 Okay. We're going to move on. Our next speaker is Rick Lifton from Yale University talking about genes, genomes, and the future of medicine. Thanks, Eric. I'm delighted to be here today to have a chance to speak to you. I want to start with the proposal that many of the goals of biomedical investigation revolve around four major prospects, understanding the pathosesiology of human disease, with the expectation that this ought to contribute to the ability to make early diagnosis, enable prevention, and enable new and effective treatments. Genetics clearly has had profound impact on our understanding of pathophysiology. And going forward, the one thing that I think we can safely say is that we have great tools to understand pathophysiology, where that will ultimately lead us in the ability to diagnose, prevent, or treat disease will really depend on what the actual pathophysiology is of each disease. And until we know that, we really can't speak, in my view, coherently about what the prospects are for any of these alternatives. I do want to reflect on this anniversary that we've been through, I believe, three distinct eras of gene discovery in this explosion, and I want to make the point that each of these has been driven by saltatory changes in technology, and all of these technological advances have been driven by NIH and NHGRI. And I don't think there is any doubt that without the focus and the commitment to build tools, that we would not be where we are today. And I think it's worth reflecting on, there are so many ways that this could have gone off the rails and not happened, that I think it's just miraculous and a real testimony to NIH's and NHGRI's commitment that we have come to where we are today. So those three, these three eras, in my view, have been the development of complete genetic maps, the development of complete lists of variation of sequence, and now the rapid advances in the technology for whole genome and whole exome sequencing. And so when one thinks about how this has transformed our thinking about disease, as has been discussed, there are thousands of diseases that have been figured out now at the molecular level, and you can't open any textbook of medicine and not find organ systems and pathways that have been profoundly changed by genetic contributions to these disorders. And this is simply a list that I was thinking of last night. Without any particular goal beforehand, I recognize that all of these have come out of the three different eras that I was just alluding to, some from the early Mendelian era, some from genome-wide association and admixture mapping, and some from whole exome sequencing. And all of them, I think, hit my list of favorites because they have pointed to potential therapeutic applications or diagnostic capabilities that have profoundly changed the way we think about these diseases clinically. BRCA1 in cancer has obviously had profound impact for the early diagnosis and prevention of adverse consequences of breast and ovarian cancer. We've changed fundamentally our view of obesity from being a matter of personal strength and character to one of biology with the recognition of the impact of leptin and mutations in the MC4R receptor to have profound effects on body weight. Our understanding of Alzheimer's has been dramatically transformed from thinking, well, maybe these amyloid plaques and the proteins in them have something to do with the disease and maybe not, to recognizing that there's a causal relationship between mutations in gamma secretase and Alzheimer's disease and similarly rare families with Alzheimer's with mutations in APP. New insight into sleep-wake regulation from discovery of the gene for narcolepsy, a recent finding that I don't know how many of you have, it's really percolated through. If you walk into any dialysis unit in the United States, you will find an excess of young African Americans on dialysis. This has been hotly debated over the years as to why this is, it's simply a matter of access to care, explaining why they end up on dialysis. We now recognize from the work of Martin Pollock published just this fall that there are common variants in APOL1 where the heterozygous state protects from infection with the particular tropanosomes in Africa. Homozygous state as a pure recessive increases the risk of going on dialysis by a factor of seven. This will profoundly change the way we think about this disease. Eric has mentioned IDH1 and glioblastoma, multi-forming, gain of function mutations with large effects for glioblastoma and we've also heard about innate immunity and autophagy in inflammatory bowel disease. And lastly, NAV1.7, if you're missing this channel altogether, you're ostensibly normal except you're insensitive to sensory pain, which obviously has interesting therapeutic potential. In my own work, we started for years with an interest in hypertension for the primary reason that this is one of the major contributors to death on the planet. The major risk factor for death from the number one and number three cause of death in the United States and worldwide and its treatment is far from optimal. Most patients are not adequately controlled and most of those who are require three or more drugs and its pathogenesis has been unknown. The reason it's been so hard to understand from physiologic analysis alone is demonstrated here in this working model for the regulation of arterial blood pressure by Arthur Geithner. And you can see that this, I think you'll concede that this is a pretty complex wiring diagram and as a consequence, it's been hotly debated as to whether this is a primary disease of the brain, heart, kidney, adrenal gland, or vasculature. And this struck us now some 20 years ago as an ideal place for genetic investigation with the idea that perhaps we could identify single lesions that would have large effects on the trait. And my one minor disagreement with Eric in thinking in his beautiful talk was he mentioned GWAS as the analog of drosophila mutagenic screed. That would be like Yanni Newsline going out to understand development by taking a collection of millions of flies and identifying subtle differences among them and then trying to understand development. That of course is not what was done. Mutagenesis was done looking for mutations with really large effect that you could pin your hat on and pin down and identify the molecular basis. And so the analog in human is to look for the extreme outliers that would be what you would get from a mutagenic screen if you could do one, which of course you can't, but the fact that there are six billion individuals walking around on the planet, virtually every base that can be mutated that's compatible with survival is walking around somewhere on the planet today. And if we cast our net wide enough and are observant enough, we can find these extreme outliers in the human population. And we have done this repeatedly. And instead of finding genes that are distributed all across the physiologic landscape, the genes we have identified that drive blood pressure to the either the extremely high or the extremely low end of the distribution all converge on a single final common pathway that regulates how the kidney handles salt. And these mutations are distributed both along the nephron, if this is one of the million nephrons in our two kidneys. There are mutations that alter salt reabsorption along the nephron. There are also mutations that lie outside the nephron in endocrine systems that regulate the activity of salt reabsorption in the kidney. So for example, there's an epithelial sodium channel that if you have loss of function mutations in that channel, you have a profound hypotensive disease that typically causes death in the first weeks of life due to an inability to maintain a blood pressure due to a profound intravascular volume depletion. Conversely, if you have activating mutations that increase the activity of that channel, you have the opposite phenotype. Severe hypertension manifest from births onward. In addition to finding genes that we already knew something about from prior physiologic analysis, there are a number of genes that were previously known of entirely unknown functions such as the wink kinases, which we've demonstrated by subsequent physiologic analysis. Our master regulators of many of these diverse salt and water retaining pathways. Importantly, these effects are not subtle and they point to a single vector. There are diverse effects on potassium, magnesium, and calcium. But if you know what is happening to sodium and chloride, you know what is happening to blood pressure. Increase net salt reabsorption, blood pressure goes up, net salt reabsorption goes down, blood pressure reduces. And these effects are individually large. So we think perhaps one of the most telling features is that there are five genes in which loss of function mutations drive blood pressure to the extreme low end of the distribution in the human population, whereas gain of function mutations in those same five genes drive blood pressure to the opposite extreme. The entire spectrum of blood pressure encompassed by the 6 billion people on the planet can be found with variations in a single gene going from loss of function to gain of function, telling you the power of the salt homeostatic pathway to regulate blood pressure. This has been turned into predictive medicine by applying diagnostic methods to identify individuals who have these mutations regardless of their family history or blood pressure. And prospective screening is quite successful in identifying these. And this is clinically important for getting these patients on the right treatment. And also in mitigating some of their consequences, many of these mutations result in cerebral hemorrhage at early ages. In this family alone, there are six people who died from cerebral hemorrhage before the age of 45. And we now prospectively screen these families for intracranial aneurysm. And in this family alone, there are two individuals who have been clipped based on this idea. So if salt is so important, why aren't the diuretics that are the most commonly used anti-hypertensives worldwide more effective as single agents? And why is the epidemiologic data relating blood pressure and salt intake so weak? This is a nice example where finding the gene suggests environmental interactors and these extreme outliers can be used profitably to understand basic aspects of biology. So this is a kindred segregating mutations for the thiozide-sensitive sodium chloride co-transporter. And if you're missing both copies of this, you have a syndrome called Gittleman syndrome, which is the most mild of the blood pressure-lowering mutations. And in this kindred, we identified 200 descendants and identified those individuals who had two mutant copies of the gene, those who had one mutant copy of the gene and those who had none. And that's a very simple question. By measuring 24-hour urinary sodium excretion, identifying how much salt these patients are actually eating on their own normal ad lib diet. What we found was that compared to their wild-type brothers and sisters, those with two or even one mutant copy of the gene are eating a lot more salt. This is not a subtle finding, and it makes a couple of important clinical points. One, when we give patients a thiozide diuretic and a gentle pat on the back and say, don't eat too much salt, we're ignoring the fact that the drug we are giving them is actually driving them to engage in the behavior that we're encouraging them not to participate in. Well, this suggests that if we really want to get where we want to go, we shouldn't be driving thiozide use to higher and higher levels without intervening with something that might blunt the drive for, to increase dietary salt consumption. And this is one of the bases for combined use of inhibitors of the renal angiotensin system, along with diuretics, which is becoming a standard of care. The second point that it makes is paradoxically, in families such as these, the individuals who eat the most salt have the lowest blood pressure. Of course, this is not a paradox when you understand why that is. They have a primary salt wasting problem, and they're compensating by eating more salt. This makes perfect sense in the light of the molecular understanding, but obviously indicates the confound in trying to simply relate dietary salt intake to blood pressure in the general population. In addition to these rare Mendelians, for diseases that we have not been successful in identifying rare Mendelian forms of disease, we have done a number of genome-wide association studies. One of these is cerebral hemorrhage, a devastating consequence frequently interacting with high blood pressure. And with Murat Gunnell and others, we have done two genome-wide association studies comprising 6,000 cases and 14,000 controls, and have a list of genes that are coming out of these that I'm very interested in. One of these is the same variant that contributes to myocardial infarction on chromosome 9. And there are others that are beginning to emerge as a pathway, as Eric was suggesting, in endothelial repair. And a lot of work will remain to prosecute this, but this is the direction that it's pointing. Others have done genome-wide association studies for hypertension, and these have been among the largest disappointments in the field. Studies of nearly 100,000 individuals have explained about 0.2% of the variation in systolic and diastolic blood pressure, and these will prove difficult to prosecute. There may be interesting genes under there, but this remains to be established. In our case, we asked a very simple question. We had three genes that, in which homozygous loss of function mutations caused extreme lowering of blood pressure. And we asked the basic question, what is the prevalence of loss of function mutations in the heterozygous state in the general population, and do these affect blood pressure? And to approach this, we sequenced 3,000 members of the Framingham Heart Study and identified likely functional variants based on variants that occur at positions that have been conserved from invertebrates to humans. And we subsequently confirmed virtually all of these biochemically as loss of function mutations. So first, about 2% of the population is heterozygous for mutations at one of these positions. Importantly, all of those that are inferred to be functionally significant are extremely rare. There are no common variants with frequency 1% or greater. These are all 1 to 2,000 to 1 in 40,000 in their prevalence. And as I indicated, we've now demonstrated that virtually all of these are biochemical loss of function mutations. Turns out that these actually have quite large effects on blood pressure. They reduce blood pressure at age 60 by about 10 millimeters of mercury. And this is sufficient to reduce the overall prevalence of hypertension by about 60% in the Framingham cohort. So these are rare variants with relatively large effects. So we're now entering into this third phase of discovery, which I'm particularly excited about because of the ability to identify rare variants with large effects and to identify previously unrecognized Mendelian traits. In our own group, Muram Choy shown here developed the whole exome sequencing approach, which is about 95% sensitive and greater than 99% specific for identifying novel variants. And the cost of this is coming down. It's about $2,500 last year. We'll likely be about $1,300 this year. So to take advantage of this, we've established a new sequencing center at Yale based on a different model. It's an open access facility. We didn't know what the demand would be in the first year of operation. 60 Yale faculty have used the high throughput sequencing facility. And our throughput has grown from virtually none to about three and a half terabases of sequence per month. So the applications will be to identify previously unmappable Mendelian loci, dominant reproductive lethals we've had almost no ability to find, recessive traits with high locus heterogeneity, somatic mutations in tumors, and then rare mutations with moderate effect in common disease, and finally clinical diagnosis. Our first proof of principle was simply a simple clinical diagnosis. A boy who was referred to us for one of these renal disorders, we simply sequenced his entire exome and nine days start to finish, identified a homozygous mutation in the gene SLC26A3, which was particularly satisfying because this is not a gene that affects salt reabsorption in the kidney, but instead affects the colonic absorption of chloride. We went back to the referring physicians, suggested a diagnostic test that confirmed the diagnosis, and this enabled institution of treatment that has been sustaining for this infant. Secondly, Murat Gunnell has a wonderful cohort of patients with arising from consanguinous union with structural brain abnormalities. This is a natural for this kind of application because novel homozygous variation is likely to be of interest. And in the third exome that we sequenced, identified a very interesting mutation. This is the MRI of a three-year-old boy. You can see the profound loss of the cerebral cortex, despite having normal cerebellum and brain stem. And you'll also note the complete loss of cortical folding compared to a control individual, a normal three-year-old subject. So this boy has a forebase deletion in the gene WDR62 of previously unknown function. And then going back to this cohort, it was interesting without the MRIs, the diagnoses were all over the place, but once you had the mutations and you got the MRIs, it turned out to be quite monotonous. They all had loss of cortical mass and loss of gyro folding. And these patients all had homozygous loss of function mutations in WDR62, which turns out to be expressed in neural progenitors and somewhat later in the maturing neurons. As an example of a new disease with a dominant de novo mutation, I'll give you this. Keith Chote came to the laboratory with these two photographs of patients with a disease called ichthyosis with confetti. The ichthyosis is the bright red skin. The confetti are all of these white spots. And when I asked him what's the genetics, he said there are seven cases in the world, they all arise from unaffected parents, no consanguinity, and they're all singletons who typically do not reproduce. And I asked about the spots and he said, well, we just biopsied one spot and it appears to be normal. So that led to the thought experiment that what could this be? Perhaps this is a disease caused by dominant gain of function mutations and all of these spots are somatic revertents. And how could you have high frequency somatic revertent? It would have to be some kind of recombination mechanism that we speculated. So we went on and demonstrated that this is in fact true, which is what's so wonderful about the technology. These patients have de novo mutations in the gene keratin 10. Interestingly, all of the mutations, although they occur from exon 6 to exon 7 and in the intron, in between, they all lead to frame shifting into the same alternative reading frame. And all of the white spots are independent revertents that have arisen by mitotic recombination having lost the mutant allele. Well, this is an example of high-frequency reversion by the mirror image of the well-known tumor suppressor mechanism and why this occurs with a keratin mutation is interesting but presently unknown. I'll close with a story that is just published in Science today. One of the major causes of severe hypertension on the planet is aldosterone-producing adenoma. This is found in 5% to 10% of patients with severe hypertension and hypertension clinics all around the world. These are interesting tumors. They are benign. They virtually never invade local structures or develop distant metastases. And this poses the question, like other endocrine tumors, is there a fundamental mechanism that links the constitutive proliferation in the tumorogenesis and the constitutive hormone production? This was, of course, a very difficult question to answer prior to the ability to do high-throughput sequencing. So we started with a very simple experiment. We thought, well, we'll start sequencing tumor normal pairs and we'll go as far as we need to go. We didn't need to go farther than four tumors to find a very interesting result. Of the first four tumors, the first thing we found was that these were quite bland tumors. They're on average two protein-altering mutations per tumor. And in the first four tumors, one gene was hit twice, the potassium channel KCNJ5. That was not uninteresting for reasons all explained subsequently. And we consequently went on and sequenced another 18 tumors. And in the first 22 tumors we sequenced, there were eight mutations in KCNJ5 and they were all one of two variants in the gene, G151R and L168R, and the likelihood of that occurring by chance is exceptionally low. So these mutations tell you what they are actually doing. The potassium channel has a selectivity filter shown here from the crystal structure by Rod McKinnon. And the reason these channels are selective for potassium is that there is a gatekeeper series of amino acids, GYG motif, that is conserved in every potassium channel from archaebacteria to humans. And Rod McKinnon back in the 1990s had demonstrated this selectivity filter and it demonstrated that the key residue G151 and this tyrosine that 152 were key to the maintaining selectivity. And these two mutations in adrenal adenomas, aldosterone producing adenomas, mutate this residue directly and residue in the second transmembrane domain that interacts directly with the adjacent tyrosine residue. So this predicted that these mutations would directly alter sodium handling. And this indeed is the case, not shown here, but we've done electrophysiology which demonstrates that indeed these mutant channels conduct sodium as well as potassium. And the consequence of this is readily understandable. The normal adrenal glomerulosa where aldosterone is made, these cells are hyperpolarized owing to constitutively open potassium channels. Aldosterone gets made normally by either closing these channels in response to angiotensin 2 or by extracellular potassium increasing, depolarizing the cell. This activates a voltage gated calcium channel raising intracellular calcium channel, calcium levels. Acutely, this causes increased secretion of aldosterone. Chronically, it causes cell proliferation. So these mutations do this by simply opening the channel, making them permeable to sodium, allowing cell depolarization, raising intracellular calcium, and leading to both chronic aldosterone production as well as cell cycling. Now, you might have argued that this is actually, may not be the complete story. There might be other mutations that contribute, and that could be the case. However, we now have five families that have inherited mutations in a Mendelian form that are associated with congenital aldosteronism. In this family, for example, this individual father who was investigated here at NIH in 1957 had profound aldosteronism at age five, needing both of his adrenal glands to be taken out. He had two daughters, both of whom had their adrenalectomy in childhood as well, and they have inherited mutations in the same potassium channel that also make the channel permeable to sodium. So past views on salt and blood pressure before the genetic era, the Salt Institute and Industry Lobby like to say one thing we know for certain, salt does not cause high blood pressure. That view has changed. The WHO and NHLBI now recognize the essential role of reducing salt balance in the treatment of hypertension. Early use of combinations of diuretics and inhibitors of the renin-angiotensin system are now recognized as a key combination. Prospective trials have shown that if you reduce dietary salt, you reduce blood pressure. This has been modeled by Lee Goldman this year in the New England Journal to show that if you reduce salt intake by 25%, that you'll have a profound effect to reduce the number of strokes, heart attacks, all cause death, and reduce healthcare costs in the country. As a consequence, the Institute of Medicine this year released a study on strategies to reduce sodium intake in the United States focusing on the fact that a very large fraction of our dietary salt intake comes from processed foods, and if you intervene there, you might be able to make progress. Lastly on therapeutics, we've identified a number of potential targets, and the most interesting thing is to find targets that have large effect on blood pressure but are neutral on the adverse effects. And for reasons that are clear from physiology, patients who have mutations in the potassium channel, ROMK, have profound reduction in salt balance, but no discernible effect on net potassium balance. And this has led to this being a very promising target for pharmaceutical development that is now in clinical trials in humans and shows great promise. So as we think about going forward in using sequencing in clinical practice, there obviously are tremendous capabilities, many potential liabilities. Who should we be sequencing and why? We should likely be using this to identify mutations that either establish diagnosis or markedly change our estimates of the susceptibility or that dictate therapy. They're already good examples as to where we might apply this. My own view is that in the near term, we will have the greatest traction when we focus on patients who are likely to have mutations that are pathogenic, but we'll see how this plays out. When should we be doing this? It depends on what the utility is. In Connecticut, we do 40 neonatal tests for about $40 on every newborn. One can imagine as costs come down, one can imagine that this might become cost effective, but a lot needs to be worked through. How do we deal with incomplete understanding of what every variant we see means? We know from the exome data that we will find handfuls of novel variants that are not present in anybody else that we can find to date. How do we communicate these results? These have profound implications for the education of healthcare professionals, patients, and social policy. And then lastly on therapeutics, we need to really help our industry colleagues to focus on the best targets and to prosecute them with passion. I think it's safe to say that the core strength of industry is medicinal chemistry and they have been late to the game in understanding the meaning and how to prosecute biology. There's an awful lot that we can contribute from the academic community to industry and without that we will fail to close the loop and with that I'll stop. Thanks very much for your attention. Time for a question if anybody wants to go to a microphone. I will then ask Rick, what's been your experience in your clinical realm with all these stuff as you're making the more general community or even non-community doc taking up the diagnostic side of what you're finding in hypertension? Yeah, so I think we spend an awful lot of time persuading people not to get molecular diagnosis because once you know enough about the molecular, about the relationship of genotype to phenotype, you very frequently can state categorically that if you have the following 10 clinical findings that everybody always collects, that that's pathognomonic for the disease and there's no need to pursue a molecular diagnosis. Interestingly, many patients want molecular diagnosis regardless and one of the areas that we are not very good at is inexpensively sequencing single genes and I think that's an area that we would clearly like to see improved. On the other hand, for the dominant traits in families, our experience overwhelmingly has been that families are interested, want to participate either in research studies or for clinical purposes and one of the barriers has been finding inexpensive, clear certified validation of these. Certainly for many of these dominant traits, that has been very important. I think in particular with this aldosterone producing adenoma, one can imagine the possibility of having a diagnostic test from the blood since there appear to be two mutations. If you can get a sensitive enough test, that could actually be very important clinically. We now spend an awful lot of resources evaluating patients by very invasive methods for these tumors, adrenal vein sampling being the dominant method and I think there are great opportunities there. All right, thank you Rick.