about some of the novel paradigms that were discussed yesterday including evolutionary approaches so there's a long-standing hypothesis in diabetes that the same variation that now increases risk for diabetes once had an evolutionary advantage for diabetes it's easy to the original hypothesis from Jim Neil the expression of variation that optimized glucose homeostasis. So broadly speaking the same hypothesis is relevant for perhaps relevant for type 1 diabetes another autoimmune and inflammatory disorders in the sense that the environment has changed with respect to exposure to pathogens and the development of the immune system. So I think evolutionary studies have some really interesting possibilities for an independent identification of regions that may be implicated in disease susceptibility. Studies in Anna DiRienza's lab on CalPain10 have led her to focus on a region within the gene where she believes there's a signature of natural selection these observations are coinciding with observations in my own lab looking at the difference in the magnitude of linkage disequilibrium between cases and controls that also localize susceptibility to this particular region of an intron and both of those are converging with studies initiated by others looking at expression as a phenotype with respect to CalPain10 so studies in Graham Bell's lab are focusing on this variation that seems to affect expression of CalPain10 and again these studies are localizing this variation to the 3 prime half of the gene with further investigations necessary to know whether it's going to localize within this same intronic region so those are the kinds of studies that I think will be important in complex disease studies in general closed by with reference to some of the things Shirley said yesterday as a quantitative as one of the more quantitative of the biological sciences genetics has always drawn a lot of purely theoretical interest but I think one of the lasting legacies of the original Watson and Crick discoveries and the the legacy of the progeny of Watson and Crick theory and genetics will not be done in a vacuum you all have generated so much data that we will never be able to use to have theory that is not responsive to all of the data that's been generated I just like to close by thanking my colleagues from both the theoretical and the molecular who keep us grounded thanks thank you very much Nancy our last speaker Adrian is professor of human genetics at the welcome trust center for human genetics at Oxford he is going to be speaking to us today about gene identification for infectious disease resistance the first 50 years thank you very much and I'd like to thank very much the organizers for inviting me to this superb celebration it's an honor and a privilege I could have the slides on please thank you I must say I found yesterday both entertaining and very enlightening and was particularly interested in the talks on comparative genomics which clearly is making huge strides very rapidly and it was clear that evolution is really conserving selecting for prioritizing certain types of genes over millions of years and those can be summarized as genes involved in immunity resistance and in reproduction so the way I would summarize that is that evolution really only cares about two things infectious disease resistance and sex and I would submit to you that if you're going to be interested in just two things in life that's not a bad selection it certainly works for me but thinking about sex I'm afraid the problem here is I've only got time to talk about some of the subjects and you're not going to get any well I see that Peter Goodfellow is up a little later just after lunch and maybe you'll be lucky then so what I'm going to talk about is this and this is really a problem infectious disease mortality and this is what happened in the last decade 1990 to 2000 TB winning the prize at the top of this gory list for the most deaths 30 million deaths but the point really is we haven't cracked these diseases there's a huge amount more we need to know about them we need new interventions we need many new vaccines and we need to understand more about their pathophysiology and of more fundamental interest we need to know more about what they have done to us and to our genomes and this field as I try to suggest in my title some events happened in the other place in Cambridge in England that lead to us being here today now for those of us well many of you will be aware that for those of us in Oxford in England there has been a certain amount of rivalry between these ancient universities over the last 700 or 800 years or so and I guess if that event happened today there would be university vice-chancellors and what we would do in 1953 life was a little simpler so what actually happened was that a young research fellow by the name of Tony Allison finishing his D. Phil studies set out for the tropics he spent most of 1953 in East Africa and did some studies that were summarized in his publication submitted by the end of the year that appeared in the British Medical Journal in February of 1954 for the Cichl hemoglobin and malaria and what Tony did with characteristically was a characteristically direct study where he identified 30 Luo individuals in Nairobi and gave them malaria these 30 were not randomly selected half of them were heterozygous for Cichl hemoglobin the others were not and he showed very dramatically and very clearly that there was a difference in susceptibility between malaria on malaria administration arguably the first demonstration of a polygene in a complex disease he also did an epidemiological type study one of many since showing that there are lower parasite rates in individuals heterozygous for Cichl hemoglobin and lower parasite densities as well so that started a bit of a cottage industry which has grown since and I think this slide shows that the hit rate in malaria and even though I wouldn't stake my life on maybe more than half of these certainly eight or nine of them are definitely right and arguably this is the best summary of many genes being involved in any complex disease you can see on the far side that there are genes involved mainly in red cell invasion and parasite growth and then of more interest to me over the years genes that affect the immune response to malaria including our own work on HLA are associated with resistance to severe malaria this we feel is more important in that we can think more easily about interventions that you might develop looking at the immunologically relevant genes like HLA B and we have tried to develop vaccines following that strategy TNF the work of Dominic Kodkowski's lab has led to another type of intervention the use of anti-TNF in African children with severe malaria but we need to see more genes we need new interventions the candidate gene approach has worked in other diseases as well this is just a brief and incomplete summary of genes identified by case control analysis in TB pneumocouple disease, HIV largely worked from NCI and other units in America and hepatitis B but the point I want to make about all of these is these are not genes with effects of the magnitude of sickle hemoglobin that gives you around about 90% of these small effects and by and large I think this is what you will find and what we do find in the study of infectious disease resistance which makes the exceptions all the more exciting there are only four I know of I've talked about one already sickle hemoglobin then Lou Miller's work on the Duffy blood group and vivax malaria most sub-Saharan Africans are Duffy negative there are 32 mutation in the chemokine receptor 5 C-C chemokine receptor and then many of you will be aware that there's a kind of long-running experiment in progress in Britain at the moment those of us who meet in the 1980s are being followed up a hundred or thereabouts of us have developed the human equivalent of mad cow disease due to a pion that also in humans gives new variant tritial Yakov disease developed new variant all had exactly the same genotype at the prion protein gene a genotype that's found in only 40% of the population so a clear absolute association with susceptibility for that genotype whether the rest of us are protected or just going to get disease later remains to be seen so given those four examples of major gene effects what tantalized us a few years ago was whether instead of carrying analyses in different diseases we could apply linkage analysis in families to pull out a few more big genes in complex infectious diseases we've taken the effect of Zipair analysis route and the first success we had really was in a rather unique disease leprosy leprosy is pretty horrible it's reducing in prevalence but not an incidence in the world two forms of interest to immunologists tuberculoid with stronger cellular immune responses more of a skin disease and leprosy was once actually considered a familial disease in the mid 19th century but then it was the first emilepre the first pathogenic to be seen down a microscope probably the first disease with the defined HLA association using family based association analysis in the early 1980s and now I'll show you the first disease where we think and this involved collecting literally hundreds of families in southern India with our collaborators a classical affected Zipair analysis on 245 Zipair's showed a peak of linkage lot score of over 4 on 10 p13 and luckily there was an obvious positional candidate gene in that region encoding the macrophage manus receptor MRC1 this is a C type lectin an entry receptor one of many and also has a key immunological role in micropanocytosis by dendritic cells leading to antigen presentation so Kerry Tosh in the lab defined lots of new slips and markers in that gene of over a hundred kb and intriguingly there was a cluster of polymorphism shown there in red which in fact are a cluster of three amino acid changes within 12 amino acids in CRD2 carbohydrate recognition domain too showed clear evidence of association within families in fact the strongest association is with the haplotype but most of the effect is due to SNP74 shown there a glycine to serine change and some elegant analysis by Long Carden in our center at Oxford showed that you could attribute essentially all of the linkage to these variants within this one gene and in fact some molecular modeling and some thermostability destabilizes this carbohydrate recognition domain interestingly from an evolutionary point of view this is a variant that is common in Caucasians if you like in Indians and in Europeans much less frequent in Africa the opposite of the geographic distribution of the malarial resistance gene sickle hemoglobin and perhaps relating to the fact that there's been far more mycobacterial disease over the centuries and hepatitis B virus infection a good illustration of a dichotomy in immune responses in humans with 15% of people failing to clear that virus those who become persistent carriers are at high risk of a paracelular carcinoma the commonest cancer in Africa and some evidence of a genetic association including intriguingly a high rate of persistence in those with trisomy 21 intriguing because that's where we found the linkage curve Q 21 Q 22 and again we were relatively lucky in that close to the peak of that linkage curve is a set of cytokine receptor type two genes including the two chains of the interferon alpha receptor interferon is a treatment for hepatitis B the second chain of the gamma receptor and the IL-10 receptor beta and again we were able to show within families this time that actually two of these genes can be used in infectious disease and can probably account for most of that linkage the IFNA R2 and the IL-10 RB gene and functional workers in progress so where we've got to is four diseases have been looked at these are chronic diseases most infections are acute so in the chronic diseases you can collect families you can do linkage studies in leprosy a gene has been identified in hepatitis B interestingly we see different genes like the Gambia the loci I talked about but Angela Fraudson's work in Italy suggesting that a quite different gene at a different location on the long arm of chromosome 6 that she's mapped and identified is the major susceptibility locus in that European population so clearly in our view genetic susceptibility to infectious diseases is polygenic mainly is accounted for by an accumulation of small effects from multiple polygenes but exceptionally you can find in linkage analysis and to finish I just want to reiterate the importance of these studies for evolution most people who work on evolution these days work with VDUs and computers looking at if you like the bodies in the evolutionary battlefield if you're looking at infectious disease and doing studies in the field where infection is still killing millions of people every year you have a chance to watch the battle and recognize the residue of the dead if you like so by looking at genes that are of key evolutionary importance I'll give you one more example you can get a flavor of what's going on these are tall receptor genes identified only in the last few years there are 10 of them scattered throughout the genome and these recognize PAMPs pathogen associated molecular patterns which are central to the recognition of pathogens and to the innate immune system tall receptor 4 and the pathway of endotoxin or lipopolysaccharide that's all been worked out very quickly and there is an amino acid change at position 299 of TLR4 which again in some interesting human challenge studies was shown to be a hypo functional allele when you give people inhaled endotoxin this variant therefore is hypo functional both in those clinical and transfection studies and other people's workers is hypo functional so why have we got a hypo functional variant in a key innate immune system receptor the answer again is probably malaria here in a study of about 4,500 children from three different countries in Africa Gambia, Kenya and Malawi we see modest but clear protection odds ratios of about 0.7 0.6 again severe manifestation of malaria in these African children so what we have is a hyper responsive allele of TLR4 that's interestingly found at higher frequencies in sub-Saharan Africa than elsewhere and malaria may have provided the counter selective force for this partially active allele in other words we may have a balanced polymorphism of the innate immune system with that variant being good for malaria but bad for gram negative bacterial infections so clearly infectious disease associations have been defined with many genes using a variety this is almost certainly just the tip of the iceberg in tuberculosis we know there's a major human susceptibility component these are four early twin studies each showing much higher concordance rates in identical to non identical twins two genome scans in western South Africa by linkage have been negative in TB as there have been in many other diseases so clearly the genes that must be there are not big enough other people do is to genome wide association studies I would argue that these are more likely to be productive in complex diseases that have selected our genome than in many other diseases simply because these are highly polygenic diseases you need to do this sort of study because the odds ratios are small you need to do big studies because almost certainly epistasis will be important that's been shown in HIV very recently by Mary Carrington's group genetic interaction so I'll leave you with the thought that there's a huge amount to do now we finally have the tools to do this with genome wide association studies in infectious disease and close by acknowledging the many people around the world who've been collaborators as well as actively involved in this research thank you very much so I'd now like to call all of our speakers for this panel up here and you can your microphone and please if you have a question in the audience there are two microphones in the aisles and I'd like to ask you to come up to the microphone and ask your questions we do have plenty of time for questions and so I want to see some people up there with your questions because I'm not asking all the questions for this panel wonderful I'd actually be curious to hear from either each of you or a subset of you you've covered broad areas of biomedical research each case illustrating quite nicely how things are about to change or just starting to change in the past couple years you've seen evidence and I would be curious to hear how since all of you have been so wildly successful doing things up until your laboratory staff in terms of their levels of expertise what they need to be doing say five years for now your programs and your respective areas what are the new tools and skill sets that are going to be needed in your laboratory that are different than what you've seen over the past say three to four years so did you want to direct your question to I mean I would actually encourage you to direct your question to oh but I like them all but I'll pick on Rick how's that okay good I think it's really boring when everybody feels obligated to answer the same question but you have to find the answer the the challenge I think is very clear the we will need to be doing much higher throughput genotyping if you imagine doing linkage disequilibrium studies will be need to be doing typical studies with millions of genotypes rather than tens of thousands of genotypes we need to be mining data we need to have much better informatics expertise is we tend to lose the idea that at the end of the day there needs to be a cortex at the end that can get around all of the data and actually interpret it and decide what it means I think there is the mistaken notion probably best exemplified in the biotechnology industry that if you simply generate a lot of data it will answer its own questions and I think that is something that clearly won't be easy to make sure that we don't build these notions that it's going to be easy if we just generate a lot of data Rick can I have one follow up because I suddenly and I guess I will point this back at you because I know it's something you're very interested in you've been so successful at getting clinically trained researchers involved in the problems that you're studying and certainly at NIH there's a tremendous amount of interest now in thinking and you raise it even yourself about you know we don't have enough clinical investigators how are we going to train what recommendations would you make for training clinical investigators who are already burdened with needing to learn clinical medicine and face all the challenges just being a clinician and being a clinical researcher but now on top of it having to gain additional skill sets to take advantage of this new genomic era. Yeah I think it's clearly among the multifactorial problems that we encounter this is one of them. The barriers I think are relatively low clinicians really come to genetics quite well the idea that you're starting with a top down strategy of a disease that they already know something about from their clinical experience is a very natural one for most clinicians and the barriers I think are creating environments where they feel comfortable and know that they actually have a career future if they follow a particular career path and work hard and are successful. Some of the barriers that we have now are clinical investigation looks enormously unattractive to many investigators given how hard it is to simply get IRB approval to start a study these days. So these I think are among the challenges that we face also loan forgiveness programs. Most clinicians finishing their training are an enormous personal debt and trying to figure out how to encourage physician scientists to take this road is a continuing challenge. The good news is this work is enormously appealing to any physician who has the faintest interest in investigation because there's nothing more satisfying than wrestling a disease that's been on the planet since the evolution of man to the ground and understanding it at a fundamental level. If I could just also echo some of what Rick said I think that there has always been something of a partnership between the scientists at NIH and the academic community and I think that's even more important now and in the future just as I illustrated our dependence on being able to go to a well curated database to then match our sequences to what was already known. I think that the efforts as outlined in the genome strategic plan for the future of trying to move as rapidly as possible to understand the complexities of the pathways within cells and the interaction of pathways is going to be critical because the more information you can have at your fingertips when you identify a gene or a sequence and go into the databases to see what's known about that and what is its functional role the more quickly you can generate some hypotheses about why that particular sequence might be affected in whatever disease it is that you find so I think that's important and I can't emphasize enough Rick's comments about clinical research and the increasing difficulty of trying to link samples from cancer patients with their clinical outcomes so that you can immediately see which genetic changes are associated with poor outcomes or aggressive disease and which ones are associated with a more response with a greater response to treatment and this is going to kill not just hamper but kill clinical research unless this is dealt with and since many of the regulations are coming from the government and from HHS it is absolutely critical that those who have some voice within that particular government program act to have people responsible for the generation of these rules understand the enormous detrimental impact that has on clinical research. This being the difficulty of getting IRB approvals for clinical research. That's right and the rules that are now out there is that if you do not have approval from a patient to use a sample for genetic studies you have to go back and get the approval and in leukemia these patients are by and large dead so you have to go find relatives to get the approval. It's catastrophic. Next question. David Valli Baltimore before I ask my question I might take my speak from my position as President of the American Society of Human Genetics the major research organization for human geneticists to second what Dr. Rowley just said. Then I have a question for Dr. Hill which I was very intrigued with his report of discovery of genes of both minor and major effect in terms of resistance to infectious disease and I wonder if he could just tell us a minute of his thoughts about how this will lead to drug development and especially to preventative approaches based on individualized risk. Thanks. As I hope I made clear that there are many reasons for doing infectious disease genetics and many reasons for doing genomics as we heard yesterday and only one of those is to try and develop new interventions but having said that that is one of the most important goals. I think the problem with most infectious diseases is that we need to move from understanding the concept of simple pathogen causing disease as being a one to one relationship to understanding the complex pathophysiology that determines why I say only 10% of people who are infected by mycobacterium tuberculosis end up with disease. So clearly there are a huge number of environmental and genetic factors along that pathway that determine whether you get a unnoticed subclinical infection or you end up dying from TB. And the tantalizing thing is that most of us can deal with most of these infections most of the time. Only 15% of people become carriers of hepatitis, only a fraction get TB, only a smaller fraction get leprosy. So we have the immune mechanisms and defense mechanisms to deal with these diseases. And it's only a small proportion that are failing. And the key question is how do we identify those people and identify what's different about those. So I think the chances of being able to modulate that subtle change in susceptibility by understanding the pathways in principle are very good. The problem is that most of the interventions we've tried like anti-TNF antibodies in severe malaria so far, and that's an intervention that came from the genetics of TNF polymorphisms, promoter polymorphisms and malaria, are applied too late. So I think the most important interventions will be preventive ones, particularly in developing countries in low cost settings. And clearly we're looking at immunomodulation, we're looking at vaccines, and the challenge now with our new tools and new vaccines that we can make is to target the right sort of immune response to prevent these particular diseases or particular manifestations or particular infections. So the short answer is an understanding of pathophysiology. And from that perspective, the more genes there are, the better off you are because the more pathways you can interrogate with millions or hundreds of polygens, then if there's only two or three genes, that's all you can learn about. Thank you. This is Bob Nussbaum from the NHGRI. Before I ask my question, I just wanted to echo Dr. Valley's comments as the President-elect of the American Society of Human Genetics, I'll be receiving the baton from his hand next year, the importance of IRB regulations and HIPAA regulations being interpreted in a way that maximizes protection at the same time that it does not interfere with the societal good that can come from this kind of research. I had actually a specific question for Drs. Cox and Hill, and that is if they wanted to comment on the importance of genome efforts in animals and animal models for disorders like infectious diseases and glucose homeostasis and how much that has helped or not helped in furthering our understanding of the human disorders. Yeah, very briefly. Mass genetics in particular has made important contributions in my own area of tuberculosis. One of the genes that has been shown to be relevant in human TB came out of mass genetics that was the NRAM gene worked on in Montreal for many years. That unfortunately is not the only example, but it's one of few, and the difficulties you have is that very often the pathogen has evolved as much as the human host. So the mouse malaria, for example, separated from the human malaria is around about the same time that mice separated from humans 50 to 100 million years ago. So there are substantial differences, and of course there are differences between different strains of mice and between other animals. So it's often hard to find the right model, particularly with the same pathogen, where you can use mouse models to identify key susceptibility genes. On the other hand, you could argue that what you're really after is pathways rather than the genes, and that will often be the case that new pathways can be identified by genome scanning in mice, and we're beginning to see more of that in malaria in TB and other infectious diseases. But if your pathogen doesn't infect mice, you have a problem. Yes, I echo that with respect to glucose homeostasis. So it's very reminiscent of the monogenic genes that have been identified in humans. Most of the genes identified in the mouse to date are not major risk factors in humans, but nevertheless have shed incredible light on new pathways and areas that we had very no knowledge of before they were identified in the mouse. So I think model organism studies will be key, not just in the mouse, but even in model organisms ranging from Drosophila to yeast, I think they'll be useful, but certainly and there are emerging models more polygenic and multifactorial for diabetes in the mouse, for example from Ron Kahn's group, that I think will be very useful in understanding how glucose homeostasis works and testing drug therapies for effectiveness. Jennifer Puck from the NHGRI, many of you, actually all of you have indicated how some of the very rare single gene disorders in humans have illuminated or started out studies of more complex and more common illnesses, but in the case of immunodeficiencies there are over 110 distinct rare single gene disorders known. How do you see these illuminating the more common diseases that we'd like to try to get a handle on? Yeah, just a quick comment on that. Of course those are by and large great candidates for diseases that are dependent on an immune response. Surprisingly, we have almost no examples of a disease where a knockout mutation causes a severe immunodeficiency in humans, and a subtle change in the expression of that gene is a polygene for a complex disease. The nearest is the interferon gamma receptor where knockout mutations of that or the IL-12RB2 gene give you susceptibility to usually non-pathogenic mycobacteria like mycobacterium avium, and there's some preliminary evidence that promoter variants of the interferon gamma receptor 1 are associated with TB susceptibility. But if you take those 100 genes as a standard, it's for common infections and it's interesting to speculate why in evolutionary terms. Certainly in terms of understanding basic biology, these are going to turn out to be enormously illuminating. That's probably not the soul fault of monogenic diseases that we haven't come up with new therapies since the entire field of immunology I think has been rather disappointing up to date in terms of coming up with new therapeutics. But just to cite one recent example, the polyglandular autoimmune disease with the mutation in the air gene. This has clearly I think been one of the spectacular new insights into the fundamental regulation of how we distinguish self from non-self and recent manuscript demonstrating this may be the transcription factor that leads to the expression of genes normally confined to specific organs in the cells that are determining self from non-self. These are I think going to be a fundamental importance. So I think the basic premise remains that understanding these rare diseases is illuminating basic biology, obviously which are going to turn out to have immediate therapeutic interventions I think is always a matter of speculation and largely of chance and you won't know until you get there. And we'll take one last question. I'm afraid it's not so much a question as a comment from Peter Goodfellow. Genetics works by actually you can only see what varies. So if in your population something isn't varying you can't see it by genetics which is another way of saying that pathway expansion is actually crucial. And similarly just because you identify a gene through that's the best point in that pathway for therapeutic intervention. Yes, I would certainly agree with that. And I think the example that I cited of the cholesterol biosynthetic pathway being inferred from mutations in the LDL receptor demonstrate that you do need to bootstrap your way from identifying a specific gene that's mutated to fleshing out the pathway. Now the fact that there are even alleles on the planet most of the genes that can be mutated are mutated but it's very fair to say that many of these will turn out to not be viable so we will never see them. And unless we have the ability to use these genomic approaches to go from understanding a particular gene to what pathway it's operating in we will likely be very unsuccessful in devising new treatments. Very good. Let's thank our panel again and then we have a break. Please be back by 10.30. While people are getting seated I'd like to tell you a little bit about my association with the genome project. I was acting director starting in 1992 when Jim Watson stepped down as director of what was then the National Center for Human Genome Research. And I must say it was a great privilege for me to interact with many of the incredibly talented architects of the genome project who you've been hearing yesterday and today and I was charged with two jobs when I took over which period of about a year. The first of my tasks was obviously to continue the momentum that had been started by Jim Watson and I just want to say that for me that was a relatively easy job because of the incredible support of the then deputy director Elkie Jordan who just retired last year and was deputy director of genome for all those years and many of the other people who were in the office when I arrived Jane Peterson who you've already met, Mark Geyer and many many of the other staff the staff of genome is absolutely fantastic and deserves a great deal of credit for the success of this project. My second job was to recruit somebody to replace Jim Watson not an easy person to replace obviously and it became clear after some searching that Francis Collins was in fact the person for the job based on his interest in human genetics, his intellect his personality and his leadership skills and also the fact that he could sing about DNA not only on key but with enormous enthusiasm and that was a unique quality that I think was the final reason why he was chosen to head the genome project. I also want to point out that I am aware perhaps more than anyone else that it was on April 4th, 1993 that Francis arrived to take the helm and become the steward of the human genome project because I was liberated. I thought at that point to go back to my lab as it turned out I was subsequently tapped for other responsibilities but Francis arrived and it was just 10 years and 11 days ago. So this is also in addition to all the 10th anniversaries that we've been celebrating. This is the 10th anniversary of Francis Collins leadership of the genome project and the modest person that he is, he never mentioned this but I thought I would bring it up and thank him both for relieving me of the job but also an incredibly a terrific job of leading a program which has been so very very successful. Now with those introductory comments some of you may have noticed that we were talking about genetics for parts of the body below the neck earlier in the morning and now we're moving above the neck to various sensory organs including the brain and I think we have some very very exciting talks to come. We have five talks. I have to ask the speakers to keep to the 15 minute time period or we will get into trouble later on in the day. The first speaker is Tony Monaco director of the Welcome Trust Center for Human Genetics. He received his M.D. and Ph.D. from Harvard Medical School and has been working on the basis of both neurological and psychiatric disorders. He will speak today on genetics of speech, language and reading disorders. Tony. For that nice introduction it's very nice to be here. I work at the Welcome Trust Center for Human Genetics and you heard from my colleague Adrian Hill we've been set up since 1994 trying to identify susceptibility genes for common disease and using the fruits of the human genome project every stage and now we look forward to using the variation that will come out of the human genome project in our goal to identify these susceptibility genes. Today I'd like to tell you about work in my own lab which aims to understand the development of the brain through molecular genetic strategies in behavioral and learning disorders in children including autism, specific language impairment, reading disorders like dyslexia and attention deficit hyperactivity disorder. But today I'm going to focus my attention on specific language impairment which affects about 5% of school children and it's really a diagnosis of exclusion where children who have difficulty acquiring expressive and receptive language despite normal nonverbal intelligence and opportunity and also in the absence of any sensory or neurologic other diagnosis. Specific language impairment or SLI is a large comorbidity with other childhood learning disorders including autism, dyslexia and ADHD and children with SLI have difficulties later in adult life with cognitive and social functions and also at increased risk of psychosis. So one of the genetic factors in SLI it's been known for a long time that speech and language problems cluster in families but of course this familial clustering can be to either genetic factors influencing the trait or the fact that individuals within a family share the same environmental factors. However twin studies pioneered by Dorothy Bishop and others have shown quite conclusively that speech and language disorders are very heritable with difficulties in monozygotic twins being much higher rate than dizygotic twins. Also if you treat language impairment as an extremely continuous trait one can get significant estimates of heritability which are over 95%. So how do you find genes for SLI? We've used two strategies one is a quantitative trait loci approach where one selects children who are at the extreme tail end of a measure of language ability and then assesses all the siblings in that family to see how the trait differences between siblings measures up to their genetic similarities across the genome. Another method we have used is to try to find rare multi-generation pedigrees that have severe speech and language problems and I'll tell you about one such family the KE family which is shown here about 50% of the individuals in the family over three generations are affected with a very severe speech and language disorder males and females equally affected it looks like a clear and simple auto-symbol dominant trait so although there's simple inheritance of the trait the phenotype is quite complex probably the most tell-tale sign of their phenotype is what's called verbal dyspraxia and that's a problem with the fine or a facial movement needed to articulate speech however it's not simply a motor problem they have deficits in multiple aspects of language abilities both receptive and expressive including the processing of phonemes and grammatical skills their nonverbal deficits are much reduced compared to the verbal deficits and they are some of the performance IQ affected members is less than you would consider normal however there's a large range of that and one would not consider the nonverbal deficits as characteristic of the phenotype in the affected members of this family and lastly brain imaging studies of the group in Institute of Child Health in London have shown both structural and functional abnormalities in affected members of this family in both cortical and sub-cortical motor related areas especially the basal ganglia and later when we'll see the expression pattern of the resultant gene Simon Fisher in my group in 1998 mapped this gene to a region of chromosome 7 Q31 right near the cystic fibrosis gene and using the first draft of the human genome sequence from that region facilitated with our collaboration with Eric Green we're about to identify a gene called Fox P2 however it was the stroke of identifying a second patient CS Jane Hurst a clinical geneticist in Oxford identified this patient who also has a verbal dyspraxia language impairment but relatively preserved nonverbal IQ and had a translocation, a de novo translocation break point which disrupted this gene of 17 exons right there in that intron that enabled us to sequence then this gene of 17 exons containing a fork head DNA in the K.E. family fork head transcription factors is a large expanding family of over 40 members in human you can see the large collection and relationship of all these fork head transcription factors throughout many species and there are known to be key members of regulators of embryogenesis and have been shown to be mutated in other human disorders including glaucoma, thyroid agenesis, ovarian failure and immune deficiency in particular though the mechanism of mutation is many times where dosage is critical in embryogenesis for the action of these genes and they usually have haploinsufficiency of one copy and are segregating as an autosomal dominant trait in the K.E. family here you can see a three-dimensional structure of the fork head transcription binding domain with the three alpha helices and the two wings and the K.E. mutation was an arginine to histidine missense mutation in an invariant residue that is conserved throughout all fork head transcription factors in all species and we did not find this mutation in any control chromosomes so what questions does the identification of this gene raise? First question is is the Fox P2 gene contributing more widely to common language related disorders such as SLI and autism? One might predict that this gene could have interesting evolutionary aspects if it's truly involved in the development of speech and language so what are the within and between species variation of Fox P2 in humans and primates where is it expressed in the developing brain? What genes as a transcription factor is it turning off or on at what time points and what happens if you disrupt Fox P2 in a mouse would you get the squeak one mouse? The first question we answered by looking at micro satellite markers a large number of them throughout this gene in families with autism or specific language impairment did not find any association we also screened all the 17 exons that we knew at that time in subsets of cases with autism or SLI and revealed no mutations we also looked at a number of SNPs not exhaustive but no significant association was found either in autism families or in SLI so our work and others I think are showing that this gene is not a major susceptibility gene for either of these two disorders and this is particularly relevant for autism because there's a very strong linkage in this same region of chromosome 7 that has been shown by a number of groups what's the expression of Fox P2 here is just a northern block not maybe what you'd expect for a gene involved in speech and language it's ubiquitously expressed but you can see here in fetal brain there's a fairly large transcript signal working with Andy Copp and the Institute of Child Health to see a lie in my group has looked at the Fox P2 expression in embryogenesis and it seems to be expressed in motor related areas here at embryonic stage 16 and a half in the mouse and we also looked at human fetal brain tissue and show that there's strong expression in the cerebellum a different phylamic nuclei the caudate putamen and also the substantia nigra and here in the inferior olivary nucleus in the medulla so it does parallel also what we see with the imaging of the human patients who have this disorder where they had motor related areas such as basal ganglia and cerebellum as being defective so what about the second the third question was what does this gene look like in evolution here just shows a more revised version of the structure of the gene there's a large number of non-coding exons and very large introns at the 5 prime end of the gene here you can see the forkhead transcription binding domain in exons 12, 13 and 14 there's a polyglutamine stretch in this gene which is actually one of the largest non-pathogenic stretcher polyglutamines ever seen in a protein but and here the KE mutation the arginine mutation in this exon the Sante Pabos group has analyzed this gene in between mouse and human there's only three amino acid changes and two of these amino acid changes have occurred since the split of chimps to human they're both in exon 7 function unknown however Sante's group has sequenced about 14 kilobases and so is another group and found similar findings that there's a lot of rare variants in these introns but very few common variants and using this information there's evidence that this gene has been highly selected in recent human history and also that this is compatible with the emergence of spoken language in the last 200,000 years so in summary FoxP2 is mutated with a rare development sort of speech and language and we think it's the insufficient functional FoxP2 copy during neurogenesis which leads to this abnormal development of the motor related neural structures that are important for speech for language as an analogy the muscurgistrophy gene is not the gene for walking we think that FoxP2 is not the gene for talking however it is for the first time an entry point into the neural pathways important for the development of speech and language if FoxP2 is not a major susceptibility gene for common forms of SLI how can we define genes for common forms of SLI if they affect several percent of school children here we've used that quantitative trait loci approach with large samples from both an epidemiologic sample from Cambridge and also a clinic sample from Guy's Hospital in London and we've used various measures of expressive and receptive language ability called the CELF the non-word repetition which is a highly heritable trait which is a measure of phonologic short-term memory various measures of reading ability and of course a full IQ measurement we've used quantitative trait loci linkage approaches as I showed you earlier that Sib pair trait difference approach but also partitioning trait variability using variance components on chromosome doing a genome wide scan in these two samples we have shown a very strong linkage a whopping LOD score here on chromosome 16Q and interestingly both the epidemiologic and the clinic based sample are contributing to this linkage which is the strongest for non-word repetition more interesting the same region of chromosome 16 shows quantitative trait linkage with various reading measures in the same families here measures of reading spelling or reading comprehension are also showing up in the same region however the expressive language and receptive language test show any strong linkage in this region however there are problems with multiple phenotypes and univariate linkage analysis of course if you're doing 8 to 10 different measures and 300 different markers across the genome there are large problems of multiple testing there's issues of power if you analyze linkage results based on independent univariate measures are you getting all the power out of your sample there's also problems with the interpretation of linkage findings so the replication of linkage findings if you find linkage to non-word repetition in this region but not expressive language does that mean you found the gene for non-word repetition and we think this is a very dangerous area where people try to look at regions where they found linkage and correlate them with only that univariate measure to overcome this problem we've worked with Angela Marlowe and Lon Kardon at the Welcome Trust Center to do a full multivariate genetic analysis here these traits are multiple correlated traits one can actually simultaneously look at the co-variation of these traits at the genetic locus that you're interested in and here you can see for that chromosome 16 locus here are all the univariate traits and then a full multivariate analysis and you can see that one can actually get a significant LOD score which is not only picking up the large LOD score we saw with non-word repetition but also many of the reading scores and also showing some co-variation we're now with expressive language and receptive language that we did not see when we looked at those as univariate traits one can also in this model drop individually each of the univariate traits and ask whether they're contributing significantly and at least at the 5% level all six measures are contributing to the linkage in this region so I'll end here with my conclusions that FoxP2 has been identified as a gene mutated in a rare monogenic speech and language disorder genetic analysis of families with common language and reading disorders has identified significant QTL linkages and we're now using the variation in the genome to look for linkage disequilibrium to identify candidate genes and associated variants the goal of this research is a knowledge of the biological pathways in the brain that are important and hopefully this will provide better diagnosis and potential results. And just quickly some acknowledges Cecilia Lai, Simon Fisher who did all the work on FoxP2 Diane Newberry, James Cleak on the SLI Genome Scan our collaborators on the statistical analysis Angela Marlowe, Lon Carden Andy Kopp's group for the study of FoxP2 Farinay Varga, Caddium and Marcus Pembrey who have been our long standing collaborators on the KE family Slanti Pabbo's group on the evolution of the genetic gene Eric Green for his help in facilitating the analysis of the chromosome 7 region and my clinical collaborators not only in the autism consortium but Patrick Bolton and Gilliam Barrett especially in the SLI consortium. Thank you there and I'll stop for it. Our next speaker is Tom Friedman chief of the laboratory molecular genetics at the National Institute of Deafness and other communication disorders. This is Tom Friedman, hereditary disorders of hearing while on faculty at Michigan State. He was recruited several years ago to the NIH and we'll talk about hereditary hearing disorders in the genomics era. Thank you very much. It's a privilege to be here and I thank the organizers for inviting me. In 1994 I was just 10 years old living in Detroit and I read an article in Scientific American reporting the discovery and although I didn't know it at the time this probably marks the beginning of my interest and an entire generations interest in the nascent field of molecular genetics. Today I work on the genetics of hearing and deafness in humans. About one in a thousand children are born with a significant hearing impairment and it's thought that about half of that has a hereditary underpinning and the other half an environmental. About 50% of older individuals have a significant hearing impairment and we really don't know what the contribution of genetics is to that. It's significantly understudied. Most inherited forms of deafness usually segregate as monogenic traits and hearing loss due to oligogenic traits or complex modes of adherence are rare and difficult to document. There are literally hundreds of syndromes that include hearing loss as one of their features and this isn't surprising since some of the steps involved in hearing are shared by other systems. In contrast there are many novel processes and structures involved in hearing and therefore it's not unexpected that there are many genes that cause isolated hearing loss as the sole clinical sign and this isolated deafness is referred to as non-syndromic deafness and I'm going to present an example of each one example of non-syndromic deafness and an example of syndromic deafness. There are predicted to be a hundred different genes which when mutated can cause non-syndromic hearing loss and in a few moments I'm going to discuss an example. It's counterintuitive but non-syndromic deafness is not a simple diagnosis to make and should be considered tentative even after the gene has been identified. Hearing loss is often detected at the stations of other organ system pathology and young children with one of these syndromes may be incompletely diagnosed with non-syndromic hearing loss. For instance, individuals who have recessively inherited Gervle and Lang Nielsen syndrome have congenital auditory dysfunction together with a heart conduction problem that is easily overlooked. I'll present a second example in a few minutes. Therefore the role of clinicians in mapping and studying the function of deafness genes is fundamental and crucial throughout these studies. In competition and in collaboration, several research together are engaged in a saturating search for genes that are necessary in humans and mice for hearing. The continuing discovery of a large number of common and rare mutations associated with deafness in humans has provided many serendipitous amounts of entry into the biology of hearing. This high degree of genetic heterogeneity of deafness reflects the great diversity of specialized proteins that are required to make sense of sound. I now want to briefly tell you how the ear makes sense of sound. Sound waves enter the external ear and the an ear drum, vibrates and the ossicles of the middle ear vibrate. One ossicle is connected to the ear drum and movement of the stapes which is attached to the oval window causes movement of the fluid that fills the cochlea or this snail shaped organ. The cochlear duct has three chambers. This image shows a cross section and this middle chamber is the organ of cordy which I'll tell you more about in a moment. The red ball represents the direction that sound waves is traveling back in the oval window in the scale of vestibule. And the blue ball is the direction the sound wave is traveling back in the scale of tympani. Fluid wave that is created in the cochlea causes an up and down movement of the basilar membrane of the organ of cordy and this movement of the basilar membrane produces a shearing force between the tips of the stereocilia these projections from the outer and inner hair cell. In an animated fashion at the stereocilia a few nanometers of deflection of these organelles the stereocilia bundle towards the tallest stereocilia and through an angle that can be less than one degree opens within a few microseconds a channel which is thought to be located either at both ends of these tip links or one or the other perhaps at the top. This mechanically gated ion channel has not been identified the molecular correlate is unknown as are the gene or genes that encode these tip links. When these channels open potassium in the endolymph flows down an electrochemical gradient into the hair cell and through a complex process which is obviously oversimplified here depolarizes the membrane. The acoustic nerve sends a neural signal to the brain that the brain interprets as sound. It is the inner hair cell that transduces hydromechanical signal to an electrochemical signal and it is the outer hair cells that further affect this signal by amplifying it in a complex way which I won't discuss. Sensor neural hearing loss occurs when the inner ear fails to send a nerve signal to the brain in response to a sound stimulus in the cochlea. The two genes that I want to discuss are involved in this process. One is involved its thought in the recycling of potassium back to the endolymph and the other one is involved in the integrity maintaining the integrity and probably important during the development of the stereocilia. This rapid progress which I'll tell you about briefly in understanding hereditary deafness has paralleled the ability of genetic maps with highly polymorphic markers the physical maps, the genomic sequence the instrumentation and bioinformatics and the transcript-owned databases as well as very important models because once you identify a gene for hereditary deafness in humans you need an animal model. The first thing that we do is to ascertain families segregating hearing loss and we ascertain them in North America and you're looking for large families in this case one that's segregating dominant disorder and in this case probably a recessive disorder. These are children of a consanguinous marriage and it's important then to obtain comprehensive clinical data and determine the mode of inheritance. In 1996 some genes for hereditary deafness had been mapped but not a single one had for non-syndromic recessive deafness but not a single one had been cloned at that point. But the next year in parallel with the ramping up of these resources, these wonderful resources, three genes had been identified or cloned in 1998 10 genes, a total of 10 in 99-15 in 2016 in 2001-23 in 2002-32 in 2003 and the orange tags represent deafness loci for which the gene has not yet been identified. In addition, on chromosome one there's a dominant modifier which acts as a suppressor of deafness for a locus on chromosome 4 DFNB-26. So you ascertain families you map the genes and you find mutations and often these mutations are private mutations or rare mutations but nevertheless they give you an entry point into studying the biology all the way from the histopathology to the structure of these proteins and their function and that leads to an integrated understanding of the biology and the pathophysiology of the auditory system and eventually to inform clinical applications one of which of course is a cochlear implant. So when common mutations are identified in these genes there are immediate benefits. I want to now tell you just about two loci one on chromosome 13 and one on chromosome 10. So in 1994 in Dr. Petit's lab at the Pasteur the first recessive non-syndromic profound congenital deafness locus was mapped to chromosome 13 and this locus was designated DFNB-1. Three years later in Kelsel's lab the gene underlying this was identified and it encodes a gap junction connexin-26. There are about 20 members of the connexin family which are phylogenetically ancient proteins found in all metazoans. Gap junctions are involved in intercellular exchange of small molecules and ions between adjacent cells as well as forming channels on the surface of cells. Also this gene has a very simple architecture there's just one coding X on. Nevertheless, mutations in this gene account for about 30 to 60% of non-syndromic autosomal recessive hearing loss in Europe and the United States in one particular mutant allele 35 del G, a deletion of a G base, accounts for the majority of mutant alleles of JGV2 in southern Europeans. Rob Morel showed that another mutant allele of this gene 167 del T is responsible for about 40 to 60% of non-syndromic deafness in the Ashkenazi Jewish population and there's a carrier frequency of about 4%. I now want to switch to the second gene that I want to tell you about and give you a little bit of introductory information about Usher's syndrome. Usher's syndrome is characterized by bilateral congenital sensorineural deafness and it's often profound at birth for the most severe form of Usher's syndrome there are two other types which I'm not going to discuss. While the loss of vision occurs perhaps at age 15 to 18 to 20 so children who are born with Usher's syndrome are thought oftentimes to have non-syndromic deafness. They also have vestibular problems which is characterized by lane ambulation. There are seven loci, there are more than seven loci but seven published loci and five of these genes have been identified. Usher 1b, 1c, 1d, 1f, and 1g one of them encodes an unconventional myosin two of them encode proteins that are involved in macromolecular complexes and two of them encode adhesion proteins catherin-23 and proto catherin-15. I just want to briefly tell you a little bit about catherin-23. In 1996 families segregating non-syndromic deafness were mapped to chromosome 10 and shortly thereafter a couple of families with Usher's syndrome were also mapped to this very same interval and it was hypothesized by this group that mutations of the same gene could be responsible for both non-syndromic and syndromic form of deafness. Julie Bork now Julie Schultz positionally cloned a gene in which he found mutations in families that were segregating non-syndromic deafness and also mutations that were responsible for Usher 1d. There's also a mouse model originally mapped by Breida and positionally cloned by Federica de Palma at the same time that the human gene was discovered. There are now 20 mutant alleles which are associated with non-syndromic deafness in this gene catherin-23 which encodes catherin-23 and there are about 21 mutant alleles which are responsible for Usher's syndrome. So what's the difference? Turns out that the mutations that are responsible or associated with non-syndromic deafness appear to be hypermorphs or amino acid substitutions which we believe do not entirely disable this protein. While truncating mutations of catherin-23 are associated with Usher's syndrome so what are the emerging possibilities? Well there are many of them but I've just listed three here. One is improved diagnostic accuracy and genetic counseling which will make it possible which will be possible when we understand more about this genotype-phenotype relationship for each of the genes that have been identified which means ascertaining their families and having many different mutant alleles. The molecular correlates of interior structures and novel processes will continue to be identified by map-based and functional identification of genes for hereditary hearing impairment in humans, mice and zebrafish and as I mentioned many of those specialized functions in the auditory system do not yet have molecular correlates. And finally an increased understanding of the function of deafness genes should facilitate better healthcare for hearing impaired individuals. So I want to thank my colleagues in the section on human genetics for their clever and hard work and I particularly want to thank Dr. Andrew Griffith who has been the clinician scientist because it is his work which represents the underpinnings of the molecular genetics that we do and our collaborator in Pakistan as well. So thank you very much. Our next speaker is Val Sheffield, Professor of Pediatrics and Director of Medical Genetics at the University of Iowa College of Medicine. Dr. Sheffield received his MD and PhD from the University of Chicago and his talk is entitled Impact of the Human Genome Project on Hereditary Blindness, Simple and Complex. So some genetic conditions are lethal whereas other conditions lead to loss of quality of life. Hereditary blindness certainly falls into the latter category. The Pulitzer Prize winning sports writer Jim Murray who I used to read his articles as a youth who later went blind in life made the express the opinion that without vision there is no quality of life. He of course was speaking of his physical loss of the sensation of vision however it would also apply to the global vision and certainly we've seen the vision of the Human Genome Project and the impact it will have on quality of life. Hereditary blindness like other types of genetic disorders can be roughly categorized into categories, single gene or monogenic Mendelian disorders of which there are many and then common complex disorders which are fewer but they are common so there's many patients. I'm going to limit my talk to age-related macular degeneration or AMD. The macula is a part of the retina in which you have the highest concentration of photoreceptor cells and hence it accounts for your central vision your most acute area of vision where you use for reading and driving and other functions. In AMD you get a degeneration of the macula. This is the most normal macula here that's shown although these lesions you don't have to be an ophthalmologist to pick out that there's abnormal lesions here that result in the photoreceptor cell death. The disorder is both clinically heterogeneous as well as genetically heterogeneous. I think you can recognize that these retina look different from one another but they would all be categorized as AMD. There's many studies that have shown that AMD increases with age hence the component of the term age related. As the population in our society increases it's becoming a bigger problem. In the year 2000 there were 16.6 million people with the disorder but it is projected that in 2025 there will be 27 million people with this disorder. Age related however does not mean aged caused. Certainly we do not want to a third of us do not want to look forward to losing our vision later in life if we escape the cardiovascular disorders and their lethal effects. And so it should not be accepted that this will become the fate of one third of the aging population. AMD in fact does have a significant genetic component as has been demonstrated by multiple different studies. So we and others have come up with hypotheses being that perhaps we studied the Mendelian monogenic forms of macular generation. They may be an allelic variant of AMD hence we could possibly get clues to which genes are involved in the late onset disorder or at least identify some of the pathologic mechanisms of AMD by studying these monogenic forms. The first goal of the human genome project is actually the high density development of a high density genetic map. This made possible the mapping of many disease loci and certainly now that we have the complete human sequence not only are we able to put phenotypes on the map using the genetic markers but we are able to now correlate that with the one base resolution physical map and identify genes that cause various macular and retinal disorders. This has been done in many laboratories across the world and as you can see here there's been a rapid increase in the number of loci that have been mapped as well as the number of genes that have been identified. This in our own lab has occurred with a worldwide collaboration. One of my favorite to point out is this collaboration we've had in Israel with both the Israeli and the Arab population. This shows a truck here says this is a University of Iowa logo here. This is a very talented Bedouin physician that was trained in Israeli hospitals and it shows the potential for breaking down political barriers by science and hopefully science will continue to be a leader in the breakdown of political barriers through these collaborations. Now we are at a point where we can test this hypothesis are some of these Mendelian disorders allelic to the common form of the disease and we selected candidate genes based on the fact that we had either mapped or cloned some of these genes and they appeared to resemble the phenotype of the age-related macular degeneration. The bottom line is that after surveying hundreds of AMD families or AMD patients for mutations in these various genes there was really no evidence that they played any role in age-related macular degeneration. This would appear to be a failure of the strategy but I would point out that there are actually many successes in that they have pointed out mechanisms they've pointed out really the genetic complexity of the disorders we've learned about basic biology and some clinical applications and I'll show an example of each of these. At the level of mechanisms one point was controversial as to which cell types in the retina are actually involved in the disorder whether it was the photoreceptor cells or the underlying retinal pigmented epithelium or numerous other cells including the vasculature of the retina and indeed there's now examples of genes expressed solely in the photoreceptor as well as genes expressed in the RPE that are involved in this disorder. Unlike the nice hypertension work that Rick Lifton showed where you could easily point to a unifying pathway such as not the case with AMD and the mechanisms as genes of different function are involved in the disorder. It's also given us clues to genetic complexity. One of the examples I like to show is illustrated in this slide where we have a single gene ELOV-L4 which is involved in fatty acid metabolism. There's a five base pair deletion mutation of this gene which multiple patients have and here we have a 62 year old patient with nearly normal retina certainly normal vision who has this deletion and here we have an 11 year old patient that has severely impaired vision with the same mutation and the abnormal retina indicating there must be modifiers to this primary gene defect be they genetic which is my guess or environmental defects but finding either the environmental or the genetic modifiers will be is a great goal and should have an impact on the understanding of this disease. The human genome project has also allowed us to come up with different strategies or paradigms for finding disease genes and here's an example of something that we're doing in our lab is that we're taking sort of a microarray approach of human DNA samples where what we can do is really take multiple DNA samples collected from individuals that have been very carefully clinically phenotyped and then instead of choosing individuals that have the same phenotype to fill up the microtider plates we select individuals that have very different phenotypes and then we can screen candidate genes in these DNA samples, identify potential mutations, go back and look at the phenotype, refine the hypothesis to select this specific phenotype to see if a given gene does play a role in the disorder. We selected one gene, the NR2e3 gene or a PNR gene which is a nuclear receptor gene on chromosome 15 and when we did the initial survey of 400 diverse patients with this disorder we identified two patients with putative mutations we could then refine our hypothesis and go back and look at mutations in individuals with the same phenotype where we found in 25 of the 29 patients with that particular phenotype they had mutations. The phenotype was particularly interesting as shown by this cartoon of the macula this shows an artist rendition of the multiple types of different photoreceptors that you see in the macula where there's red, green and the most rare type of photoreceptor, the blue photoreceptor in a normal macula however in this particular disease which is called enhanced S cone syndrome the macula is packed solely with blue cones the exclusion of the red and the green cones and this can be shown electrophysiologically here is a normal response to blue light which is this small electrophysiological response whereas these patients with enhanced S cone syndrome have this massive response this can be confirmed histologically on one disease patient that they do in fact have cones packed with blue photoreceptors this suggests then that the progenitor cells that gives rise to cones have a default pathway to the S cone type of photoreceptor so what about therapy many think that the goal is gene therapy but the goal is really the proper application of the proper therapy to the correct individual patient I like to give here is that in studying the Bedouin population we initially studied three different kindreds that we initially thought had the same genetic condition as we identified the loci and the genes it became clear that they all had different molecular causes of this disorder we could then go back and ask the question well do they really differ phenotypically and it turned out that two of the kindreds had severe retinal degeneration but the other kindred had mild retinal degeneration but they had myopia on top of that and so they were being treated as if they could not be as if their vision could not be corrected with eyeglasses but indeed they could be corrected by eyeglasses and it made a big impact on these children just having the lion's society donate eyeglasses another example of this is a disorder Sorsby's macular dystrophy which is caused by mutations in the TMP3 gene now that you can molecularly determine molecularly diagnose individuals with this disorder by finding mutations in the TMP3 gene then you can ask the question what do their phenotype what does our phenotype look like one of our collaborators Sam Jacobson noticed that these patients look like they had a phenotype similar to people with vitamin A deficiency a drastic vitamin A deficiency so we treated these people with vitamin A and indeed their sensitivity to light improved so here it shows a baseline when we give them over this period of time high dose vitamin A their light sensitivity increases five orders of magnitude you take away the vitamin A and it gradually falls off you give it back it improves and here you write the paper not that gene therapy won't be possible in these disorders that it may be there's one nice article by Gene Bennett's group showing gene therapy in the dog model of a severe childhood form of blindness in summary AMD has a significant genetic component disease gene discovery has led to macular disease mechanisms but not yet to a major AMD gene and there's real value in identifying clinical subtypes and identification of genetic modifiers and I think novel treatments will be a focus of future work thank you our next speaker is John Hardy chief of the laboratory neuro genetics at the national institute on aging at the NIH he received his PhD from imperial college in London and he's well known for studying the genetic basis of dementia his talk is entitled genetic analysis reveals unexpected connections between neurodegenerative diseases thanks I remember a couple of things from 1964 to I remember performing she loves you help and twist and shout for my parents with my younger brother and I also remember reading children's encyclopedia of knowledge which was a weekly magazine and trying to understand the basis of the code on usage in DNA and it's really nice for me it's all to be able to give a talk at this 50th anniversary almost as nice as if I had actually managed to get onto a Beatles reunion tour so that's how thrilled I am so why my intention is to my intention is to convey a little bit of the excitement in neurology and your neurological neurodegenerative disease research because we're really in the third golden era in my view of the study of neurodegenerative disease the first golden era really came about also because of technological improvements of course with the description of the diseases the descriptions of Lewy bodies tangles, plaques and the description of these diseases as clinical pathological entities the second golden era really I think the high point of the second golden era really came with the transmitter based treatments for these diseases Eldopa therapy I think is the perhaps the best the most really dramatic therapy understanding the transmitter basis of the phenotypes in the disease I think that was a wonderful era and now the genetics based understanding of diseases and understanding of the etiology really offers the hope that we'll have knowledge based therapies close of course the key in my view the key findings were the Huntingdon's linkage I remember that really influencing my career decisions the cloning of the prion gene by a couple of groups and the amyloid gene by Bayreuter's group of course the triplet repeat diseases I remember being told that human genetics wasn't worth studying because you could do everything much better in mice because you could organise the breeding but of course we would never have found triplet repeats diseases so I think this is a great time this is a great time to be doing research and it's a great time because of the genome project in its broader sense now the usefulness of genetics you know pity the poor pathologist all the pathologist gets to see and this is how diseases were looked at is the credits of a movie all he gets to see is that Schwartz and Egger starred in the movie but he doesn't know for sure and the nice thing about genetics is it tells you exactly the opposite it tells you where the movie starts so at least now you've got to see where the movie starts when you find the mutations you know where it ends and then you trust the molecular biologist to fill in the gaps it's a great time now these are the diseases I'm going to try and skip through very quickly Alzheimer's disease, 4 million Americans although this is a slightly prejudiced thing for me to say as an English guy to say 4 million Americans with Alzheimer's disease but of course it's all over the world the pathology there is a plaques made up of a beta I'll show you some of those in a minute in mice tangles made up of a protein called tau and I'll also say that there's often Lewy bodies I'll try and get back to that at my last slide if I have time because the pathology is a little more complex pre on diseases I'll talk very briefly about perhaps in my last slide you know pre on diseases sometimes have tangles and very interestingly some of the hereditary pre on diseases which have tangles also have Lewy bodies I'll try and get back to that I'll also talk about PICS disease you know I like these old German type names for diseases now we have committees to name diseases front of temple dementia with Parkinsonism link to chromosome 17 now there is a committee decision if you ever heard one I'll also talk about if I'll try and talk about Parkinson's disease and Lewy body dementia I want to make the point I think the Parkinson's disease in my view is not the etiologic entity clearly it's the treatment entity in other words Eldopa treatment but you know there are families and we've studied a family extensively in which some individuals have Parkinson's disease and other individuals have Lewy bodies in other areas particularly the cortex and have a dementing syndrome and I would argue that that's the same disease from an etiologic perspective of course now we treat with Eldopa the treatments are different obviously the dementia doesn't respond to Eldopa but when we have treatments based upon genetic knowledge then we'll treat these two currently current different entities from a treatment perspective in the same way I'll also try and talk about progressive super nuclear palsy and corticobasal degeneration these are sporadic Parkinsonian like diseases in which the pathology tangles now this is where we start I guess with the identification first of all of mutations in the amyloid precursor protein gene we don't know the function of this gene Conrad Bayreuter first suggested when he cloned it that it was probably a self surface receptor and now I think there's data this last year saying he's almost certainly right we've now identified a whole series of mutations in the amyloid precursor protein gene this is the whole protein this is this part of the molecule blown up I'll mention this mutation first this mutation was found by Blasphrangianis group in a disorder different from Alzheimer's disease but in which amyloid deposition occurred in brain blood vessels and really that was a very important finding and led us to start to look for amyloid one of the reasons we started to look for amyloid mutations in Alzheimer's disease and now these there's a quite a number of mutations in Alzheimer's disease being found this is the cluster of mutations at the C terminal end of the amyloid fragment here which all lead to Alzheimer's disease these two lead to Alzheimer's disease and a couple of kindreds one of which is in Belgium and then this double mutation here occurs in a large Swedish kindred with Alzheimer's disease so these mutations all cause Alzheimer's disease typically with an age of onset in the 50s and you can see that there's clear clustering and this clustering happens to be at the points where the amyloid precursor protein is metabolised here's the amyloid precursor protein again and these are the two normal pathways of metabolism the alpha pathway and the beta pathway here these pathways both occur in all of ourselves in all of us all the enzymes responsible for these pathways have been cloned in the last four or five years I won't have time to talk about those enzymes particularly but in the cartoon language we use this is the good pathway because it doesn't lead to amyloid production and this is the bad pathway the first mutation to be worked out was that Swedish double mutation and a very elegant series of experiments Martin Citrin who was then in Dennis Selco's lap showed that the Swedish mutation was an excellent substrate a better substrate for the beta pathway and it's almost certain then the reason this mutation leads to disease is because for that very simple reason it leads to more a beta production he did exhaustive mutagenesis here and showed that only the Swedish mutation had this effect and so the reason that only one mutation has been found at that position is presumably because that's the only mutation which behaves in this way the mutation we first described here at the C terminal doesn't alter the flux through the pathways what it does do and this was particularly worked from Steve Junkins group what it does do is alter the proportion of a beta 40 and a beta 42 I've been slightly over simplifying the metabolism all of us in this room I hope will be producing 98% a beta 40 and about 2% a beta 42 the individuals with this mutation instead of producing 98 to 2 produce about 95 to 5 so they slightly increase the amount of a beta 42 and this is a less soluble peptide and therefore more prone to deposition finally those mutations at least some of them in the middle of the a beta molecule seem to inhibit the alpha pathway and therefore push metabolism down a beta like pathway so the three cluster the reason I've rather overemphasized this work even though it's four or five years old six or seven years old actually is because it shows that the three clusters of mutations have different molecular effects and have the same overall outcome which is that a beta deposition is a more likely event and this is the basis for the what I think is the current major theory of Alzheimer's disease which is called the amyloid hypothesis and it's illustrated diagrammatically and oversimplistically here saying that all the known genetic all the known causes at that time led to this process becoming more likely you know with the APP mutations were in fact incredibly rare we had 30 families with early onset autostomal dominant Alzheimer's disease and only found three mutations in those three of the 30 families and so there are clearly other loci and the other locus was linked by Jerry Schellenberger though entertainingly many years before nine years before using protein data this group had suggested linkage to exactly the right place in fact though they made the mistake of reporting linkage to two chromosomes in one family and so it wasn't given as much notice as maybe it should have been and this is the the documentary evidence for the race to clone the gene here the original reports here gradually narrowing down the region as the genome project put in more and more markers down to 6.4 centimorgans here and then Peter Hislop cloned the gene in a tour de force really and identified families with six mutations this is the gene he cloned Precinal in one and a very large number of mutations now have been described about 100 mutations nearly all of them miss sense and they all have no not all but most of them have an onset age of about 30 to 40 years we looked in the databases protein on chromosome one which also turned out to have mutations half a dozen mutations have been described and these have very variable onsets all of these mutations well this leads to the function of Precinal in what is the function of Precinal in it almost certainly now is the almost certainly is gamma secretase which is responsible for this cleavage of the APP molecule so in fact it's a you know an enzyme substrate relationship and mutations in either the substrate or the enzyme seem to lead to Alzheimer's disease and this is allowed us to make animal models at least of amyloid deposition here with which show these and transgenic mice with APP genes showing plaques and APP and Precinal in genes showing more plaques however these mice didn't develop the other pathology the tangles and the cell loss but still they've been very useful and they used extensively now for drug development we started to work on this project thinking it was entirely separate from Alzheimer's disease I have to say we collected a large family in Australia in Crocodile D country which showed linkage to an area that Wilhamsen and Lynch had previously shown linkage on chromosome 17 this was a frontal temple dementia and this is our family others had also localized disease genes here and the tau gene was here right in the middle of the molecule an extraordinarily important paper Maria Spilentini showed that all of these families had very some are very subtle but all of them had tau pathology I don't have time to deal with that unfortunately this is an extraordinarily important paper very elegant work showing that they all have pathology this is the tau gene here showing alternate splicing of the tau gene I just want you to look at alternate splicing of chromosome 10 here all of us will be making about equal amounts of both X on 10 in and X on 10 out protein and this turns out to be very important we all miss tau mutations on the first time round but in fact we miss them because the tau many of the families have mutations in just outside the X and I wanted to show this because I just think this is a beautiful bit of where really understanding the base pairing really made the difference this is the end of X on 10 in the heteronuclear RNA and what we observed here was this is a palindromic sequence and in many of the families the palindrome was destroyed by these intronic mutations and therefore altered the splicing mechanism so individuals with these palindromic destroying mutations make only 4 repeat tau from their mutant allele and this is what causes them to get disease so these are non-coding changes they lead to instead of these individuals instead of producing 50 percent 3 repeat and 50 percent 4 repeat tau they produce 25 75 of course they've got a normal allele as well and that really that's why they get sick there are however coding mutations as well which have been now described by many groups including ours I'll just mention one P301 allele this is the most common mutation and I'll come back to that in a second while we were doing the sequencing we noticed that in the Caucasian population there are two tau genes which don't differ in a amino acid sequence they differ in intron size wobble bases and so on and so forth and cytos group had shown that an association between marker in the tau gene and PSP and we were able to show that this was actually an indicator of a more extended haplotype association the tau gene there are two versions of the tau gene in the Caucasian population as I say differing only in wobble bases and intron sizes 70 percent of chromosomes carry an H1 locus that means 50 percent of Caucasians about our H1 homozygous however when you look at people with progressive epilepsy and corticobasal degeneration 95 percent of them are tau homozygous H1 homozygous and so tau is a risk factor allele for the sporadic these sporadic tau opethes these sporadic tangle diseases coming back to the P301 L mutation we took this mutation and put it into mice and we're able to make mice with tangles this was J. De Lewis's work in Mike Hutton's lab these mice were driven by the preamp promoter they had tangles only basically in the midbrain and when we crossed those mice to mice with an APP transgene we were able to move the pathology up to the cortex so here we finally demonstrate a link between the two pathologies and this allows us to really at last have a real working model of the experimental evidence for the amyloid hypothesis this is a diagram of that showing that in Alzheimer's disease you have the whole process getting on the Mississippi River so to say up in Minnesota entering down here in New Orleans with cell death and dementia sounds more entertaining than being up in Minnesota I have to say however you can also get on the Mississippi River in St. Louis with tau mutations and still end up in New Orleans with cell death and dementia I'm being told to stop but I just want to make clear in the last few seconds that there is entirely the same types of observations have been made with synuclein and Lewy body diseases the key finding here was made by my colleague here Bob Nussbaum who found synuclein mutations in Parkinson's disease that led Michelle Goethe to Marius Bellantini to show that synuclein was the protein involved in Lewy bodies and there is an entirely parallel literature between synuclein and tau every finding I have said about tau you can almost say the same thing about synuclein so for example synuclein haplotypes are associated with sporadic Parkinson's disease in the same way as tau haplotypes are associated with sporadic tangle diseases so there is an entirely parallel literature between the synuclein diseases and the tau diseases and this is summarized here and how are they all related I feel as if I'm on a kind of a sprint here here is Alzheimer's disease I've already said this is the amyloid hypothesis here for Alzheimer's disease remember I said that you could have Lewy bodies in Alzheimer's disease as well so clearly the amyloid or likely the amyloid pathology can drive a Lewy body formation as well as well as it can drive tangle formation I mentioned that pre on diseases which have tangles also have Lewy bodies and so you can map these onto this pathway and of course you can map the hereditary versions as I did before for the for the tangle diseases here and likewise you can map the Lewy body disorders here I should point out that especially for the Lewy body diseases there is at least 5 extent loci so finally then we have now I think genetics together with an understanding of the pathology of the disease has led us to a point where we can see that are those these diseases have many converging etiologies the number of pathways to cell death is very very very few tau pathway, tau tangle pathway and a sonuclein Lewy body pathway and now really is the era in which we hope that this knowledge will start to lead to therapies. Thanks very much The final speaker in this morning session is Aravinda Shakravarti professor and director of the McCusick Nathan's Institute of Genetic Medicine at Johns Hopkins School of Medicine he has been a major contributor to the study of complex genetic diseases and will speak today on genomic views of psychiatric illness Thank you Mike believe me I'm sort of mindful of the time the fact that you're hungry I'm just as hungry as the rest of you so I try to be on time and to try and tell you some of the features that perhaps are a little different than the previous speakers have outlined I can't really proceed without thanking the organizers and to really tell you what a great thing it is to participate in the celebration of the years as a birthday of DNA but rather when the human genome project or ideas began as you heard yesterday the number of people who really did support and went forward were really few and it's sort of fantastic to sort of see where we've come in what really appears to be a very short period of time it also reminds me that in looking at complex genetic disorders that if you think in completing the human genetic human project that we've stood on very many tall shoulders that we need to do even with much greater frequency in the years ahead so I'm going to try and talk largely about neuropsychiatric disorders and how we can go forward when we know in fact that the large majority of cases of these disorders and this is taken from a recent review of the phenotypes and the prevalences from at least the best prevalences and incidences that we can find but I think the significant feature is that almost all studies show that when you do a variety of family studies that the heritabilities of these phenotypes in fact is really considerable and in many cases very large despite the commonly assumed sort of difficulties in assessing psychiatric phenotypes this is in fact a large burden of disease in our population in fact in many populations around the world and I want to emphasize the fact that over the last many decades a variety of studies have in fact emphasize the importance of genes in the etiology of these disorders clearly twin adoption and various kinds of family studies and all the phenotypic resemblance does not immediately lead us to the conclusion that genes involved in fact we have to be very mindful of common shared environments and in fact we have to be quite creative in measuring common environment in much greater detail than we do there's no doubt that despite the non-Mindelian nature that genes in fact have been involved the first phase of these studies by very many people in fact I would say has had really a single-minded focus on the search for single major genes in a whole variety of psychiatric phenotypes perhaps early on we assume that we would find genes such as those in fragile X that clearly are a single major gene however it does not produce a Mendelian phenotype but we know that these phenotypes would be much more complex there's often been a search in various of these traits in fact the search for structural phenotypes and this represents some very nice work and beautiful work from Nancy Anderson Iowa where she's used MRI as a way to find structural abnormalities of persistent sort of average differences between control and patients in this case schizophrenics to try and understand some structural pathology in these phenotypes but it's been clear that this hasn't been a very successful in fact approach and even if it were that's what we really basically need to understand which is one of the functional correlates and one of the rate limiting steps whose function when compromised in the human brain lead to psychiatric disorders I think it's fair to say that genetic mapping has confirmed not that they aren't genetic heritabilities in these phenotypes in a whole variety of them but rather than single gene variants cannot explain the majority of psychiatric illness that we see in the world but that of course should not and does not deter geneticists and psychiatrists of various sorts and proceeded and this has led to individuals proceeding to perform studies in a whole variety of families structures, small families nuclear families including affected sibling pairs and this represents data from a variety of investigators summarized in a meta-analysis by Judy Badd and Elliot Gershon of the University of Chicago giving us a sense of what the current status of these various genomic studies are or genetic linkage studies so what I'm showing you here is a cartoon of each chromosome chromosome one going to chromosome 22 and the X and blue and green representing sites where there has been significant evidence in this case with a p-value of one in a thousand or smaller significant evidence of the location of a susceptibility gene across the studies published in the literature the blank represents or this gap represents a centromere in fact the size of the bar here represents the total genetic evidence the Lord's score it's associated p-value if you will I think we all come away from this kind of a summary to say that there are many genes that are likely possible there's no doubt that there are probably some false positives in these kinds of scans and that the genes that in this case correspond to susceptibility factors either for bipolar disease or from schizophrenia are probably separate however it's quite interesting that the strongest two strongest signals across all studies come from human chromosome 13 and human chromosome 22 when in fact although not completely there's essentially overlap of the intervals in which susceptibility factors for both bipolar disease and schizophrenia in fact do occur suggesting that perhaps similar factors or at least similar genomic regions might be responsible and be a basis for both of these phenotypes I think the total field's impression is sort of summarized in this lesson 3 that the genetic factors are probably many and because of that and the total risk that we know the effects of any single factor in any given location of the genome are quite modest or small and if in fact we took these effects into account it would appear that no single variant or no single gene carrying variants for these phenotypes are likely to be necessary or sufficient and in many cases of course this would tend to imply that the susceptibility variants would have to be common in the population but this is really the hypothesis this is the grand challenge that we have not only in psychiatric disorders but in all common disorders or all chronic common complex disorders as you've heard throughout this morning and in fact most of the sequence that we're going to use to move forward has to be to refine this problem and focus on this towards the solution so progress has been slow but I would say that this is not a unique property of psychiatric illness one might raise the question then that the special nature of these phenotypes these phenotypes that are so uniquely human and the parents that we see of the intense heterogeneity across families is it because these genes have different genetic properties this idea has been raised many times just as there's been an overrepresentation of triple repeat mutations in various neurogenetic disorders but it appears that the answer to this is unlikely to play a very large role in our understanding of psychiatric disease over the last year or couple of years there's been a number of studies in which I think the role of some specific genes have become clear the evidence in fact is all going in the right direction and the evidence often is quite exciting and interesting from a number of different points of view they are now however just a small bit of the puzzle we are only explaining the segregation of individual genes that contribute to a disorder not really the segregation of the phenotype which is what we need to do but the genes that have been found so far in fact look like any other gene in fact the mutations of variants associated with them are likely to be no different if one takes these genes that have been identified by a variety of studies one could ask that well it's quite likely that our inability to detect mutations come from the very special nature we ascribe to mutations that they have to be in certain functionally recognizable regions of genes that they have to be unique to patients and for a common disorder for as most psychiatric disorders are but say for bipolar disease or schizophrenia this might not be the case so this is a study that represents work that we and others have done that represents a sampling of genes and resequencing these genes in volunteers in fact in controls I shouldn't say controls they just random volunteers whose phenotypes are unknown and this cartoon for these 40 genes shows only the coding sequence that is the length of the coding sequence that's been sort of sequenced the red and blue representing replacement changes and synonymous changes in the predicted protein and I would say that so far as the distribution the number and the frequencies of the individual variants that we find either by looking at all single nucleotide polymorphisms of those that really do alter the amino acid sequence of the protein appear to be no different than any other genes that we and others have seen and publish this is not to say that the dissection of complex traits such as this either by looking at genes directly that we've heard from a number of talks before or in looking at candidate genes or looking at all genes at a given location is likely to be extremely difficult as someone mentioned today morning I think we have to be patient for a little while longer because the task ahead is considerably more difficult this is a simple example where formal genetic methods as we have today in our hands can be used for a genetic dissection in this case of an oligogenetic disorder called a CNS or not a psychiatric illness but of the peripheral nervous system which is congenital agangoonosis and where we've been able to find some lessons that we're doing linkage studies doing genome wide association studies in a variety of populations what we can find is that hypomorphic mutations in the coding sequence as we always search for mutations that are outside the coding sequence that are likely regulatory and in these conserved genomic segments that you've talked about throughout this meeting as well as changes in this case on human chromosome 21 that represent altered dosage most probably as a current evidence suggests of the copper zinc superoxide dismutase that it's this collaboration of mutations between two kinds of receptors a tyrosine kinase a G protein copper receptor and dosage changes in this gene that all tend to be required to cause in this case congenital agangoonosis and that tells us I think that if this that tells us that this might be an important lesson that we should not be looking for a single kind of genetic change in fact the entire history of human genetics and looking at Mendelian disorders suggests that if any kind of mutation could occur it probably has in fact occurred and when we're looking at a multigenic disorder this is what we should search for so another lesson from what we know about the genetic psychiatric illness in fact in some kind of a summary is that given what we know given the extreme heterogeneity that we might expect for a phenotype that would be expressed in the brain and what we know about the genome structure and function even as of today I would say that our current genome searches for these kinds of complicated traits in fact warfully inadequate so here's the challenge where we take a phenotype expressed at the level of the whole genome and trying to get it to a single gene resolution finding individual variants associated with that gene and as I've talked to you before and others have pointed out that there are a number of impediments along the way having multiple genes those genes having low penetrance being common having the possibility of intense genetic interaction as well as something perhaps not emphasized of having epigenetic effects which are turning out to be such an important part of the regulation of genes in many genomes I think what I'd like to emphasize that the way forward in disease gene discovery will require two kinds of scans and we increasingly can do that and we should do that the first are scans of structure clearly those that we currently do in terms of genetic markers at increasing and higher resolution but clearly other kinds of scans of segmental aneuploidy that we know do have an impact on human disease and there are various kinds of studies which in fact can avoid this kind of a search functional scans of course are always on our mind and these currently of course can be done in cases where in fact one has access to the tissue but other kinds of genomic modifications such as methylation are likely to be very important and I think looking at the problem of complex disease to the lens of a single kind of technique is likely not to be successful I'm not going to go into the details because so many speakers yesterday have mentioned how comparative genomics is likely in fact to give us a small fraction of the genome in which most of the function is likely to be embedded and you've heard Eric Green talk about the work of his laboratory and collaborators that lead to identification not only of course of known exons but other perhaps alternatively supplies to cryptic exons as well as other functional elements whose functions sometimes can be predicted or sometimes can in fact be identified but often not in one way what we need to do is extract from the three gigabases of sequence a first line smaller synthesis if you will of where to search for functional differences and this is again a cartoon showing that if we did in fact restrict our attention to a very small segment of the genome of a few megabases where it trade locus or multiple genes that predispose to a trade map then we have to search not only the known parts that we currently do the estimated one and a half percent in coding exons but increasingly probably a minimum of three percent that would be various kinds of regulatory elements and epigenetic marks or to identify perhaps not sequence changes but other kinds of epigenetic and other changes that in fact can be inherited over short periods of time. I like to at the end point out that many authors have perhaps not pointed out that despite the difficulty of doing a genetic dissection of psychiatric trades that there are increasingly methods in this case using functional MRI for testing the functional effects of individual polymorphisms in genes that show an effect for these phenotypes. So this represents some very exciting work that's going on in the laboratory of Danny Weinberger here at the NIH in which case polymorphism in a gene that is very strongly suspected to be involved in schizophrenia that is brain derived neurotrophic factor that a polymorphism in code 66 can be used to find individuals who are both homozygotes for either type as well as heterozygotes give them specific kinds of phenotypic tasks or specific kinds of tasks to do and measure changes in the function of the brain in this case given the polymorphic sites that we know. So briefly the last couple of slides I think emphasizes what we need to do. I think often in various kinds of genome searches the lessons that we've learned from the mapping and eventual cloning of single gene disorders likely needs to be sort of expanded quite a bit. I really believe that we need to do linkage and or association studies at very high resolution and at the level of single genes and what I mean by that is 300 or 400 markers that led to the positional cloning and then mapping and positional cloning of single gene disorders is unlikely to work for complex traits because vast regions of the genome remain in fact beyond reach. I think justifiably there's been a strong concentration on making new kinds of maps of single nucleotide polymorphism which are the most common kind of genetic variants in all of us but insertions deletions are still 10 to 15% genetic polymorphism in humans and I do not think that ignoring them is quite useful. The other is I mentioned functional scans that these need to be done at single gene resolution and perhaps what we should do in the short term is scan using annotations of the human genome function that is the 5% or so that we think we do understand the function of although I clearly take Janet's point that there's so much we still don't know. Testing of candidate genes by phenotypic effects in fact is going to be a very important part of how we test candidate genes for these kinds of phenotypes and finally I think there are two parts of the changing culture that clearly has happened within the human genome project that needs to be adopted and really discussed within the much larger disease sort of community or disease related investigations. The first is our ability to collect and analyze very large amounts of data and very large amounts of accurate data in a quantitative way has to become the norm and this is often still quite a bottleneck. And finally changing the culture that is the collaborations that are so evident between various kinds of laboratories I think now needs to be broadened simply because many of these problems are so important and so difficult and the biology likely to be so exciting that it might not be the realm of any single laboratory alone. So thank you very much. I'd like to invite all of this morning speakers to the stage and while I'm doing that thank them all for very fascinating I think and interesting presentations. Each was asked to cover a huge field and I also am very delighted they were able to do that roughly within the time limits. So because we have a long time for lunch and we don't need to return to 1.30 I think we can spend 10 or 15 minutes now profitably answering questions. So I'd like people again to come to the microphones and I will point out that there is a microphone in the balcony so although people think they're hiding up there we can hear you and if you have questions please feel free to ask them. Yes. Tom can you turn on your microphone? I'm sorry how do you identify deafness in a zebra fish and I don't work with zebra fish but Sean Burgess who's here works with them and my understanding is that when the fish are in a dish and they're quiet if you tap the sides a hearing zebra fish will sort of form a sea shape and then swim away quickly and a deaf fish will sort of remain in place and if you touch them or spray water on them they'll move but they don't move the sound. Is that correct Sean? Okay. Rick. I have two questions first for John Hardy one of the few common variants that contributes to a common disease is ApoE4 an Alzheimer's disease could you comment on how that may intersect with the pathways you described? Yeah thanks for that so how does ApoE ApoE is really one of the best defined risk factors for any common disease E4 is associated with disease E2 is associated with reduce risk for disease Alan Roses who was the discoverer of this is in the audience so I'm sure that he might want to chip in I don't think it's known how this interacts with what the causes are there's been I would say three broad theories one the first that is interacts somehow that its role in cholesterol metabolism is the key feature and somehow cholesterol metabolism is involved in APP metabolism secondly there might be a direct interaction between ApoE and amyloid and thirdly there's a direct interaction between amyloid between ApoE and tau. My own view which is not worth any more than anyone else is really that I think that in the last year or two years the cholesterol collection has become more credible it seems as if cholesterol it does in fact influence APP metabolism and so my suspicion and it's not really based on any work of ours is that ApoE's effect is somehow to do with its cholesterol homeostasis in the brain but if Alan wants to say anything or I think he's saying he doesn't want to say anything which is remarkable too. I had a second question for Aravinda and Tony so as you see that many of these diseases may turn out to be fantastically heterogeneous with contributions from many genes that each have very small incremental effects individually how do you think this will play out in terms of diagnosis and treatment? Well I mean I think it's an assumption that it is so heterogeneous that it depends on what your assumption of small is the other is that when we are talking of small or large effects I think there is a major distinction which is that large effects are large in every family or case when effects are small what we currently don't know whether that small effect is sort of disproportionately distributed in a much smaller group of individuals I think so it's very possible to identify variants that in fact will have a phenotypic effect in a much smaller group of individuals and the implications for that will be very different but I think ultimately I think most of this knowledge will go to really understand what the rate limiting steps are I think that's where the main body of knowledge will point to and there's nothing to say that diagnosis of either with regard to prevention or any other you know part of the spectrum really has to be on a gene by gene basis for every single gene that is involved in a phenotype the diagnosis may have to do with some other surrogate that is a much much better clinical marker for what is to come Clem Perlong University Washington question for both our vendor and John as we see more and more effects of say intronic and exonic splice enhancer sequences and subtle effects on binding to the complex how do you think we're going to be able to screen for these are you going to have to bring in some proteomics to start looking at these very subtle differences and subtle effects by these polymorphisms I think that this is a very difficult question to answer it's a very I think that you know what we do to assess promoter variants at the moment is take a horrible little bit of DNA from the five you know from just upstream of the gene stick it stick it together with a horrible reporter gene put it in the wrong cell and then try and see how it affects expression and make credible assumptions about 10% effects it's horrible this is a real bottleneck in technology I think there's been a couple of papers this last year or two years and also we did some work a while ago looking at allial specific associations by promoter by promoter variants so looking at individuals who are doubly heterozygote heterozygote in a promoter for example and heterozygote for something in the open reading frame and doing allial specific expression studies and this is something which I we did this some years ago for apoE and it's been done more generally for between mouse strains by the MIT group and also by the cancer people from Hopkins whose names I'm blocking on now in terms of a genome wide type of approach too and I think that this is going to be a difficult technology to apply but certainly more biologically satisfying than promoter reporter assays so I think that there is the possibility to do this sort of stuff using allial specific expression studies but it's going to be tough. Just one other thing to add I think one of the ways forward is for us to really find what this 5% of the genome that appears to be under quite strong selection is. I think we have to sort of confirm the fact that that's the major hypothesis for the observations but if we could do that then one of the things we have to really learn how to do is intense scrutiny of that 5% and how looking at genetic changes that is sequence changes of those regions really lead to a phenotype. I think that's the major lesson we've got to learn but you're right, we haven't quite learned that very well. The group at Hopkins was headed by Bert Vogelstein. There was a very nice simple one page paper in science about six months ago dealing with genetic variability and expression and I think that that type of approach is going to be very important. My name is Venkat. My question to Dr. Hardee is the role of microglial cells, dendritic cells in not eliminating the precursors of the amyloid proteins may be contributing for the amyloid formation practically in almost all the Alzheimer's and then also prion diseases and it's the genetically defect microglial cells role in neurodegenerative diseases. Do you have any comment on that? I think the question was what does the role of microglial cells in a beta metabolism, in amyloid metabolism. And also there are any defect in certain gene expression by the microglial cells there, you know, fc receptor and then complement receptors associated with the genetic component of the cell itself. I'm not going to give you an intelligent answer actually, that's what's going to happen here. I can feel it coming. The there is I mean first of all I'm not an expert in the area but there is now of course a lot more interest in the role of microglial in clearing of plaques because what I didn't get time to mention is the strategies for treating Alzheimer's disease which have involved immunization and in the immunized mice and actually in the very recent autopsy report of an individual who'd received a beta immunization, it's clear that microglial do play a role and this role can be potentiated by by immunization in amyloid clearance but you know I'm not I'm not an expert. I'll just say that person who's very important for me and for English research in genetics was Bob Williamson. He was my head of department for seven years in England and he said the nice thing about genetics is that knowledge is a handicap because the less you know about a disease the more effective you can be in studying it because you don't go forward with prejudice and I've followed that advice ever since. Thank you. This is for Irvinda and John. A number of the studies that you guys both talked about identified haplotypes as conferring risk to psychiatric illnesses to neurodegenerative disorders and some of these involve genes that have been intensively studied with respect to the variation. Do you think that the effects are haplotype based or are they simply telling us about variation that hasn't been surveyed yet that's likely to be on those haplotypes? My prejudice is that it'll be to do with specific variability. That would be my prejudice but you're right we don't know and certainly with regard to the tau haplotype for example in PSP there's two possibilities in my view. The first is that it's just simply H1 haplotype expresses the variance of the H1 haplotype which are higher expressors or that there's variants which alter splicing and Mike Hutton's group is working very hard to distinguish between those two possibilities which might then allow in that specific case us to have a more molecular answer down to a single base change. With regard to the synuclein haplotype it looks as if in fact it's a promoter variability. The haplotype that we've shown is associated with sporadic disease the genetic variability in the promoter Bob Nussbaum's lab has shown is associated with not it's associated in one of those reporter essays shows a higher expression and that would be a very satisfying outcome. So I suspect that it's going to be specific molecular changes which are associated with disease we just got to find them. One last question. Can I add one other thing to it which is that it's quite possible that we'd have to back off from this finding of a single variant in a single gene that leads to the phenotype of interest and in fact you could argue that the APOL easy story is one such where there are at least two sites within that gene that's really responsible for it and you can imagine all kinds of scenarios given the identical functioning of those changes that would lead to different kinds of haplotype structures and the detection of association. So perhaps we should do it using markers to focus on the region to find the functional differences. Yes. I just want to ask about the connection between these kind of complex diseases like Parkinson's or Alzheimer's or diabetes. Lately we have been hearing a lot about from the media attention from embryonic stem cell search and also cell-based therapies for many of these disorders. Do we see a connection here? How much of this understanding will help us in cell-based therapies or are we really going too far ahead with cell therapies before understanding the disease completely? I'm a knowledge-based therapy type of guy myself, a small molecule. I don't really see stem cell research personally having an important role and certainly in Alzheimer's disease or really even in Parkinson's disease the risk of winding people up. I think we're looking for knowledge-based really understanding, once we understand the genetic basis of disease I think what we want, at the moment the dream I think is to have small molecules. That's what I think we should be aiming for at the moment. I think that cell therapies are much further away. Everyone believes we should have knowledge before we move forward? Okay, well two people have appeared. Are they short questions? Yeah. If you are disposed to develop a neurodegenerative disease or Go ahead. We'll get all three. Three quick questions. If we are disposed to develop a neurodegenerative disease or macular degeneration how do you explain the late onset of these diseases and can you elaborate also on the influence of environmental factor? Yeah, there's been some studies to look at the environmental factors but clearly there may be some components like smoking for example may have a role. Regarding the late onset of the disorder I think again it's the issue there are some alleles that have mild effect and that's an incredible feature of the disorder to me how people will live for 60-70 years perfectly fine and then they'll go blind we have a case of identical twins actually that had onset of the disease after 70 years within a few months of each other but I don't have an explanation for that. Can you speculate on the role of the vitamin C or the mechanism in improving that condition? You mentioned in vitamin A. Well vitamin A is needed for normal photoreceptor function and the thinking there was that if you study the histology of the patients with sword speeds macular degeneration the histology looks like there's a thickening of a basement membrane where actually the vitamins wouldn't be as accessible as photoreceptors so by helping the dose the hope was that it would get to the photoreceptors and that was the theory and at least works although it does not prove the hypothesis as to mechanism. Okay the last question from the balcony. Thank you I'm a nurse researcher just butting and wanting to have a better idea of how to conduct design international collaborative research. I think the model Nancy Wexler proposed and demonstrated very clearly and that is family gathering studies. I wonder whether any of you have any suggestions any speculation about how we can move forward not just with that design but any other kinds of designs to include the rest of the world in genomic research. Wow that's a tough question actually and it's getting more difficult. I think that international collaboration for genetics research is absolutely essential for several reasons. First of all family sizes here in the states are tiny so we can't really do it's getting more and more difficult. Cousin marriages are pretty rare here in the states and so we really need international collaborations for selfish reasons as well as of course to help people in other countries with diseases and it's getting more and more the paperwork is getting more and more difficult to do this sort of stuff and I heard the model that we're trying to adopt in our lab is to have researchers come from different countries all over the world and bring their samples and work on their samples with us and then go home afterwards with the samples and with the results but of course even this is getting more and more difficult with visa issues so this is a real headache which the politicians need to help us with I think. Tom did you want to comment? Okay well let me thank the speakers this morning for stimulating sessions. I believe there is lunch in the atrium is that correct? Yes and you should be back at 1.30 for the session on the implication of genomics for health care. Good afternoon we're going to get started. My name is Ray Nard-Kington I'm the Deputy Director of the National Institutes of Health let me first bring to your attention these two meeting scheduled for tomorrow genetic variation and gene time environment interactions in human health and disease at Missouri auditorium and genes brain behavior before and beyond genomics in Wilson Hall and Building 1 there are still openings for both of those sessions there is no NiAA so unless we grew a new institute in the last day or so it was just possible but I think that meant NiAA welcome to this session which is to focus on implications of genomics for health care we're at an interesting point in our health care system in the United States we have a large percentage and growing percentage of our gross domestic product devoted to health care now around 14% or 1.4 trillion dollars double the percentage in 1970 we have a large number of individuals who remain uninsured probably around 15% or so we have growing concern about how this health care system functions particularly with regard to quality of care and growing concern and evidence of differentials in treatment within our health care system across subgroups of the population so in the midst of all of those things we have the genomics revolution overlaid and today we'll have two talks to address this issue first Dr. Wiley Burd whose chair of the department of medical history and ethics at the University of Washington thanks very much I appreciate the opportunity to talk with you and I think I need some help on my slides here genomic health care as we begin to think about how genomics is already starting to impact health care as we project to the future I think it's important to think about two pathways from genetics to health benefit the first pathway is the one that has received most discussion and I want to reflect upon it but I think it's very important for us to keep within our sites the second pathway as well so the first pathway is the concept of reducing risk in people with genetics susceptibility basically tailoring health care to the person with the particular genetic susceptibilities what's often called personalized health care another strategy is one in which we use genomic information to develop new treatments innovative approaches to disease where ultimately we're going to be tailoring the treatment to the disease rather than to the person and I'm going to reflect a bit upon these two differences as far as tailoring the treatment to the person is concerned we already have some very robust models for how this might work newborn screening being an obvious one we find infants with phenylketonuria we do this by a mandated screening program present in all the states in this country and we give them a phenylalanine poor diet because we've been able to predict from their genetic predisposition that that's where the problem lies and indeed we have been able to prevent mental retardation as a result so that represents a kind of gold standard to bring out what the genetic problem is and then addressing it with an intervention here's a more recent example and I think it represents a very robust area of growth in health care practice right now this is the example of inherited medullary carcinoma the thyroid familial MCT this is an autosomal dominant condition and what we see here is a grandfather who died of medullary thyroid carcinoma his son has the same problem and of course had a 50% chance of inheriting that condition and now we know that this condition is caused by mutations in the ret gene so he's been tested the majority of people with this condition can have an identifiable mutation in the ret gene we find it it doesn't really change his diagnosis or his care but it gives us the ability now to test his children and to offer them a preventive therapy in the form of prophylactic thyroidectomy with thyroid hormone replacement so again this is finding the genetic problem finding the genetic susceptibility in the individual and then tailoring treatment to that individual where this model gets complex is with most diseases because most diseases that are of public health significance in this country are what we commonly refer to as the common complex diseases where there are multiple different causes where genetics is a contributor and a very important identifier of risk in some in some cases where there are major gene effects background gene effects gene interactions contributing to risk and this is true of cancer diabetes heart disease and most diseases generally but where environmental factors exposures diet lifestyle play a major role as well the reason why this multi factorial model is extremely important to us as we think about the implications of genomics for health care is that increasingly we are going to be identifying those major and minor genetic effects that contribute to risk for common diseases and we need to start thinking about when is that risk information useful and when is it not useful so just as an example of this paradigm we have factor 5 Leiden this is a genetic trait present in 1-5% of the population and it is clearly unambiguously associated with an increased risk of venous thrombosis if you look at published data you can see that in the general population something like 4 people in 10,000 are going to have a blood clot within the coming year that rate gets higher if you look at people within the population who are also exposed to non-genetic risk factors that promote clot the baseline risk is higher in people with factor 5 Leiden and among that population higher yet if they are exposed to a risk factor indeed when we look at causes of venous thrombosis they follow exactly that same complex causality model that we can see for common diseases generally and in the case of venous thrombosis we have identified several genes that are associated with an increased risk so mutations in factor 5 in the prothrombin gene and antithrombin 3 deficiency protein C and protein S are all associated with an increased risk of venous thromboembolism but in addition we have very well defined risk factors that are within the environment non-genetic risk factors that promote or increase the likelihood of venous thrombosis one of them is simply age one's risk of developing a blood clot gets higher with age surgery, trauma, exposure to hormones pregnancy, prolonged rest, etc so this is a complex causal model and it raises two questions concurrently one is what's the value of finding someone with factor 5 Leiden what are we going to do differently and in particular to what extent are we able to manipulate factors in order to change risk for people with genetic susceptibility or for people generally one of the particularly interesting complexities when you look at something like factor 5 Leiden where we've identified other genetic risk factors is the degree to which we're not sure what genetic profile is going to be the most useful clinically to identify here's an interesting study in which people relatives of people who had clot were assessed so these were people that had blood clots and were found to have either factor 5 Leiden or prothrombin variant which is present in one percent of people and has increased risk for blood clot and having found those individuals the study went out to family members first degree relatives of these individuals and asked what's the risk for blood clot in people that have the same genetic and you can see first of all that the prevalence is increased above general population risk just as we saw in the previous numbers but not hugely increased and in fact the highest risk amongst relatives was people who had both traits together so are we interested in finding people with factor 5 Leiden to do something different prothrombin variant to do something different or are we mostly interested in the more significantly increased risk among the smaller groups outside of people who have both as we think about using risk information we really have to think about what threshold of risk is going to lead or should lead us to do something different in order to ameliorate risk and of course what we want to know then is what are our options to ameliorate risk and what are the dangers of doing those interventions and so what we have is a variety of potential interventions for people with factor 5 Leiden we have anticoagulant prophylaxis either long-term or episodic around the exposure of some risk situation like prolonged rest or surgery we have obviously avoidance of other risk factors to the extent that that's possible oral contraceptives being one that's been quite a bit discussed but for every intervention that or at least every medical intervention that we might make anticoagulant prophylaxis avoidance of oral contraceptives there are certainly potential adverse consequences and we can emphasize that particularly with anticoagulant therapy where there are bleeding complications and so what we'd really like to know is if we use some form of anticoagulant prophylaxis in someone because that person has an increased risk on a genetic basis of having blood clot are we certain that that person is better off we may be reasonably certain based on current clinical data that we're going to reduce blood clot but are we going to introduce bleeding complications if we're talking about avoiding oral contraceptives what are the consequences of not using what is currently the most effective contraceptive therapy particularly given that pregnancy is associated with increased risk for blood clots and here we have a tremendous data gap and this is generally going to be true this is called the therapeutic gap this is the problem of increasing rapidly increasing information about risk and relatively little information about outcomes in particular very few controlled data that enable us to measure the balance of risks and benefits as we identify an increasing number of risks for common conditions let me now talk about colorectal cancer and elaborate this point just a bit we've known for some time that there are genetic factors contributing to colorectal cancer risk one piece of evidence is epidemiologic data that summarizes that shows us very consistently that people with a family history of a first degree relative with colorectal cancer have approximately a two fold increased risk above the general population and what these data show us as well is that that risk kicks in a little bit earlier generally if you have a first degree relative with colorectal cancer your risk at age 40 is quantitatively similar to the risk of someone without such family history at age 50 this is an important context because current screening recommendations are to screen everyone at age 50 and so one of the conclusions that come from these observational data is that perhaps we should screen people with a family history earlier but of course it's more complicated than that when you look at family history of colorectal cancer you look at a continuum so at one end of the extreme we might have someone whose grandmother had colorectal cancer at age 80 and in fact statistically that person's risk of colorectal cancer is average there really isn't any significant bump up someone whose mother had colorectal cancer at age 60 now we're talking about someone who probably does have at least a two-fold increased risk even if that's the only family history the person has and then you have a very small subset of families that begin to look like that familial medullary thyroid carcinoma and these are people with inherited colorectal cancer we have two syndromes familial adenomatous polyposis and hereditary non-polyposis colorectal cancer and families where there are sequential generations with colorectal cancer where the onset of cancer occurs at an early age are families in which we strongly suspect that this genetic syndrome is present so here are the two syndromes with familial adenomatous polyposis virtually 100% lifetime risk of colorectal cancer we treat these individuals with prophylactic colectomies and we know the gene, APC in PCC the lifetime risk is not quite as high we also have identified now at least five genes and these are all genes that have a particular function of mismatch repair that is there involved in DNA repair functions so we're beginning not only to be able to identify these families by pedigree assessment but to be able to offer genetic testing within the family as we did with the to identify prospectively persons at risk and we have complexities part of bringing genomic medicine into health care wanting to identify these high risk people in order to tailor care for them is making sure that clinicians have appropriate skills in taking family history in the case of inherited colorectal cancer syndrome this includes the complexity in HMPCC that endometrial cancer is also part of the syndrome approximately 40 percent lifetime risk in female carriers this particular pedigree wouldn't quite make the clinical definition of HMPCC if we base it only on colorectal cancer but when we include endometrial cancer in the definition it does meet the criteria and we can say this is an HMPCC family so we're looking at a need to increase family history taking skills in order to pick up the consequences as clinicians begin thinking about this though they need to think about relative prevalence somewhere between 7 and 8 percent of the population have a first degree relative with family history of colorectal cancer might be candidates for early onset screening at age 40 instead of age 50 and a much smaller subset have the kind of pedigrees that raise the question of HMPCC and FAP now as the clinician gets ready to tailor therapy needs to be able to make these kinds of distinctions so that we have a sort of triage function here people at average risk will be offered colorectal cancer screening at age 50 if they have moderate risk they've got that first degree relative but they don't meet the high risk criteria then we're going to start screening at 40 instead of 50 and if high risk we consider genetic testing we start screening at 20 in the case of HMPCC and in the case of familial adenomatous polyposis we would go further and offer prophylactic colectomy so the tailoring therapy to the individual gets increasingly complex as we gather additional information but what's also interesting and extremely exciting about the colorectal cancer story and the increasing knowledge about genetics of colorectal cancer is that we're now beginning to understand a carcinogenic process at a molecular level there's been a lot of work starting with seminal work by Dr. Vogelsen and colleagues to define a sequence of events that occur in the development of colorectal cancer and the sequence of events basically is a process by which normal epithelium develops a small polyp the small gets to be intermediate and then large and cancerous changes occur we have cancer and then at a certain point we have metastasis and it's increasingly clear that this process of carcinogenic development is a process whereby a cell accumulates mutations in a series of genes it also seems increasingly clear that there is an orderly sequence that is that certain genes need to mutate early in the process and are permissive of later stages of the process and that other genes may play a greater role either in the transition to cancer or the transition to metastasis and we've identified some of those genes and what's extremely important here is that as we identify these genes the genes involved in the inherited syndromes are both implicated mutations in the APC gene are an early event in colorectal carcinogenesis in people without an inherited susceptibility and that enables us to understand why familial adenomatous polyposis creates a cancer susceptibility people who've inherited a mutation in the APC gene are basically one up in the process they're already one step down this pathway toward cancer in all their cells and fewer events are required to go on to cancer the mismatch repair genes appear to promote mutation basically by failing to repair DNA errors, errors in replication and there is evidence that mismatch repair may be particularly important in the APC gene in repair of mutations in the APC gene but perhaps also in other genes as well so it appears that mismatch repair mutations have an effect at multiple places along this pathway well that's important in a variety of ways the more we understand molecularly what's going on in these inherited syndromes the better we may be able to refine the tailoring of treatment to the people with susceptibility but when we think about the big picture of health care concerns and improvement in health care outcomes what may be even more important here is the way in which studies that started with work on inherited syndromes now will open up new avenues for addressing the problem of colorectal cancer generally in the population so as we look at what's coming in the future first of all we can say that this kind of molecular work is likely to provide us with much better explanations than we've had so far for a variety of observations that are important in understanding colorectal cancer why for example is there an association with intake of folate and colorectal cancer risk to what extent are there other genes or is there an interaction between folate and the genes already identified that may explain this why do non-steroidals and cox2 inhibitors seem to inhibit polyp promotion what's going on there what kinds of interactions are occurring between those drugs that we will increasingly identify as we understand this pathway better and maybe a really key question here an important illustration of how we can learn from inherited syndromes what what is the explanation for the variable outcome in hereditary non-polyposis colorectal cancer in familial adenomatous polyposis cancer mutation carriers virtually always get colorectal cancer the figure is only 70% in HMPCC 40% of women get endometrial cancer 60% don't what's the difference what kind of modifying factors result in some people with a genetic predisposition getting cancer and others not and if we understood those better could we manipulate those for either prevention or better therapy and then obviously the more we understand the molecular events we'll be able to identify increasingly better screening tools there's already work going on identifying the presence of mutations in this carcinogenic pathway in stool samples as a potential screening mechanism for colorectal cancer generally as we think about where this is going in the future we're clearly looking at genomic research as a pathway to new therapeutic hypotheses obviously we'll then need to be evaluated tested and hopefully a fair number of them will actually become new treatments the steps along this pathway are clearly to define a complex biologic system we can already see from what we know about colorectal cancer that there are multiple different functions that interact with one another we know that there are non-genetic modifiers of colorectal cancer risk we suspect that there are a variety of complex interactions between genes and between genes and environment and we need to find those modifiers and identify them because they're likely to lead to new drug targets new opportunities for intervention that may be entirely innovative therapies understanding at the molecular level is opening the door to think about new ways to treat and as we do so I think we can have reasonable optimism that there will be an opportunity to expand the benefits well beyond people with particular inherited syndromes that we're now beginning to understand obviously if we have truly innovative methods for prevention and early treatment the more accessible the safer they are the more we have potential for improved outcomes for all people with colorectal cancer or all people who might have gotten colorectal cancer and not just for people with a particular genetic acceptability so I just want to return to this question tailoring health care to the patient or the disease and I want to propose two things first that both models are going to be extremely important as we think about applying genomic knowledge to health care that when we think about tailoring care to the person that's going to be our first opportunity and likely to remain the most important opportunity for people with high risk that is unique ways to start screening early or apply particular prophylactic procedures are likely for the foreseeable future to remain extremely important for people at high risk but the genome project I would propose is likely to have its greatest impact as it develops an increasing understanding of disease processes themselves and therefore ways to apply benefits broadly in the population obviously in order to accomplish this we have some tasks and they all come under the heading of rigorous evaluation one of the things that we have to be very careful about as we see expansion of the model of identifying genetic risk and tailoring therapies is that we need to continually be thinking about the potential harms of the label we don't want to label people unless we can offer them a net benefit and so there's going to be a constant evaluation both of how much benefit we can offer based on knowledge of genetic risk and how much harm or potential harm comes into play because of that label and that harm clearly is partly social concerns with genetic discrimination but it's also medical to what extent does someone who's labeled as increased risk then become exposed to a lot of unproven therapy on the hope that it might work so we need to keep thinking about that we also need to think carefully and continually I think about how does new genetic knowledge as it comes forward have the potential to be translated what are the translation opportunities how can we use it most effectively to develop robust hypotheses and test them efficiently to get to those innovative prevention methods and therapeutic methods that are going to be the big gains and obviously as we do so how do we ensure fair access to all for this promising new technology thank you our next speaker will be Dr. Peter Goodfellow Senior Vice President of Discovery Research at GlaxoSmithCline the chairman was teasing me as I walked up because we discussed my previous career in the circus and he wondered whether he should mention that before I started talking and I encouraged him to do so but he decided it would lower the tone of the meeting so I'm in a philosophical mood and I was even thinking about doing an experiment where I asked the audience whether they would like to vote for philosophy or science but I realized that if I gave you that opportunity you would all vote for science and you wouldn't get the philosophical bit so I'm going to give you a little philosophy and then I'm going to get to a little bit of science I have said and I believe that the most important discovery of the last century was the determination of the structure of DNA I also think the most important discovery of this century will be the complete sequence of the human genome and it is an unbelievable pleasure and honor to be studying biology and the biology of humans during this period of time and it made me sort of think back when I heard the other people this morning talking about when they first heard of DNA and I tried to they were very astute individuals who read it in their comics or heard it on the radio I actually learned about DNA actually in a sexual context and let me explain this to you when I was 14 years old in 1954 I was in love with a girl called Rowena the problem was Rowena was in love with her horse and the Minaja Chua wasn't really working out so Rowena being a kind girl decided to set me up with a blind date with one of her friends so I went to a coffee bar and was introduced to a young lady called Gabrielle and Gabrielle said to me what do you do? I said well I'm a biologist and she said oh that's really good my father's a biologist too he's discovered the structure of DNA now I must have looked rather blank because she said you know the double helix and I realized I needed to say something so I said oh that DNA so I went home and looked up what DNA was and ever since then in my mind DNA and sex have been intertwined although I never had another date with Gabrielle so I didn't get to be Francis Crick's son-in-law but that's how life goes if we look if we look at how we fund research I think we have a problem and the problem is this this is NIH funding and this year I think it will be 26 billion dollars will be spent on basic biomedical research and I would argue that the reason that society gives us this money is because they expect us to produce something and what they expect us to produce is new treatments and this is the problem we're actually not producing more drugs each year we're producing less drugs each year so that in the year 2002 last year there were less drugs produced than any time per year since 1980 now you could say that's because the pharmaceutical industry isn't playing its part it's cut its investment in R&D to make new drugs but actually that's not true the investment in R&D in the pharmaceutical and biotech industry has gone up not quite as much as it's gone up per NIH but has gone up I think we have a real problem that someday we may walk into an argument that says hey we've been giving you all this money not actually giving us what we paid for and I think that problem could get very acute because there is a problem which is inherent in how we discover drugs if you have an idea today and you go into the lab and start working on that idea it will take you roughly 20 years before you have a therapy derived from that idea so when we get up and say we've sequenced the human genome and we're going to bring you wonderful opportunities you need to just put in brackets sometime probably after one decade and if we're lucky before two decades and I think we have to put that message involved in the talks that we give in public so the drugs that we have on the market today which we use to treat disease mostly come from chemistry and work which was done in the 1970s it's true that there's a little bit of work which is now based on the gifts of the 1980s cloning is beginning to make a contribution monoclonal antibodies are beginning to make a contribution but the work that we're doing today in the labs and genetics and genomics in terms of producing new therapies perhaps it's different for using the therapies we already have but in terms of producing new therapies may well take us another 10, 15, 20 years why does it take so long well this is how you make a drug you start with a protein and you mix that with chemical diversity GSK we have a file library of about a million small molecules which are the precursors of drugs and we mix each one of those small molecules with our protein target looking for something which will bind to the protein when we find something which binds to the protein and alters its properties increases them or decreases them we then go through a process of lead optimization there at the molecular level you engineer into that small molecule the properties that you like to have in your drugs oral bioavailability once a day dosing and so on after that you take your molecule you test it for safety and efficacy in animals if you have an animal model and then you take it to the clinic to go from the thought through to the point where you start to do your experiments in animal models and it goes right 5 years from the animal models to being sold having gone through the clinical evaluation 7 years 12 years if everything goes right and if there's one lesson I have learned in the last 10 years of working in the pharmaceutical industry it never goes right it's an experiment and it's just like all experiments scientists do they don't work most of the time but nevertheless our aim is to translate new information into new therapies and what we're trying to do one of the themes that we are trying to explore at GSK is to actually copy what we've done as a community in the human genome project I would say the human genome project actually started with the hypothesis that it's just worth doing so let's do it we started on a process of a gene by gene basis but it was inefficient so we got together and said let's just do everything second thing that we did was we actually realized that machines automation should mix things together in test tubes that's a much more efficient way of mixing things together in test tubes than having people do it so automation and a global approach are two potential ways of making the process of drug discovery more efficient and so if you look at the drugs which are currently on the market today and ask the question what do they interact with what you can do is you can identify classes of targets which historically have proved to be suitable targets for small molecule intervention and these are just listed here about 40% of all the drugs on the market target GPCRs G protein coupled receptors the next most common class are hormones, nuclear receptors and so on if you take those tractable classes of targets and add up all the genes which are in the human genome there's about 2,000 of them so instead of sequentially trying to pick the latest, the best target why not turn the whole thing on its head and say let's just make a drug for each of those 2,000 we can use the drugs we've got to test hypotheses in animal models and clinic or if somebody comes up with some great data saying hey that's a good target then it's instead of phone a clone as we used to it's now phone a drug and you have that drug available to test the hypothesis that you'd like to test so I'm going to give you some examples of that type of genomic approach to drug discovery I'm going to talk about GPCRs a little bit about nuclear receptors and finish off with kinases each one of those classes are at a different stage of their development GPCRs the problem with GPCRs is that although we have them all cloned we don't actually know what the ligands are which activate the receptors and so systematically we set together a process whereby we could identify all of the ligands for all GPCRs and just to make the point if you look at those 40% of all the drugs currently on the market today they actually come from only 30 members of the family A class of the GPCRs so here's the strategy let's identify all the GPCRs let's clone them let's put them into an expression system where they're coupled such that we can recognize when they're activated then let's find a large bank of potential ligands anything that you can imagine that might activate let's use those and then when we find activating ligands we've now got a tool where we can start biological evaluation and also incidentally set up a high throughput screen to make the drug and this is just examples of all the different molecules which we put into our ligand bank and here is a list of some of the orphan receptors which have been paired with ligands in the last 10 years and several of these like the erections, uretensin work which was done at GSK have now got small molecules which are in the clinic I was going to go through a recent example, nicotinic acid activates receptor HM74 however that activation is nothing like as potent as that you measure in native rat tissues and the reason for this is because it turns out that HM74 is a member of a small gene family and the real receptor is HM74A this is an example nicotinic acid is used in the clinic it increases HDL levels and now we know what the receptor is, we should be in the position where we can make better molecules than nicotinic acid so here, I don't know if there's anybody in the audience who's a structural biologist hands up if you're an x-ray crystallographer anybody I was going to give you a really good tip here if you want to get a no, I think you're a cameraman sir if you want to get a Nobel Prize then if you can determine with ease the structure of GPCRs bound to ligands then you deserve a Nobel Prize and you will transform this particular area of drug discovery that would be the single largest contribution that could be made I think by structural biology to drug discovery in the nuclear receptors there's instead of the 400 or so GPCRs there's only 46 48 nuclear receptors in the human genome and we gave ourselves the task of trying to both clone express and get the structural information on the ligand binding domain for all of these and we gave ourselves the task of doing that before the end of this year the green dot means that we've done it either in house or it's in the literature two of the red dots we've converted into green dots in the last couple of months so we're well on the way of having this complete target class available with structural information and that structural information is already beginning to give us clues about the biology for example if you look in the ligand binding domain it turns out there are two classes of receptors there are those which have four conserved phenylalanines which obscure the binding the binding pocket those are non-ligand regulated constitutively active transcription factors and they are strongly conserved across vertebrates and invertebrates the other families are those where they are ligand regulated and are used as sensors for hormones and for dietary and xenobiotic ligands and these are rapidly evolving and again if there are any structural biologists in the audience you can get another Nobel Prize for solving the structure of not just the ligand binding domain of nuclear receptors but the nuclear receptor ligand binding domain complexed with the DNA binding domain complexed with DNA complexed with the activation and the repressor factors which go to make up the machinery of the transcription factor actually working and finally kinase targets where we have managed to link structure and chemistry so that drug discovery for kinases is getting to be much more predictable so again the same idea holds lets clone all of the kinases and when have them all available have those kinases identified structural information available on them and look for binding of the kinase to the ligand and repeat the process such that you look at lots of different kinases bound with lots of different ligands so solve the same structural templates with many different substitutions and solve the same template with many different kinases because of the conservation in the ATP binding site although it's conserved it's not optimized for ATP there's space in there which enables you to build into your compounds the specificity that you need which gives you targeting for individual kinases and because you have all kinases available to you, you can test how specific your molecules actually are and it leads to another type of approach where we can actually decide to try and do an experiment which is the nirvana for I think structural biology is it possible to predict tertiary structure from primary structure can we tell the shape of a protein based on the amino acid sequence or so far no one's been able to do that but here's an experimental approach to that problem let's group all of the kinases into related families and draw up trees based on the sequence alignment in the primary amino acids in the ATP binding site if we do that what you'll see is a sequence based grouping now let's do the same experiment but test different kinases with lots and lots and lots of different small molecules which bind in the kinase binding point in the ATP binding site in kinases if you repeatedly do that experiment there's a point where the tree that's built on sequence matches the tree that is built on small molecule relatedness in the ATP binding site at that point you will be able to predict structure from the primary amino acid sequence now I'd like to be able to say well here it is we've done it actually I think we've still got quite a way to go this is based on selection of 10,000 small molecules which have been bound to different kinases but you can see the trees don't yet look as though they're overlapping so here is my prediction for the future I actually think within the next decade we have a real chance of being able to link genomics to chemistry by using the bridge of structural biology and if we can do that and as I've already said one of the big barriers is going to be GPCRs if we can do that within the next decade it'll be possible to make drugs for all members of the currently tractable classes of drug targets if we can do that then we will cut at least 5 years off that drug discovery time frame because we'll already have the small molecules ready to go when you have a hypothesis that you want to test in the clinic I'd like to thank my colleagues at GSK who did all the experiments I talked about and provided the slides I'd like to particularly thank those people who were prepared to take time to teach a failed academic the you're supposed to laugh at that let's try that again I'd like to thank them for taking the time to teach a failed academic how to, thanks guys how to make drugs and I'm particularly grateful to SB and GSK for allowing me to play with a big chemistry set thank you very much we now have time for questions questions let me go on please Dr. Burke you've just gone through an experience of serving on a panel to consider some of the issues that you discussed and then found that panel dissolved by the new administration with some newly formulated entity coming into being I wondered if you could just give us a up-to-date version of what's happening and what if any fallout you see from that experience or that you're concerned about two comments in terms of the comments you make there is a new advisory committee on genetics that's been formed to follow up on the secretary's advisory committee and genetic testing and that committee has a broader purview it's not just going to look at oversight of genetic tests it hasn't met yet so I think we simply have to follow and see how the work of that committee plays out what I think I learned from being on the secretary's advisory committee of genetic testing is how important it is to look in a very broad way at oversight of genetic tests so if we just look narrowly at what genetic testing and what can we do as a society to make sure that genetic testing goes forward in a responsible manner what I think the secretary's advisory committee in genetic testing did was unearth a layout for discussion a variety of different issues the committee did look at regulatory oversight and came to the conclusion that there was an opportunity for regulatory oversight by the FDA for pre-market evaluation of genetic tests and that recommendation was made but at this point does not look like that recommendations going forward but I think what is most important is to recognize that whatever one might get from pre-market evaluation it's just a small part of the picture of what needs to happen in a society to make sure that genetic tests are used responsibly what I felt I learned was that pre-market evaluation would probably lead to a careful evaluation of laboratory methods we can legitimately say that the CLEA certification process may also provide that and would lead to good labeling of genetic tests which is a very important property that we'd like to see developed but it isn't going to address the issue of preparing health care providers for the use of a big new complex technology it doesn't address the issue of informed consent practices and the fact that different genetic tests have different requirements for informed consent it doesn't address the issues of public education so I think there's a whole lot of other things that are important and perhaps better addressed in a non-regulatory model Peter I'd like to ask you with this approach of developing drugs for all of the tractable targets what will be the approach to trying to identify their indications I tried to intimate that at one level you can say you use the information which is generated by the whole world it's very similar to the period of time when cloning started to become obsolete because people had already cloned the gene and therefore when you wanted to test a hypothesis you picked up the phone and you asked for the cloned gene to be delivered to you obviously within GSK and many other companies and academia a lot of emphasis is being placed on genetics and Alan Roses is using a large program to try and identify appropriate targets for therapeutic intervention as well as both safety and efficacy issues with the drugs we already have using genetic approaches the other possibility in this also has a long distinguished track record is if you have a small molecule which is specific for a particular target you can use that itself for target validation and I think eventually we'll want to be in a position where we can test directly in the clinic whether a hypothesis is correct or not a number of people have raised the concern that genetic services now are primarily utilized by an affluent educated patient population what are your thoughts about ensuring better access to genetic technology and testing as we move forward it's clearly a complicated problem and it's made more complicated first of all by inequities in access that exist now as you're referring to it so I think we have to be very concerned to look carefully at those inequities and ask how we might for example increase a workforce that provides genetic services and increase reimbursement for genetic services but I suspect that a very important other part of it is really being very thoughtful about what in what ways we could integrate genetics into clinical practice clearly they're going to be continuing there's going to be an important continuing need for services that we call clinical genetic services for genetic counselors for medical geneticists providing specialized services but as we increasingly identify genetic risk factors for common diseases the question I think has to be how can these these genetic tests and ensuing interventions become part of routine clinical care I think there's a clinician education issue here I think there needs to be a very careful effort of geneticists and other clinicians working together to develop clinical practice guidelines and I think that will at least have some potential to improve access clearly we have major problems with access in our healthcare system generally and as with any new technology that will become a concern with genetics Hi I'm Ron Johnson from Human Genome Sciences I have a question for Peter so you talked a lot about known drugable targets and I wanted to know what your optimism was in identifying new classes of targets that are drugable and how will genomics be now that we have the Human Genome what approaches will be used to identify those Well some classes I didn't mention are clearly already tractable phosphodiesterases would be one class for example there will always be cases where the biological rationale for a particular target is so strong that we should try and find a way of intervening therapeutically at that point even if it is not very tractable but I think I would still argue very strongly that if we can systematically approach the problems that we have then we can be much more effective and that's also true of the target classes which are at the moment not particularly tractable and in an analysis we've done you can actually point out what is the thing that needs to be solved in order to move a particular area forward so iron channels has been held up for a long time by the lack of high throughput patch clamping technology those have now been developed so it really depends on the class by class and I'm not arguing that one should walk away from the rest of the genome I'm arguing that there's some low hanging fruit and why don't we pick the low hanging fruit I'm Patty Walter and I'm from the University of Northern Colorado School of Nursing but I'm also a patient who's very thankful for the discovery of monoclonal antibodies and as a person who's being treated with that we have significant family history and so genetics testing is very important to me and I feel a responsibility for my two sisters who have three children who are girls each of them and what are we at in the social status of the ramifications of insurability not only for ourselves but our children we find out what our genetic predispositions are clearly there is a lot of concern about the potential that genetic information genetic risk information may affect people's access to health insurance to other kinds of insurance to employment the fact that a majority of states have passed anti-discrimination legislation the fact that there's a continuing interest in trying to develop some form of federal legislation I think speaks very strongly to a societal consensus that there are certain uses of genetic information that are unfair I really don't think there's any question that there's a consensus on that point what I think is the problem is figuring out how best to enforce that protection so the genetic anti-discrimination laws that are on the books haven't really been tested in court there's a lot of concern I understand from legal colleagues that the bills that exist may not provide the protection that we hope that they'll provide and there's also a sort of concern under the heading of genetic exceptionalism is it appropriate to protect someone from discrimination because they have a particular genetic factor that identifies risk and not protect them from discrimination from another non-genetic factor that identifies a similar risk and I guess what I think we need to move forward with is that dialogue as a society and looking very vigorously at what policy options we have to prevent what I think we all agree is unfair which is the prevention of good health insurance coverage to people on the basis of some sort of medical risk information I was going to comment that in the UK we took a very similar approach where there is a moratorium for five years on the use of genetic information for insurance purposes and I think that this was designed well it was designed specifically to give us a breathing space where we could debate the issues which are raised by having the potential of more predictable tests available Hello my name is Gerardo I'm from the City College of New York I have two questions for Dr Goodfellow How do you foresee a knowledge or development of the proposed database linking the primary and tertiary structures will generate faster drug development and secondly how might one make effective use of the molecular systematics data we have right now to advance such a Could you repeat the second one again or let me answer the first one and then repeat the second one I was just playing when I talked about linking primary structure with tertiary structure what I was saying was that I could see an experimental approach which theoretically could link the two together and might solve this problem but we're nowhere near solving the problem although I did want to stress the fact that having three dimensional structures of proteins is absolutely key to linking genomics and chemistry together and all of advances that we're making in the drug discovery process are rooted in that link between chemistry and biology when we're dealing with small molecule therapeutics What was the second question How might one make effective use of the molecular systematics that we have right now to arrive at this end to arrive at the development of a database linking the primary and tertiary structures and that leading to drug development I'm tempted to say that following on from one of the speakers this morning he's absolutely right about your hearing going once you reach the age of 50 but I think I'll answer the question which is the and again this goes back to something one of the speakers said this morning we are generating enormous datasets which are very rich and we need to develop the IT infrastructure which enables us to handle those data points so for example in GSK last year we generated about 150 million data points so if you can't look at that with a piece of paper and a pencil you have to have new ways of visualizing the data manipulating the data because as he also said at the end of the day it's the cortex which is the important part of scientists and that's the bit that you have to be able to interface with the data so one last reference to earlier on that's the people need doing science going forward I think creative scientists have to be trained in interfacing their brains with data using IT methods so if I was having my time again I would be doing computer science learning informatics as a way forward Mike Iderola from NIDCR so if chemical space is infinite does a 1 million compound library cover all the bases and does it generate robust leads for you to follow up what kind of constraints does it impose as well well it depends on what you define by infinite some people think 10 to the 60 is the number of opportunities which is kind of infinite like I think you can collapse that down to about 10 to the 20 which is still sort of more molecules than on the planet earth so that's still kind of a large opportunity I think the serious answer is that you don't have to sample all of chemical space in order to find a chemical which will do the job that you want doing so yes you're right when there's 10 to the 20 possibilities and you're only looking at one in 10 to the 14 of them you've got a serious problem which is another way of saying eventually we have to learn enough about being able to predict what we want that we can do this in silica but we're nowhere near that yet we have to collect more data to make what we do more predictable a question a couple questions for Dr Burke one of the key points you said to moving genomic healthcare forward was to try to resist labeling the concept of labeling I was wondering from a sociological point of view if you might have any advice as to how to resist what I think is a human weakness to label well it's an interesting conundrum in a variety of ways first of all we have to label people if we're going to use the model of using genetic susceptibility to tailor care so labeling is an inherent part of personalized medical care and I think it represents an example of the fact that labeling can sometimes be for a very powerful good and if you take that concept to a larger sort of stage I think what we say is labeling can be good or bad depending upon context so what we really want to do is use labels for the good which includes in the genomic model using genetic susceptibility testing when it has the potential for good and looking critically at that what goods are we accomplishing on the other side I think we have to be mindful of what you just referred to which is we do have the potential we know there's a societal potential to use labels adversely as well and we need to basically define that first that is we need to understand the ways in which labels can be used negatively in order to counteract or avoid them so I think first there's a basically a research process and I think for example the genomic LC program has as a very important concern understanding how labels can be used adversely sometimes the solution is going to be don't use the label there because there's more harm than good sometimes the solution is going to be let's recast how we're using this label let's try and either put protections in place or think as a society as to how we can use this information differently and some of this will be very complex and difficult to do often the labels actually are reflecting prejudices which already exist in the population and I always tell my students you don't need to use high-tech approaches in order to be discriminatory on that note in the session a little bit early we have a break and 3 o'clock return thank you welcome to the closing session of this remarkable symposium what we're going to talk about now are the implications of genomics for society and in many ways these are some of the images that have most caught the public eye we see people talking about the intersection of race genetics and we see white supremacist websites like Stormfront trying to use genomics to support their agendas we see on the one hand silly cartoons that suggest that we have genes that predispose us to talk about the weather to think we have stock market savvy to buy large SUVs and complain about stock prices we see movies like Gattaca that raise the nightmare of genetic determinism we see the important use of DNA technology both to convict into free suspected criminals but we also see people using genetic defenses to crimes increasingly this array of ways in which genomic information has really fed the public imagination often in quite hyperbolic ways is an incredibly important topic and I think the challenge, one of the major challenges for us is to separate the hype from the real biology and then once we understand the real biology to recognize that deciding what the implications of these technologies are depends importantly on us making informed political and ethical choices happily today we have we have a number of really wonderful speakers to help us try to understand some of these important conundrums and the first of these is Dr. Harold Freeman who is Associate Director of the National Cancer Institute and the Director of NCI Center to Reduce Cancer Health Disparities and he is going to speak to us today about race science and genomics Thank you very much for that introduction I'm going to take on a difficult subject in about 15 minutes so it means I'm going to have to move rather rapidly through this reflections on race science and society I think we could agree that race is perhaps the single most defining issue in the history of American society and let me tell you why let's take the issue that Columbus I say was discovered in America by the Native Americans in 1492 and here we have about 500 years of history before you I will not go into this history but it also includes a declaration of independence written by Thomas Jefferson who apparently had a hundred slaves himself and so that's a Jeffersonian dilemma to start with a civil war was fought slaves were emancipated Plessy versus Ferguson decision separate is equal Supreme Court decision finally Brown versus Board of Education and the civil rights movement so we're going through some very critical changes with respect to race historically in this country Charles Darwin apparently didn't think much of race I quote him in the Senate of Man 1871 he said the variability of all the characteristic differences between races cannot be of much importance he didn't know about the human genome or anything else or even Mendel's work at that time in 1997 the American Association of Physical Anthropologists indicated that races do not exist and the philosopher Sandra Harding put it this way from California she says race is not a thing it's a relationship between groups of people that reflects a cultural framework of societal institutional and civilizational values she indicates that this framework and cultures is within the sciences and affects scientific investigation how she goes on to say how we select the problems to be considered worthy of research how do we derive our concepts develop hypotheses and design all affected because scientists like everyone else are socialized before they become scientists and acculturated so we all of us have to look at what baggage we might bring on this issue to science here's a chart that has to do with the concept of racialization which occurred in over 500 years or so in our society there was a time when so-called white people didn't call themselves white for example back in 1400s 1500s when the Europeans came to this country saw people that were different and believed they should distinguish themselves they began to call themselves white and the others Native Americans and the slaves were black so it wasn't an innate classification even for white Americans who then called themselves English or Christian or some other designation in order to understand what has happened we have time today to go into this we have to understand human evolution we have to understand migration patterns population genetics genomics and the social history of America leading into the assumptions that we all make about each other as we look at each other through what I call the powerful lens of race and so racial classifications in America and elsewhere have not been determined according to genetics or genomics but have been determined according to society and politics the one drop rule a very fascinating manifestation of how people are classified originated in the 1600s agreed to by everyone by 1920 this rule which is still in effect as I speak here today in America defines as black any American has one black ancestor irrespective of how far back in time regardless of physical appearance so you could appear to be white and you could be black this rule is a very important rule to understand because when we talk about census categories now in America under directive number 15 we talk about black and white and other classes of people the black classification has to do with this rule hardly a scientific rule and so the issue then is the need to disentangle the social and political meaning of race which has developed over hundreds of years from what is often presumed to be a biological meaning of race which is a debate going on right now certainly genetic techniques that we talk about now with the great knowledge we have in genome and this is a wonderful progress these techniques could allow us to understand populations of people that are genetically different there's no question that there are such populations the issue that I'm bringing out here is though that race category does not help us to make this distinction yet those distinctions do need to be made there are diseases that occur in groups of people like Taysox disease circle cell disease, other diseases that we need to focus down on to help to identify the groups of people who have this problem but those groups of people that will come out in that analysis will not be equal to race clearly there is a manifestation of differences in population related to the parts of the world from which people have come over hundreds of years Africa, Asia and so forth this seems to be true but when you come to America and you define black people as any people who have one black ancestor certainly that's not the same phenomenon as geographical commonality recent findings by Feldman and others have shown that genetic variation occurs more within so called groups than between 5% genetic variation within groups only 5% between groups and there's more of a chance or at least an equal chance that there will be a genetic variation comparing one person to another within the same group as across groups also what has come up recently in debates particularly in the New England Journal of Medicine about two weeks ago between Neil Rich and Cooper has to do with defining that statistically within a group like white and black you may find statistical differences genetically but even finding that it's not an indication that race is the cause of that difference a very important point and so it is up to this point I believe that we need to determine the precise variables which determine the varying burdens of disease not go with crude categories like black and white but focus down on real categories that really make a difference just as we're focused down through tissues to cells to molecules we need to focus down on the populations and not use crude terms like race we should not assume that culturally or politically defined groups define development variables for disease but here's a way to look at this causes of disparities you could debate this but I believe that there are three major causes of disparities and they include low economic status lack of resources, poverty as one circle overlapped by a circle of culture indicating risk promoting lifestyle although culture can also be positive and overlapped again by a circle of social injustice injustice focused through years of injustice plus current social injustice that exists even today in a racial profiling manifestation if this is correct then I wish it to indicate that these circles may change over time and in the future or in the past the predominance of one circle over another that may be different I think this is a concept that should be considered it has nothing to do with genetics another way to look at this and that the chief driving force of disparities in my view is poverty itself causing negative human events as listed in these boxes which include inadequate physical and social environment poor housing lack of knowledge risk promoting lifestyle and diminished access to preventive health care decreased survival occurs culture I see as a prism through which poverty reflects itself and if that is true then culture has the opportunity to either accentuate or diminish poverty's expected negative effects there is a need to take into account racial classifications I think still today in order to address historical inequities from a civil rights perspective still we haven't gone to the point I think where we can say we can be color blind but to somewhat of a catch 22 here we have to measure call people races to measure injustice but races don't really exist we're in a bind here how can we get out of it it's going to be a tough thing to get out of it we need race to find injustice doesn't exist I catch 22 I believe that to retain race as a variable and I believe it should be treated as a variable instead of just a vague category we say sex is male or female we say age is 40 or 50 we say race we say nothing so let science come to the table and say if I use the category of race let me define what I mean by the category and then we can debate whether we agree with that so to retain race as a variable with a bearing on scientific research it must be recognized that race is a social construct and it's determined by how people see each other we also must clearly define when we use race what is being measured and for what purpose we must point out again there's no genetic basis for racial classification it's a social and political determination here's a good point particularly with the knowledge that we have today the thing we're celebrating the human genome double helix 50 years of science this is powerful science and I believe that the power of science can be used to eliminate false perceptions of the meaning of race and could prove pivotal in moving our society toward racial justice and reconciliation and this is something that I believe I'm looking forward for us to do to eliminate false perceptions and to create a concept of one race the human race and so to get to the conclusion I think that the scientific truth which we're all interested in must always be wedded to social justice should not be separated in our laboratories even those of us who do the most basic research have to stay connected to human beings and human benefit and not just say Eureka I found it let's go on to the next question I think one of the great purposes for science of people we need to think about that in our society I believe that we see value and behavior toward one another through a powerful lens of race and it causes us to profile looking either way and the profiling is very critical because it causes us to make assumptions about people without knowing who they really are as opposed to profiling and I guess I'll end with this slide the great scientist Albert Einstein showed that in this four dimensional time space continuum that we call the universe if two people are seeing things from a different position they will see things differently in other words what you see depends on where you stand our next speaker will be Dr. Maria Freer who is the chief executive officer of the global alliance for tuberculosis drug development which is a non for profit international organization that is working toward the development of faster acting and affordable anti tuberculosis drugs a need that is brought even more closely into focus given the talk this morning that demonstrates that TB is still an infectious disease killer in the world her topic is intellectual property a boon or a barrier to genomics good afternoon it is wonderful to be here after all of these marvelous talks on the human genome it is particularly wonderful to be here because in my new incarnation as the CEO of the global alliance for TB drug development I hope to use the power of genomics for targets for developing new tuberculosis drugs and if Peter's talk is of any inspiration to us just imagine the development and the complexity for a drug for which there is no market and so our previous speaker spoke about poverty and I hope our last speaker today will address genomics for global health I have however to speak to you about a topic that has been amazingly silent over the past day and a half for those of us who were involved very early on with the issue of intellectual property and the human genome it has been a wonderful ride when Harold Varmas recruited me to come and direct the office of technology transfer at NIH in 1995 I was three days in the job and I got this call and he says this is Francis Collins I'm calling you from a taxi cab I'm about to take a plane I need to talk to you about patenting genes and that's how it started and it was a marvelous and remarkable ride difficult at times but exciting what I'd like to do today is sort of remind you of the chronology of what has happened in the last few years with respect to intellectual property I want to talk a little bit about capitalizing on research and I will speak about the lessons I learned with this grand experiment that was the issues relating to the human genome so let's talk about the chronology and I want you to notice these arrows in the bottom if you can read them because they are they look the same size but the dates are quite different if we look at this first arrow we really realize that from 1953 all the way to 1979 there were amazing changes in biology and we've heard about this for the last day and a half but it was remarkably naive when it came to intellectual property in biology but there was one key event and that was the formation of course of Genentech the years that followed particularly 1980 were incredibly exciting and pivotal for those of us that look at the interface between intellectual property and science and one of the key things that happened in 1980 was of course the passage of the Baidol Act and I will not belabor the point here because I'll be talking about the Baidol Act and the implications of Baidol during my talk but I thought I'd highlight for you some of the things that happened during that decade and of course this list is by no means comprehensive but it does give you a sense of movement the Chakrabarty decision for the first time we were told by the Supreme Court that in fact if the hand of man intervenes in a biological environment in a mouse or in a cell that is patentable and as you can see some very critical patents came into being the famous Colin Boyer patent the co-transformation patent at Columbia PCR the insulin gene remember when it took tremendous amount of time energy and effort to clone and actually find the use of these genes by Genentech was considered a critical event and last but certainly not least technologies like Oncomas for those of us who have tried to make these technologies available we certainly remember these patents with great presence and not so much glee sometimes but in 1991 to 1995 the NIH files the first patents on ESTs and for those of you who may not remember Craig Venter was in fact a member of the NIH community at that time and these patents were filed it was a very controversial filing and when Harold started his directorship at NIH he decided that it was really against the public good to sequester this information and these patents for one particular organization and institution but clearly others follow suit and there were and still are filings for EST sequences in the patent and trademark office the flavor saver tomato and I will have another example of the golden rice in this chronology we're also examples of biotechnology in balancing with commercial applications I will not delve too much into the food part of the discussion I will focus on the biomedical aspects but certainly there was a revolution taking place and continues to take place to this day now if we go back look at what happened in the following years up to 2000 we will notice that the scientists for the first time realize that we were now at an interface intellectual property and the genes and the human genome came together in not necessarily a cohesive way and the first reaction to this duality came with the famous Bermuda principles a meeting that was held in Bermuda supported by the welcome trust in which scientists working in this field got together and said there has to be a better way we have to be able to deposit these sequences and make them available for all of us the good of the community is greater than the good of the individual I will go back again to these principles but it is important to also remind you that at that time the patent and trademark office put forth some very broad lines that in fact allowed for the patenting of gene sequences not only were those gene sequences allowed but if you had a portion of the sequence the language could allow you to have the entire sequence and the utility associated with this as we will see was not necessarily very well defined or had to be particularly well defined in the world we had GATT we had the tariff regulations with respect to intellectual property there was a change there was a complete sea change with the globalization coming to the fore and we also had the SNP consortium formation again we see action and reaction with respect to the human genome with respect to the discovery and with respect to the management of intellectual property more patents on more genes the ex vivo gene therapy patent that was awarded to NIH came into being the kinase patent that was awarded to INSIDE and the primates themselves patent all of which of course have their own histories and we can follow them through the years to come in 2000 and 2002 there was the Blair Clinton announcement and for those of you who need to be reminded the markets are labelled that was a very important morning and a very difficult time in which the markets plummeted because there was an erroneous assumption that there were going to be no more gene patents but a year later the first draft was put in place we had a wonderful celebration but in order for science to publish the sequence it was required that we signed a material transfer agreement to access that data an unprecedented move for the first time a scientific journal was asking people who wanted the data to sign an agreement that was traditionally used and seen for the transfer of intellectual property so here we are with patents on human stem cells the Cox II patents and we also start seeing although we've had litigation all through these 20 years we start seeing increased litigation and for those of you who follow the press you will know that there was litigation between the University of Rochester and Pharmacia and Pfizer and the Cox II patents so this year 2003 and beyond where are we there were modifications or enhancements to the Bermuda rules in the famous Fort Lauderdale meeting and of course we're here to celebrate the first edition as Francis calls it of the human genome but what next, where do we go next that's a question that I'm afraid I can't answer necessarily for you today but I will remind you of some of the principles that we had to grapple with when we came up with how it was we were going to manage the intellectual property I promise that I would tell you about Baidol Baidol allowed for the first time to capitalize on the research and to actually make money for universities from research supported by the Federal Government this led to a dramatic increase in the number of patents given to academia and it was seen by universities as a new source of revenue there was an added university and business symbiosis a symbiosis that by the way has served us very very well there were spinoffs from universities to companies Genentech of course was the predecessor the granddaddy of them all but we see them coming to the fore all the time there is personnel exchange scientific advisors for companies also have academic appointments in fact the majority of them do or are very closely related to licensing agreements between industry and academia all the time but as Peter mentioned earlier the time to get from a discovery to a product is enormous 12 years for pharmaceutical almost 20 years for a vaccine so what became extremely important is that the intellectual property itself the fact that you actually had ownership or title to the patent became it became a commodity we had no products but we had the intellectual property rights and intellectual property rights for the first time in biology were really seen as a source of wealth now there were some warning signals before the genome the patent of the genome ever came to the fore we saw the material transfer agreements for those scientists in the room that have tried to get materials from one laboratory to the other or one company to to their labs you will notice that it takes a tremendous amount of time energy and effort to sign the terms of engagement by which we were going to transfer one material to the other it came to the point where the experiment took less time to do than negotiating these agreements there were patents on research tools all of a sudden we were finding that in order to access a very basic tool we had to pay royalties to the owners of the tool and the the publication rights and luckily the academy and industry understood and realized that knowledge dissemination was a critical aspect for both and so even though now publication rights are protected in order to have people file appropriate intellectual property this is something that we have I believe successfully conquered we did see overreaching licenses the so-called reach through rights and we of course had to deal for the first time with very clear conflict of interest relating to these academic and industry relationships so when the genome came when we were talking about gene patents we had an unknown landscape there was a revolution on technology first of all and it has been mentioned before for the first time the technology was such it wasn't like sequencing the insulin gene that took tremendous amount of time energy and effort the automation was there the ability to generate massive amounts of data was there and the data did not match the utility so the quandary was if we allowed the data to go public we lost all patent rights if we had to hold on to our patent rights we didn't have an appropriate utility so we saw patent filings that were essentially for trivial uses so we knew the sequences were novel we knew the sequences were not obvious but how did we determine utility and that was really the quandary that was the critical issue how, at what point does this sequence become something that we support in terms of patenting so we started asking some very tough questions and it wasn't only in the United States by the way this was an international issue we wanted to know particularly in Europe the Europeans were very thoughtful and they wondered can you actually invent a gene the definition because of Chakrabarti in the United States was clear but the definition in Europe was not that clear did we really invent genes and therefore warrant patents or did we simply discover them how do you invent around a gene sequence because patents are meant for you to disclose the information so others can invent around them but if you have a sequence how do you do that who owns the gene of course we've heard that many many times and how does science continue unencumbered if somebody already if you go and you put your flag and you claim the territory for all of, for you how does the rest of society work at what point are you working on a sequence that somebody else has intellectual property rights for which you had no knowledge that they existed and at the end of the day of course it's ethical and we will have other speakers talk on this issue so there was a unique and remarkable response the people working in the human genome project throughout the world came together as I mentioned before and put forth these Bermuda principles these principles have been truly remarkable and they have changed the landscape of how science and academia and business interact government and funding agencies also became very aware and came forth with policies and practices to ensure that when they funded the research that the research would be available to all and I think we have to give a lot of credit here to groups like the MRC and the Wellcome Trust in the UK NIH for the first time implements policies on access and we implement the famous research tools guidelines and for those of you who haven't read them I will humbly say that I consider that to be one of the key accomplishments of my time at NIH it has been used and these guidelines are used quite often not only by academics but also by people that are NIH funded but I'm proud to say by people from industry and academia all over the world and certainly the public pressure did not go unheeded with the USPTO and they did in fact change their guidelines and tighten the guidelines so when the draft came in in 2001 we know the results I mentioned them earlier the sequence was made public on the web from the international public consortium the Solera information was deemed to be so important that Don Kennedy agreed to have the material transfer agreement put forth before they could be accessed so what are the lessons learned when I look back at my tenure at NIH I think what was it that really made the the whole thought process work we had two core principles one was supportive of the academic mission two was the understanding and comprehension of the balance between the public and commercial benefit and that both were important we clearly had to support academic freedom it was critical to have and continue to have publication rights and any deal we cut and any patent we put forth and any agreement to which we held ourselves had to be supportive of the research enterprise of the educational enterprise and to ensure that the right people got the credit if we allowed our students to work on technologies that were patented or protected could they publish a thesis that was a quandary that we could not and would not accept if we were going to continue to move the scientific enterprise forward with respect to the balance of commercial and public benefits we understood that patents are important as strategic tools but that appropriate patenting goes hand in glove with that use of intellectual property and licensing as well it's not only the patent itself but it's what you do with the patent and how you utilize those intellectual property rights it's one thing to have a baseball bat and you can galvanize a whole stadium by hitting a homerun but you can also use that baseball bat to hit somebody and inflict damage that's sort of the way you can use patents so strategic licensing and your intellectual property has become and will continue to be very critical to ensure that this public and private balance is maintained. Access of research tools again is something that we share scientists everywhere know and understand that if they do not have access to the most basic and important technology so that they can do their research research is encumbered and certainly we have to facilitate the transfer. With respect to patenting we consider them to be critical elements for further research and development but we believe that the scope of the patents should be commensurate with the invention itself in an academic setting blocking patents or defensive patents or stealth patents which are sometimes used perhaps legitimately as strategic moves in industry are not the way really to allow the process and for those of us who have to deal with transferring technology sometimes patents are not necessary for transfer technologies that are ripe to be put in use already. In terms of gene patents full length sequence of patents of genes with known utility are fairly well accepted there has been little support for patents on ESTs or partial sequences and there certainly is understanding that the breadth should be commensurate with the inventive exercise. So finally the points to consider. I think the human genome and the patents on genes have provided us with a unique laboratory on how to handle intellectual property and will continue to provide us with a unique laboratory of how to handle intellectual property in the context of business and in the context of academia. Vigilance is critical and one of the most important players in this whole arena were the scientists themselves the scientists who decided to auto-regulate and to know how to get the whole better that the whole was more important than the individuals. We also have importance on flexibility the symbiosis between the academy of industry as I spoke about but the last point I want to make is that of course we have to stick to the principles these are critical and these will guide us there will be many many more intellectual property battles to be fought as all these interactions that were spoken about before come to the fore. Let me acknowledge just a handful of people that in this historic event I won't be recognized. Jack Spiegel at OTT and Barbara McGarry John Stewart from the Worldcome Trust Rachel Levinson from the Noel de Noire from the European Commission that put forth on her shoulders all of the discussion of the European with respect to gene patents. Thank you very much. Our next speaker is Paul Miller. He is a graduate of the Harvard Law School and currently a commissioner of the Equal Employment Opportunity Commission and he's going to speak with us today about the incredibly important issue of discrimination and employment. We've already heard from comments from the audience today that this is a major topic of public concern. This is just so cool. I feel like Captain Kirk up here. It is a true pleasure and honor to be I'm just going to play with this for my ten minutes. It's really a pleasure and an honor and a privilege to to be up here at this symposium. It is such a remarkable couple of days and to have been a part of it in some small way is really a tremendous highlight for me. We talked a lot about over the past couple of days about the discovery of the structure of the double helix and I think that the profound developments that have taken place since the discovery of the structure of the double helix 50 years ago are really quite staggering and more staggering still as we've learned are the potential benefits and the boundless horizons and promised and unimagined applications for the science of genetics that are still yet to come and yet as with most technological advances there are significant risks for misuse and abuse and history has taught us that the wonders of science hold unique power to sway and seduce but also too often to corrupt and to corrupt the course of human nature and I'm concerned about those who might be marginalized by such scientific progress. Now with apologies to Aldous Huxley we stand at the precipice of a brave new world and sometimes though I also worry that we will find ourselves in George Orwell's world as well and the challenge for all of us scientists, ethicists, lawyers and policy makers is how best to balance the rights of the individual against the advance of science and technology in this rapidly changing world. Now scientific advances never exist in a vacuum and must always be viewed through the social, ethical legal and political prism. In fact, not too long ago Darwin's revolutionary theories on natural selection and the evolution of species that sparked so many wonderful scientific advances and led to a greater understanding of mankind's place in the natural world was dreadfully misapplied through the prism of bigoted notions of the underlying causes of class, social and biological differences bringing us the legitimized atrocities of the eugenics movement. Science can be easily corrupted, experts are often wrong and governments sometimes traffic and cruelty and ignorance and what is interesting to me is that whether or not the scientific community is mobilized to derive something as junk science or voodoo genetics may ultimately not matter. Sometimes public opinion and market forces prevail regardless of whether something is scientifically rational or not. As the science of genetics explodes and the technology becomes more accessible the issue of how society protects its workers from the misuse of genetic information will become more important and while I certainly applaud the emergence of new genetic technologies my fear and a fear that is shared by others is that employers will attempt to exclude qualified workers from employment due to real or perceived genetic predispositions or their carrier status. Now this concern applies to those who either have a predisposition to or who are carriers of markers for both rare and common disorders alike. I think that those that the particular risk of discrimination is higher for people with genetic disabilities because of the stigma attached to such disorders. Genetic discrimination in my mind concerns an employer taking an adverse employment action based upon an asymptomatic genetic predisposition to a disease or a medical condition. What we're talking about is the increasing ability due to advances in genetics of predicting who may become ill in the future. Most genetic markers cannot predict that an individual will in fact get sick in most instances only that there is a greater likelihood that he or she will actually fall ill. I believe that to exclude qualified employees or applicants from employment opportunities due to such a genetic marker indicating a potential medical condition constitutes illegal discrimination. Now it's difficult to know precisely how prevalent the use of genetic testing or genetic discrimination in the workplace is today. There have been studies that have provided anecdotal evidence of instances of genetic information and yet in my mind even if such evidence does not point to a widespread problem today, I believe that as the cost of genetic testing decreases and the practice becomes more commonplace, the potential for real discrimination will dramatically increase. Moreover, the mere fear of discrimination may cause people to become reluctant to take advantage of the growing array of genetic tests that can identify vulnerability to specific diseases and may in fact hamper one's willingness to get involved in being genetically tested for research purposes. Genetic discrimination though I think it's important is a new application of an old violation of law. The entire body of American workplace anti-discrimination law is based upon the premise that applicants and employees must be selected and evaluated based upon their ability to do the job and not based upon the myths or fears or stereotypes made about that person due to his or her race or gender or age or religion or disability. Society faces the question of whether employers should be able to consider genetic predisposition information in making employment decisions and if so how the law should protect workers from the misuse or the potential misuse of such information. While some, including my agency, the EEOC have argued that the Americans with Disabilities Act, the law which prohibits discrimination on the base of disability protects people from employment discrimination based upon one's genetic predispositions. There are those who are concerned that the courts will find that the ADA does not cover genetic predisposition discrimination. Others believe that genetic discrimination is so different from traditional disability discrimination that the ADA does not provide a satisfactory framework and no court has yet ruled on this issue. Nonetheless, the principle of genetic non-discrimination in employment enjoys wide bipartisan support. Legislation has been introduced in both Houses of Congress by both Republican and Democratic members which specifically prohibit discrimination based upon genetic information by employers. Moreover, President Bush has expressed his support for the principle of genetic non-discrimination. However, as those of you who travel in Washington know, the devil is in the details. And as of yet, the Republican and Democratic sides have not gotten together and reached an agreement as to how to craft and how to draft such a genetic discrimination bill. However, I am hopeful, I am still hopeful that Congress may pass a new genetic non-discrimination law, in fact, this term. And I also want to note that towards the end of his term President Clinton signed an executive order which banned the use of genetic information and employment decisions in the federal government. Now, the first and only lawsuit alleging genetic employment discrimination was filed by the EEOC against a company called Burlington Northern and Santa Fe Railroad. And because it was settled before a trial on the merits, no court ruled on the applicability of our theory. And the facts of the case are simple and it got a lot of press and you might have heard of it. And the facts of the case are such, the EEOC alleged that Burlington Northern or BNSF, the railroad subjected its employees to surreptitious genetic testing. BNSF was testing to identify a genetic marker for carpal tunnel syndrome to address the high incidence of repetitive stress injuries among its employees. The EEOC further alleged that at least one employee was threatened with discipline and possible termination for refusing to take the genetic test once it was discovered. Now, shortly after filing the case, the EEOC and Burlington Northern announced a mediated settlement in the amount of $2.2 million. The railroad also agreed to halt any genetic testing that may or may not have been going on at the company. While the result was overwhelmingly positive in that the EEOC achieved everything that it sought in its lawsuit, by resolving the lawsuit informally, which was the right thing to do, it was never necessary for a court to rule formally on our legal theory in the case. But what was most particularly reassuring about this Burlington Northern case was that no one, not the business groups, not the employer groups, not the scientists, not the press, not the politicians, not even the talking heads on Fox News Channel, nobody thought that surreptitious genetic testing of employees and taking adverse actions against those who had the wrong genetic marker should be allowed. In closing, let me say that with advances in genetic technology, we will soon realize that everyone has genetic predispositions for one genetic condition or another. Mapping the human genome changes the way we understand who is quote normal and who is quote disabled. If we all have genetic misspellings, how do we now define who is healthy and who is not? If we all have genetic conditions that are just waiting to express themselves sometimes in the future, aren't we all in some ways truly disabled? As we will have the knowledge of the potential genetic disorder that each of us harbors, disabled people may no longer remain stigmatized as the other within society. Thank you very much. Our final speaker for this session is Dr. Robert Schaler, who is the director of forensic biology at the office of the chief medical examiner of the city of New York, which is the largest public forensic DNA laboratory in the U.S. today about forensic DNA testing and human identification. Thank you very much. I have to say that it's a privilege and an honor to be able to take part in the celebration. I had to think back when the first time I had a chance to take a look at DNA, and that was in June of 1964, my first day of graduate school and it was quite exciting, I have to say. And so I spent the next four years working with DNA in graduate school and I didn't go back to it again until 1986 when forensic science began to think about DNA as something that might be important. The world of forensic science the world of forensic sciences is clearly a world that's a little different than we've been talking about the last few days. Forensic DNA analysis had its origins about the same time as the discussions about sequencing the human genome were taking place. I didn't realize that until yesterday but it dawned on me that that's about true. Alec Jeffries in the mid 1980s coined the term DNA fingerprinting and so to speak with respect to law enforcement that's the starting point and once the prosecutors and the police jumped on board it was up to the scientists to keep the ball rolling so to speak and we had a tumultuous time of it as well. We had a lot of court decisions. Eric Lander we put another hat on his head from yesterday was one of the defense experts and one of the first famous cases the Castro case in New York City when he was a forensic expert and now he's also forensic scientist as well as a well-known geneticist and genome person and I was talking to him last night about that and we had a little bit of a chuckle over it because I was on the other side of the fence. The Castro case was an important one and Judge Jerry Shindlin whom you probably have seen on TV is Judge Jerry his wife is Judge Judy was the presiding judge and he came up with an interesting decision which basically said the work that was done by Life Codes Corporation could not be used to include anybody but it could be used to exclude someone and what he also said was that there was a science of DNA technology forensic DNA technology that could be used to make identifications and that was critical. In New York a few weeks ago at the 50th celebration Dr. Watson made a statement that one of the things he's most proud of is the application of DNA to law enforcement and although Dr. Watson is not here today I would like to say that this is for him. Law enforcement with respect to DNA testing we're involved in rapes and homicides and burglaries and property crimes as you know we're involved in mass fatality identification the victim identification and the re-association remains and I'm going to talk a little bit more specifically about these in a few minutes and then of course we're also concerned with identification from parentage testing and more specifically I'm more concerned with the identification of rapes as a result of identifying from the products of conception and in those instances the husband is not really a husband he's in a salient. Dr. Charles Hirsch who's the chief medical examiner for the city of New York described forensic DNA testing as the single most important law enforcement tool in the 20th century and I think he's correct and I think from all the newspaper hype and everything that you hear about it on television I don't think there's in any law enforcement tool that you've heard more about on March and in May in 1994 two 15 year old girls were raped September 11th 1997 Joe Halis Castro who was 19 was raped murdered and set on fire on a rooftop in April 19th the 13 year old girl was coming home from the skating rink and she was attacked at knife point and raped and her rape has said be quiet and take it like a woman on June 2nd 1998 Rashida Washington was 18 years old she was raped sodomized strangled and dumped in a stairwell on September 25th 1998 a 15 year old girl was abducted at knife point and her rape has said you're lucky to be raped by such a good looking guy September 16th 1998 a 14 year old girl was raped and sodomized and he said to her act like you love me and then we go all the way back to 1991 in May Jonathan Hayes was conducting an autopsy of a young woman who was 13 years old who had been stabbed in the left breast strangled and dumped on the street the police had no knowledge that any of these cases were connected they just didn't know they had some theories that several of them were linked together by at least and they thought maybe there was as many as three people committing these crimes in the April 19th 1998 case there was a composite drawing of the person who attacked her and the murder of June 1998 there was a sweatshirt that was found at the scene this is the CSI stuff by the way and Jonathan Hayes removed a pubicare from the vagina of Payola Ilarra who was 13 years old but still the police had no idea who committed these crimes and again still believing that there was more than one person on the sweatshirt there was a laundry tag and the police followed the trail of the laundry tag because it went back to one particular laundry and although there was a long list of people who had the tag they were able to identify the mother or a woman who had that laundry tag about a half an hour before they about the same time as they got the information on the laundry tag there was an anonymous tip from the composite drawing where a woman said that there was a guy who lived in her apartment building who matched the assailants description and the laundry tag belonged came from a woman named Key who lived in that same apartment building and the guy who lived in the apartment building had the name Ace and he was in apartment 1910 it's a good looking guy this statement comes from the lady in my laboratory who testified against him I think something happened here but he was non-threatening appearing and it was someone your daughter might date so we have all of these rapes and we have suspicions that this one person was at least involved in one perhaps two of them but science has to come to the rescue as you know the murder of in 1998 the June 2nd 1998 murder and the September 28th, 25th 1998 murder rape case were linked by DNA testing and so that was the first clue that other cases were involved in these cases what is interesting is in the murder case three suspects and a boyfriend were eliminated the suspects were under intense investigation by the police in the April 19th 1998 case two suspects were eliminated by DNA testing in the September 25th 1998 case a man by the name of Daniel Simmons was two weeks away from going to trial for that rape and DNA eliminated him the prosecutors were so convinced that if he didn't do it someone in his family must have that they had four of his brothers tested and we were using YSTR testing to eliminate them the Johalis Castro who was the 1997 murder victim had not yet been identified her body had been burned beyond recognition but there were suspicions of who she might be and so she was identified by doing parents testing by DNA testing on the March 1994 case the rape kit had been discarded by this time there were suspicions that maybe this one guy had done all of these and so we were going back and looking at all of the evidence trying to piece it all together but the rape kit had been discarded but the hospital had made a slide that had the sperm on it so we were able to get a DNA type from the sperm there was no evidence in the 1991 case except the vaginal pubicare that was removed by Jonathan Hayes in the New York City Medical Examiner's office and this was eventually subjected to mitochondrial DNA sequencing we still had to prove the case we had the guy we thought we had the guy but we had to get his DNA to make the comparisons and this is a smart guy he didn't want us to test his DNA so the police at one point the police had actually tried to trick him into giving a blood sample for TB testing and he said he's a Jehovah's Witness but he can't give his blood and he refused to do it so they gave him a cup of coffee and the cup of coffee came back and it wasn't his it was quite a surprise turns out that he had exchanged the coffee the coffee matched the rape pattern and the murder pattern but the coffee cup came from the cell mark because this Aaron Key was the guy's name and he had exchanged the coffee cup with his roommate so the police had to go and get other other examples to try to get his DNA in another way so they ended up going more cups of coffee they finally gave a cup of coffee to the cell mate and they knew it was his cup of coffee and so we had to match his DNA to that which means that Aaron Key's DNA was on the other coffee cup one DNA series matched the coffee vendor that they got the coffee from which means he was probably taking sips out of cups of coffee that he was giving out to people during the trial he was convicted he got 200 plus years in prison so he'll be in jail for the rest of his life and during the prosecutor's summation Aaron Key said he didn't want to listen to what the prosecutor had to say about him and the judge told him to be quiet and take it like a man now there's no DNA index at this time which we call CODIS combined DNA indexing system which is established by the FBI and it's a nationwide system right now two of those murders wouldn't have happened and three of those rapes would not have happened because Aaron Key had been put in jail in 1992 and his DNA would have been taken and so we would have known right away who it came from but the system works pretty well my laboratory were uploading cases twice a week the system is something like this if we take a look at New York State for example New York City is considered a local database we upload our DNA profiles from crime scenes to Albany which is the state system it's called the S-DIS and then from there it goes to Washington DC and so we can search locally within the city itself we can search statewide and we can search nationally and we've had hits nationally as well as internationally and we're identifying people behind bars just as importantly we're identifying people who do not commit crimes and we're able to release them as I showed in the example of the Aaron Key case so we were happily going along at this point in time doing rape cases and doing homicide cases doing some crime scene work and I was feeling very much like a forensic scientist and then this happened in case you didn't know the chief medical examiner in any jurisdiction has the responsibility of making the identifications in any mass disaster whether it's an airplane crash or a fire or in this instance where two airplanes crash into two buildings and the buildings fall down this perhaps is the most complex use of DNA for human identification that I know of we had 2,792 missing people and about 19,930 remains and some of these have gone up because we had to go back to the remains and try to find where we have co-mingling and we have been able to separate out a number of those we originally thought were from one person but now we know they're from more than one person at that time it was pretty chaotic as you can well imagine we had to reorganize the laboratory and because the chief medical examiner is responsible for making identifications that put me in responsible for making the identifications using DNA so we had to reorganize the laboratory in fact we had to reorganize the way we think about doing business and a number of things came to light pretty early one of we had no idea how many samples we were going to be getting but we certainly knew that we weren't going to be able to do all the work ourselves so we began to think about how we could get other people to help us to do this work the New York state police came on board on September 14th and we split the responsibilities of them they took the DNA responsibilities for the family samples and we took the DNA responsibilities for the samples coming from the World Trade Center site on September 12th I got a phone call from Craig Venter and he offered the services to do mitochondrial sequencing you have to understand we were looking for ways to do a lot of DNA analysis in a short of time as possible we were looking for high throughput capabilities the New York state police had an existing contract with myriad and myriad had a high throughput capability of doing STR analysis which is the typical forensic test that is used for the CODIS system as well as all forensic laboratories pretty much throughout the world and so this became the way we set things up originally we had a lot of help we had a help from the national institute of health and I thank everyone here who was I know people here in the audience who have helped us tremendous help the national institute of justice set up a committee to help us convene 30 scientists and these 30 scientists were able to help us in various areas we needed help in certain aspects initially and one aspect we needed was we had no way to make identifications we didn't have the software tools to do it number one and number two if you can make an identification using STR the paradigm at that time was 13 loci makes an identification what if you don't have 13 loci what are your statistical barriers so you know that typically what happened to those matching anti-mortem DNA profiles with post-mortem DNA profiles gives you a match and then you have your kinship analysis which is basically a paternity test and then of course you have to make the notifications so the families can have closure the preset statistical barriers were set up by the NIJ committee which is called the kinship and data analysis panel which is affectionately known as the KADAP we had a number of challenges I don't know why those pictures don't stay up there but what happened was we were getting all things and all kinds of samples preserved samples, baked samples decomposed samples, bones, bone splinters mummified samples samples which looked like they had just been buried for a short period of time chunks, small pieces and what faced us most was the fact that degradation was occurring we had the possibility that this process would take over a year to recover the samples it actually took about nine months but as you can well imagine human tissue doesn't like to stick around in the normal environment for that long and so by the time the end of September came around we realized that at least 25% of the samples were giving no DNA test results at all and another 25% were giving partial DNA test results so we had to figure out a way to deal with decomposing tissue and what was happening was this we were having PCR reaction failure for a number of reasons we had a lot of co-mingling remains so a lot of our samples were mixed samples because of degradation and the extent of degradation we were getting limited target DNA so we were getting stochastic sampling and we were getting strange effects the typical DNA degradation looks something like this where you begin to lose signal eventually you lose low size sometimes you lose alleles and sometimes you lose alleles that they're not in a very predictable way the typical mass disaster testing paradigm is do the typical forensic STR test maybe go back and do some reanalysis and do selective mitochondrial DNA testing and this was done in the American Airlines Flight 587 which fell in New York at the same time we were doing the World Trade Center and so we did that TWA, Flight 800 Swiss Air and Alaska Air we changed the world for this process because of the World Trade Center we did the forensic STR testing we had really aggressive, we had to re-do re-learn how to extract DNA from degrading tissue we set up mini-plex STR reactions we did massive mitochondrial DNA testing and we're still in the process of doing SNP testing I'm going to give you an example this is the number of DNA identifications we've made you can see how they've fallen off over time but we're still doing it we've created samples something like this this is the original testing on the bones and this is not working basically what you're seeing here is a sequence of using mini-STR mini-plexes where we've reduced the primers so that we get a smaller amplicon size and we're able to fill in the blanks of some missing loci that way and then we can match it up to a toothbrush which is the top line and you can see by making a virtual profile from all the original testing we're able to show that the STR profiles are the same and can therefore make the identification and the statistics fall within the guidelines of the cadap which basically, like I said, helped us out a lot we're also using mitochondrial sequencing to help us sort through the large numbers of samples that have bad data and we can choose a name basically every person who died has an RM number which is a reported missing number we can select a disaster sample we can look at the DNA, the mitochondrial sequencing on that sample and then we can check the STR see whether or not we have a match and you can see here we have a potential match but we're missing a lot of alleles we can also check the against the personal effect with the original electrophereograms to see whether or not we have alleles that should have been called that weren't called because of some threshold barriers and we can also look at the mitochondrial sequencing and you can see here that we're missing a number of positions where calls were made so we have to go back and try to re-sequence this anyways, this last example is one which is in progress we have not made the identification we believe we know who it is but this requires more analysis and probably another year before we're able to get it all completed and I thank you for your time and I think that's it I don't think I have time to talk about SNPs I can just show you, we kind of look at SNPs in a little different way I will that's it, thank you don't go down I'd like to invite our speakers up here for questions and answers from the audience with the rich amount of material that's been discussed in the last hour plus that we'll have plenty of questions I do regret that we're only going to have 14 minutes left for questions so well, so please keep your questions brief do you have a question? Please I wanted to ask Paul Miller examples are often given in this discussion about the workplace where testing for predisposition may be a benefit, the example that's often used is if a particular gene predisposes for sudden cardiac arrest wouldn't a airline want to test its pilots for such a test, also the Department of Energy currently has experiment going on in testing for those who might have a predisposition working with beryllium for beryllium of this disease how do you handle in your context those kinds of situations where there might be a benefit where predisposition would show that's a really important question an important issue I think that is important is first that the individual worker have the flexibility that becomes the worker's choice and that their their decision forced and consent whether to be tested or not I think that people have the right to engage in sort of activities that may not be healthy for them being a cop for example may not be a particularly non-threatening job but people need to do that I think that where we get into trouble is to use predisposition information that is some likelihood that they may get sick as an opportunity or to screen people out of jobs that they're qualified for in the law we have something called the direct threat standard and that is if somebody poses a direct threat to others or to themselves now under the law that an employer may have an argument not to not to have to hire them and I think that that direct threat standard will probably apply here in the instance of genetic predisposition testing but what I think is really important about genetic predisposition testing is that in most instances genetic predisposition information doesn't tell somebody that they are going to get sick, that they are going to have a heart attack just that they may have a higher likelihood from other people from the general population in having a heart attack and I don't think that that is should be enough to exclude people from opportunities in employment the gentleman up in the balcony has a question for Dr. Schaller DNA evidence was subjected to a great deal of scientific validation before it was fully accepted by the courts some people have been arguing recently that other forms of forensic evidence should have similar kinds of scientific examination and validation including fingerprints and I'm wondering what your reaction is to that proposal and why I agree I think that you know we take fingerprints for granted because they've been around for so long but nobody's really ever done an exhaustive study by the same token I don't think we have any indications that fingerprint analysis if done correctly gives wrong identifications but I think what we're dealing with is a situation in which we have people doing fingerprint analysis and I do know that a lot of fingerprint analyses are challenged and some of these have been found to be errant simply because the examiner either didn't utilize enough points of comparison to make come to a conclusion or maybe the field itself needs more clarification as to what constitutes an identification with safest semi-automated fingerprint analysis there's imaging which matches things up and that does a database type of function but then a person takes the matches that are brought to the top and then does a comparison with them so once again you're back in the realm of a person doing that kind of comparison and by and large these people are not scientists yet. Yes ma'am. Hi, Leah Hardy from Pfizer. I have a question for Dr. Freeman. First I'd like to thank you and all the speakers for very thoughtful and thought-provoking presentations. My question is this. As you describe race as a very crude variable for categorizing people for scientific study and one alternative that's been proposed to self-describe race is the use of genetic markers that are exclusively or at very high frequencies in certain sub-populations and in light of your comments and given that the world is increasingly becoming a smaller place and people are increasingly mingling with each other and populations are becoming less distinct I'd be interested in your thoughts about the use of these markers in scientific research. A good question. I think that the issue of using markers in scientific research for outcomes that are designed to help people is perfectly perfectly good. The danger is though I think in defining who the populations really are from a genetic perspective and the danger also is been brought out in my colleagues concerned about if you do find something what does it mean with respect to the rights of that person. I believe that we should look down deeper into populations from crude levels of understanding now to refine our understanding genetically of populations of people who are subject to increase the risk for certain problems and there should be nothing to stop us from doing that but it has to be done with the sensitivity that I mentioned and in particular I don't believe there's any evidence that if you do that analysis that the groups that you will finally discover with respect to some particular measure that you're trying to understand say from a medical point of view those categories will be much different from what we call race. Yes. Hello, today Hamilton University Pennsylvania School of Ethics. This question is for Dr. Schaler. I was wondering if you could speak to how the FBI's DNA database impacts racial profiling. Have there been conversations in the law enforcement community on if there's a positive or negative impact on racial profiling? The National DNA Index CODIS utilizes markers which specifically do not affect racial profiling and so when samples are put into the National DNA Index there's no indication of race at all so it doesn't impact it at all. I guess I'm speaking more from a societal perspective in the black community you often hear conversations that maybe African-American men are targeted more often and some of this is hype but some of it is also reality so I guess more from a societal perspective have there been conversations with relationships with African-American men or women, African-Americans in general in the law enforcement community. Have you been privy to those conversations or do you know if those conversations are going on in the law enforcement community? I personally have not been privy to those conversations but I do know that there are efforts out there by scientists, forensic scientists to try to get more information from the human genome that we're using right now. People will be talking about. Up in the balcony. My name is Eric Ornman from Applied Biosystems I had a question for Dr. Scheer. I'm most curious about with all the successes of CODIS they're still quite small compared to the successes in the UK in terms of number of cold hits per day or per week. Can you comment about the likelihood of the US getting more serious about forensics the CODIS is rapidly expanding right now the limiting number of profiles that is keeping us back from getting more hits are the number of crime scene samples that are in the index we have a lot of convicted felons who have their DNA profiles in the index right now thousands and thousands of them tens of thousands of them we have relatively few crime scene samples to compare them against and so that's really what limits you in my laboratory we upload several hundred profiles a month and we're getting approximately a hundred hits a month and we have a relatively small number of samples in there so I think you'll see as we get tens of thousands of crime scene samples in there you're going to see some dramatic numbers Hi, I'm Will Fitzhugh from Solera the question is about patents I was wondering I wanted to make sure I understood what you said about EST patents they weren't enforceable or weren't invalid is that what you said and the follow up question is then if there's quite a continuum between ESTs and complete accurate gene sequences so what are the standards for enforcing patents with slight errors and polymorphisms and things like that Yes, I was talking historically on the filing of the EST patents and the guidelines of the PTO as you all know them as well as I do I went into the news this morning and one of the anchorman said I'm sure that after the break I'll have something thoughtful to tell you so stick around let me think about what I need to say about that and that's sort of the reply because you're absolutely right there is a continuum there and the utility of course keeps getting is something that needs to be associated with it it is an interesting observation and I think for those who are interested in this field we will see how these patents eventually are issued and how the guidelines are implemented so right now you and I can speculate I think all we want but the proof is going to be in the pudding and then it's going to be more important as to what the courts are going to say I hope you all will join me in thanking these wonderful presenters this afternoon our final presentation today I think is really fitting today's investigators from around the world as we've talked about the development of the human genome sequence and we've heard today about the importance of thinking about genetic variations that contributes to disease around the world both from Doctors Hill and Dr. Freer this afternoon and so it is fitting that our final speaker is Sir David Weatherall the Regius Professor of Medicine Emeritus from the University of Oxford University who is going to speak to us now about genomics and world health Well thanks very much indeed for inviting me to what's been an enormously exciting few days I'm a little bit nervous about speaking as the last speaker I've had one or two bad experiences over the years when I was a very young man I was in a position like this and the hall emptied except for a group of old ladies and I thought well good pro as I am I better give them their talk and after about ten minutes two of them got up and I could see they were carrying buckets and I realised that I was lecturing to a group of people who were in there to clean the hall so if you're feel like getting up just keep your buckets on there quietly I'm also archaic I hate anything modern and so I'm using good old fashioned slides and the two little men at the back I hope who are going to put them on for me could I see the first slide please I think the reason I was asked to give this talk was because for reasons best known to herself the director of the general of the world health organisation asked me years back if I would be the lead writer that's the person who does all the work on a report and genomics and the world health and this has been published now in a fairly massive report so I strongly advise you not to read it but you can actually buy the summary and I found it quite honestly one of the most difficult things I've ever done although I've actually worked quite a lot in the banking world and we have a big group of people in Oxford who do that and the reasons that it's so difficult I think are shown on the next slide this is a picture of the state of disease really in 1990 or 1992 and shows disease in what are called Dali's disability adjusted life years I think health demographers are not terribly interested in whether we're dead or alive but whether we're the nuisance we are when we're alive but it gives you a very good idea of the I think it gives you a good idea of the frequency of death and disease in the community and the red shaded areas are the diseases of sub-Saharan Africa malaria diarrheal illness perinatal deaths HIV and the sad thing is that same diagram compared with the disease pattern in the established market economies because to be very similar right across the richer world hasn't really changed since the 1960s or 1970s it's changed a little and it certainly hasn't changed over the last 10 years HIV would have come shooting over tuberculosis would certainly be back in this list probably the most frightening statistic and the kind of thing I think any of us who have been in clinical practice would go to bed every night with a bad conscience about the figures from 1998 of infection and the analysis of deaths in childhood from infection of about 12 million deaths indicated that about 5 million of those were due to diseases where we have vaccines or easy methods of prevention so when you actually go into the world public health and against this frightful situation of gross poverty and dysfunctional health care systems and you look at the kind of things we've been talking about for the last 2 days they seem almost irrelevant and that was really the problem that I had if we look at the next slide the other problem is that if you're really kind of honest with yourself and you ask what has been the impact of genomics on world health so far the answer is it's been very limited it undoubtedly has altered practice in single gene disorders and in clinical genetics but in communicable disease malignant disease it's given us huge insights into mechanisms, quite extraordinary and to some degree in the other common diseases of western society but if you took day-to-day clinical practice because these diseases make up a very small part it's not yet fulfilled its promises so it's against that background that we have to kind of think now should we what should we be doing about thinking about the future for not just our own countries but the developing countries and I'd like to take just two issues really this evening can we see the next slide please one is the approach to the analysis of genetic disease and also the second point is can we look at genomics and ask the question is some of it already ready for transportation and can that act as a kind of future model if we look at the approaches to determine of the genetic component of disease we've talked over the last two days about two major approaches the kind of genome down the kind of genome to phenotype approach which has been so successful in the diseases where we knew absolutely nothing about the kind of cause and path of physiology then there's the genome up approach the phenotype down approach which seems to me to be of so critical importance and I think so many of our younger people coming into this field now don't really realize the importance and I'd like to show that by one or two examples particularly in my own field and then the other approach which again we've neglected in many directions we've looked at twins for common disease but we're now about to spend hundreds of thousands maybe millions of dollars analyzing monogenic disease for modifiers where we haven't got one good twin study and we've totally neglected the kind of classical approach of epidemiology the analysis of genetic disease in a changing environment so let's ask the question can we look at an up-down approach and can we apply any of these approaches as models in the developing countries we'll see the next slide I'm going to just briefly remind you about the hemoglobin disorders because they're monogenic diseases which are now completely forgotten in the kind of western genetics but they are the commonest ones in the world population and the WHO reckons about 7% of the world population are carriers they're either genetic diseases of the synthesis of hemoglobin or of the structure of hemoglobin and I'd just like to focus on these two here at the alpha and beta thalassemia just for one minute if we could see the next slide this is the world distribution of these diseases and these are the estimated births of new cases each year and all this data is completely wrong it's inaccurate what is fault particularly is WHO data but what we know now is that if you go into these countries and you micro-map for these disorders you find that the distribution is grossly uneven right across all these regions so these bland figures from one center or two centers don't give any idea of the frequency we've recently been micro-mapping this whole region with our Indonesian colleagues and the heterogeneity is extraordinary but the overall data suggests that we tell the Indonesian government that to treat these children over the next 20 years they will require about 1.8 million bottles of bloody year and currently in Thailand it's estimated about nearly 3 quarters of a million of these kids so that's a monogenic disease in a setting of a developing country now the central question and problem for this disease and like all monogenic diseases is the following if we could see the next slide this is a child who's got the very worst clinical pattern of beta thalassemia he was born well profoundly anemic within six months in a country where they didn't have enough blood to give him to keep his blood up and frightful deformities of the bones, numerable fractures his hands over his painful liver he's got a huge painful spleen under this and it's a kind of living death it's not that bad if he's adequately treated and this child is homozygous for beta thalassemia and so is the next child who comes from a few miles away and both these children are homozygous for the same mutation and it's this level of heterogeneity of all monogenic disease which again you step back and you say really are we ready yet for putting this kind of information into a kind of a simple population like many of the developing countries well if we look at the next slide which tells us in very simple terms and pre-genomic terms of about 30 years ago the value really of having some at least primitif idea of the pathophysiology of a disease when one is looking for genetic complexity these diseases are diseases of hemoglobin and humans have two alpha chains and two beta chains as adults so that boy is born fine he switches over to make beta globin he can't and this disease is not a disease of hemoglobin production but it's an imbalanced globin production because these excess chains precipitate damage the red cell precursors I was talking about this disease in Washington a few months back and I pointed out that what happens to these bone marrow of these poor kids they're a bit like a committee and I just likened it unwisely to an NIH committee where there's enormous input and enormous proliferation but no output and that's the problem now if you understood just the outlines of this pathophysiology you could sit back and say well I could understand the heterogeneity of this disease difference defects at this locus of different severity maybe children who inherited a defect in alpha production would be better off if you got two forms of thalassemia they might be better than one or maybe children who were able to make a little more fetal hemoglobin after birth and so one already had a clear view of what the genetic identifiers might be and if you take this old picture and you then transfer it to 19 or to 2000 wherever we are in the next slide what genomics has done for us it's not changed our kind of overall views of the principle of these sick kids it's just added of some levels of complexity so here we are with our excess alpha chains and our inclusions we now know that there are about 250 mutations here of varying severity those two children have the same mutations so we can't account for the differences here but we now know that there's enormous diversity in the human alpha globin genes you and I have four hundreds of millions of people in the world population are missing either two or one alpha thalassemia is by far the commonest mild genetic disorder so again some populations have more alpha genes so just up here this combination and this combination enormous diversity we now know that there are genes which make it more likely or less likely to make fetal hemoglobin after birth and that's heterogeneous we also know that there's another layer of complexity that the complications which the anemia causes for these poor kids, their bone disease their loading iron their jaundice are all under genetic variability even infection because of co-selection different populations have different back genetic setups now you might argue so what we know there's going to be genetic fine tuning but if we just take one example jaundice when I went to Sri Lanka a couple of years back I saw mild cases of this disease but the kids were all deep yellow and had never seen this before and there was no good reason for this when we knew about that there were promoter variants in this gene which is involved glucuronidization and the handling of bilirubin and any child with this kind of disease makes a little bit more bilirubin we looked at the promoter polymorphisms in Sri Lanka we found remarkably that about 30% of Sri Lankans are homozygous for the long polymorphism in this gene which actually makes them handle bilirubin less effectively now this is an extremely distressing symptom and these kids are deep yellow, they're almost green if they're homozygous for this polymorphism with this disease now the two messages here if you are looking for monogenic diseases in different parts, in populations you do need to know the frequency of these modifiers and in this case it did occur to me that if we knew nothing about the cause of this disease and we'd done a genome search for these mildly anemic children we might have picked up the beta-globin genes the one thing we certainly would have picked up would have been this gene and if we'd done that and we thought that this was the cause of this disease or even related to the cause we'd have set the field back about 70 years because the early people who investigated these diseases oh yeah, they may be pigment disorders now can you take this complexity into the developing world well I think you can you don't need to know a lot you need to know the mutations and it turns out that every country with these diseases has a different set of mutations you need to have a rough idea of the frequency of the major modifiers and so what's happened over the last 10 years or so who see the next slide is that programs have been developed as partnerships between universities in the richer and poorer countries and they've worked extremely well as local simple partnerships where people have come over from the high frequency countries learned simple DNA technology taken it back with them developed their own programs and at the moment we're trying to extend this type of idea by developing regional networks so in Asia for example that they can take these kind of this simple technology and help countries for example in Thailand they have enormous expertise now and they can generate that around countries like Cambodia where these diseases are so frequent and if you look at the next slide in the Mediterranean island populations this is Cyprus where they decided because the high frequency of this disease if they treated all their patients with blood they would by within 15 years 40% of the island would be blood donors and they would be spending twice the island's total health budget on this one disease and they decided on a preventive program of prenatal detection and counseling and you can see what's happened with this type of joint training program to the frequency of the disease over that period and as the Mediterranean islands developed these simple technology if we look at the next slide it gradually spread in exactly the same pattern north simple collaboration between laboratories training go back, set up programs in their own countries I won't go into the kind of reasons for this variability because there are enormous kind of differences in the distribution in these countries of populations for example in Thailand about 80% of the population are rural and there have been all sorts of important ethical and organizational issues but this list is now very much longer so can this kind of simple university or other institution north-south partnership be applied to the broader, some of the broader diseases of the developing world just to introduce simple DNA technology we see the next slide if we look at genomics and infectious disease obviously we've got several genomes pathogen, vectors and the human genome and we've already got some hints today about the possibilities here but there's still only possibilities so let's just go to simple DNA diagnostics and the human, what we know from as Adrian was telling us this morning about variability and susceptibility other things here that are already a value for the developing countries if we look at the next slide the of course most of you know that DNA diagnostics in communicable disease have taken off in quite a big way and we don't need to go through these in detail at all either for detection or for analysis of resistant treatment regimens there's a beautiful paper in Lancet which may not have got to over here from last week from the Oxford group but we'll see on the use of the ability to identify genes specific to mycobacteria tuberculosis rather than bovis and the development of a very beautiful simple fast technique which seems to be much more accurate than simple skin testing for identifying people with latent TB now question is are they cost effective are they too expensive for the developing countries and exactly the same technique method has been used the north south simple collaborative method between universities in this field as in the globing field if we see the next slide and this is Eva Harris's program for the Sustainable Science Institute and it's been exactly the same type of program you take groups of scientists you develop training programs you develop workshops and what you try to do is to develop simple cost effective DNA diagnostics many of these DNA diagnostics are ridiculously expensive and honestly they're not much better than standard steam microbiology but where you've got organisms that are difficult to type difficult to grow and the elegant outcome of this program has been the extraordinary value for diagnostics for specific organisms and this is mostly work in South America for the early identification of dengue types leishmaniasis and leptospirosis which has been a notoriously difficult organism so I think that's a beautiful example of when DNA technology is ready is it cost effective in the setting of a developing country yes it is and so I think one should not be negative about taking these technologies to the developing world if we could see the next slide the Adrian talked to us this morning about resistance and susceptibility to infection there are early hints now that this type of information may be useful in clinical practice some very nice studies from West Africa the MDR polymorphisms this is a gene which is involved with the identification of people like a protein probably polymorphic become poly polymorphic due to gut infections in Africa and this seems to play an important part in some of the susceptibility to AIDS therapy particularly with the protease inhibitors the very nice data from Marley suggesting that the polymorphisms in the malarial parasite which this is the one that denades chloroquine resistance drug resistance may be again very valuable in early identification of resistance in communities and then we talked a little bit this morning Adrian about the polymorphisms of the malaria resistance which have become so of such interest in recent years and the next slide just recapitulates I think Adrian's list I don't think you can see this Paul might but if you can bad cerebral malaria and what Adrian pointed out this morning was that so much of our recent work looking at these polymorphisms has taken WHO criteria this is the worst thing that can happen to you with malaria or profound anemia and if we look at the next slide this is the list that I think Adrian showed this morning the increasing list of human polymorphisms in relationship to resistance or susceptibility now as Adrian said lot of these are mild they need confirmation and so on but if sickle does offer 80% resistance to getting in the state of that child and the band 3 mutation that's so common in parts of Melanesia seems to cause complete resistance to cerebral malaria if we just look at the next slide this is a case control study in this region of Papua New Guinea where alpha thalassemia is compared with normals in a case control study what is the likelihood of you being admitted to hospital with these severe complications and if you have got homozygous for the mild forms of thalassemia the likelihood is to 0.4 and in heterozygo it's about 0.6 and the fascinating thing about these data and I think about emerging data on sickle are the fact that these children seem to be protected not only against malaria but against other severe community infections whether that's because the protection against malaria makes them less likely to pick up other community infections is not clear but these are very big effects and if one was then moving into these populations to test a vaccine, an attenuating vaccine one really has to know about these issues like sickle the polymorphisms with really big effects well of course the big science approach just to finish the next slide please over the last year the complete sequence of the malarial parasite P. falciparum has been announced the sequence of anopolis gambia which carries the wretched beast around quite remarkably actually a lot of sequence information on this form of plasmodium which is the mouse plasmodium and already a lot of information about the stages of the life cycle in terms of proteomics and patterns of protein again a lot of information appearing already about metabolic pathways which are obviously making people very very excited about the possibility of new targets for therapy and vaccines but as we heard from Peter earlier this afternoon this is all going to take a while I do get lucky breaks and my last example is next slide please was such a break which happened just a few couple of years back searching through the falciparum genome finding a metabolic pathway which is used by the parasite discovering that a useless antibiotic had been designed to target this pathway and it was no longer in use finding very quickly and remember this all relates to the genome sequence that it was effective in murine malaria and within two years the first human trials and I gather from Thailand this is looking quite promising so you will get lucky breaks along the way so I just tried to give some examples of DNA technology that is already a value in the developing countries and if we just summarize see the next slide I think the way we have to go in the future it's clear for us as well as just decent human beings really that the fear out there wherever I went doing this exercise was you're going to have new more high technology medicine in the richer countries which is just going to make the guide worse and we have to do something about that in reducing this awful gap where 90% of our research at the moment is directed at 10% of our diseases now I don't want to talk about action of governments and non-government organizations tonight they're really self evident and we've heard something very interesting information about patents at least today I just want to leave this message actually for the younger scientists in the audience and for those who are involved in education I think there's an enormous opportunity for developing these partnerships and what it requires really is a much greater understanding of the global aspects of disease in our medical schools departments among our young people it's going to take a big change in emphasis but you can do it we started doing it in Oxford about 10, 20 years ago and now we have groups all over the world who come back each year about 200 young people and we have this combination of units in the developing countries together with these types of partnerships I've tried to describe this afternoon and what we should be doing is asking the question what is cost effective DNA technology for the developing world because once you've introduced that you've got a basis and as further developments from genomics come these countries can develop their own kind of technology basis to also be done obviously at the industry level and so on I don't want to prolong that tonight I think to do that we will have to change our outlook on education and try to excite young people into more work in this field that will require revitalizing clinical research if we're going to make the best of translational research and above all it's going to be a big education problem in trying to integrate a public health world with genomics but I think we start in this small way this is the way forward now it didn't seem right for an old man like myself to have the last word at this meeting can we see the last slide please there are two messages I had tonight really one was the vital importance of phenomics as much as genomics and the we've heard about it during the day the vital importance of not ignoring clinical research if we don't accurately define phenotypes and if we don't look for the rare experiment of nature we're going to miss out William Harvey realized this in one of his last letters to a friend and the other key issue is because of the extraordinary complexity of human disease probably a sick human being is the most complex biological organism imaginable and I think Francis Crick very much realized that but I'm not going to give Francis Crick the last word about two years ago Jim Watson came to Oxford on sabbatical and he's great value for the youngsters and on the very last day was the on the last week I said to him look Jim would you give a lecture in the medical school and we'll get the entire medical school every young person and the place was packed and they're all sitting on the edge of their seats and he gave this lecture and at the end he said take some questions and everybody was terrified and then one young man put his hand up and said excuse me Dr. Watson but you made the greatest discovery in biology in the last century perhaps ever could you tell us what's going to be a discovery of equal importance in the next century and there was a long silence and the eyes rolled and you could hear all these bottoms moving along to the edge of their seat in expectation and then he said there won't be one and I think he may be right in the medical world it was 300 years after William Harvey before the circulation of the blood had any effect on cardiology it was 70 years I think it was 1880 when William Koch announced the discovery of tuberculosis 70 years until Waxman and Penicillin Streptomycin and another 70 years and we're still in trouble with tuberculosis medical advances follow quite a while after laboratory basic laboratory science advances and we shouldn't be expecting the whole of medicine to change overnight and I think this is a very important message that we have to get it's not an active message to our grantors or to the news press but we've oversold but there are things that we can do now simple things and we can do them at individual scientist individual institution lesson to start to move into the developing countries it's not too early to do that and it will form a basis for the tremendous excitement which I'm sure will come from this field in the future thank you well thank you Dr. Widerall this was an inspiring lecture but all good things have an end and I'm here to close the symposium there's no question over the past two days you've heard a dazzling array of insights into the past and the present and the future of genomics and scientific research I'm sorry that I couldn't attend today's sessions but these sessions were terrific I really think that we owe a great gratitude to the National Human Genome Research Institute and also the foundation for NIH without those two partners I can assure you that you wouldn't have seen the symposium program all the associated programs and the great event last night there's no doubt that without the support of the many government partners and private partners we couldn't have made this happen now Dr. Widerall has helped emphasize something that is dear to me and that is that in fact we need to think now of the planet as one and we need to invest across the board when we invest in science to make sure that the 90% which is not researched with the resources that we have does get researched and for someone who comes from the developing world I think this is a very important message I'd like to really give you some personal thoughts about what happened between yesterday and today and our release at noon yesterday of the human genome fundamentally I think this is going to define science in the 21st century and you've heard that many times but why is it I personally think it's going to be the challenge of the century when you look at the burden of disease in our country here and over the world when you look at the costs to social to national budgets and the social consequences of not tackling the growing burden of disease you realize that we're engaged in a race but then when you think about what humankind has accomplished through science and why we need to accomplish more you realize that in fact it is the disequilibrium between environmental conditions and humankind that we as an intelligent species are in fact responsible for that generates what I would consider an evolutionary accelerator the changes between our genes and environment and lifestyles the speed at which they're occurring is pretty unique in evolutionary history and when you if you thought about an event that occurred in the environment and you thought about the number of generations it would take evolution to adapt us to mutation and selection you were talking about 5,000 to 10,000 years of evolution for any particular change in the environment well this is exactly what we're facing we're facing rapid changes in our environment that have occurred over the past 100 years would it be food supply with the emergence of obesity would it be globalization with the emergence of general diseases and exposure to agents that we're not exposed to so at the end of the day you realize that there is a need for an acceleration of adaptation and science in fact plays that role in fact understanding the genome and having the ability to modify our own gene environment interaction scientifically is going to be a requirement for survival of the species in some ways but at the end of the day unless we do this we're engaged in a race I think that could not be one without the human genome sequencing being done first so I think there's no question that today and yesterday will be in the scientific calendar of great days so I'd like to make an announcement that we have decided to change actually the scientific calendar as you know we are on the Gregorian calendar and stars January 1 so I'd like to announce to you that today at the end of the symposium we've just finished day 1 AG this is day 1 after genome and yesterday we were basically last day BG and with that I'd like to really thank you all for your participation thank the organizers and thank Dr. Weatherall for his magnificent lecture and all the other speakers please enjoy and let's go on AG day 2