Scientist Stories: Euan Ashley, Precision Medicine for Rare Childhood Disease

Good morning, everyone. My name is Julien Sage. I'm a faculty in the departments of pediatrics and genetics here at Stanford. And I'm here to chair this next session and then to introduce the speakers. It's a session on precision medicine. And I would bet if I would just ask How many people in the room have you heard about precision medicine? I think everybody in the room has heard about it. Even President Obama has heard about it. And he actually launched his own precision medicine initiative earlier this year during his State of the Union address. So it's something that a lot of people have heard about. You also probably have heard about maybe slightly different names for precision medicine— personalized medicine, or molecular medicine, genomic medicine. And today in this introduction, I don't want to tell you what the maybe subtle differences are in what these different names mean. But maybe just quickly redefine what precision medicine is. And really, it's just customizing healthcare. In other words, it's tailoring medical care to individual patients, but not only based on the clinical features of their disease, but also on molecular, genetic, and/or genomic characteristics that we can now obtain from really new methodologies and new approaches. So this is really a new way of doing clinical care in this world of new clinical care. So today, we're going to be discussing precision medicine with an outstanding group of speakers. Actually, all the speakers here in the keynote and all the shorter presentations, all of these speakers have given hour-long lectures at international meetings and are really specialists in different areas of precision medicine. Unfortunately, they're going to have to give really short talks today. And I really want to thank them for agreeing to come here and giving this short presentation. I think it's going to be very stimulating, and I think it should be very exciting. So I'm looking forward to discussion at the end of the session. OK, so this is just a brief introduction I wanted to give. Next, I want to introduce our keynote speaker today here. Our keynote speaker is Dr. Euan Ashley. He's a Stanford faculty. He's in the Department of Medicine. He's a physician scientist, and he's the director of the Stanford Center for Inherited Cardiovascular Diseases. Heart doctor, basically. Yuan is also part of this new NIH network called NIH Network for Undiagnosed Diseases. So it's actually a real specialist in rare diseases. And Maisie, if I were to summarize his expertise in the lab in very few words, he's really good at clinical interpretation of genomic data. So if we acquire new genomic data on patients, how do we translate that information for clinical care? So I'm going to stop here and I want to ask Euan to come on stage and then give you the keynote presentation for this session. Welcome.

Thank you. Thank you. Thank you. Well, it's an absolute pleasure to be here today. Thank you for joining us here at this incredibly exciting conference. It's a real honor to be here in front of you and have the opportunity to give this keynote presentation today. Thank you to Julianne as well for the introduction. As he mentioned, I'm going to talk a little bit over the next 20 minutes or so about precision medicine. And as he noted already in the introduction, what most people think when they hear the term precision medicine is, what on earth is that? And so what is precision medicine anyway? He already gave you a little bit of a description of what it was. And I think he mentioned that even the president had coined this phrase from his State of the Union address. And here's what he actually said. I can't do an Obama accent. I should mention that to you first of all. I do a Scottish accent quite well. And if you push me hard at the end, I might do an American accent. But I'll probably stick with my native for now. I want the country that eliminated polio and mapped the human genome to lead a new era of medicine, one that delivers the right treatment at the right time. And this is the essence of precision medicine. We'd heard maybe more about personalized medicine or individualized medicine before this. Where did the term precision medicine come from? Well, it came from this document, which was published 2 or 3 years ago from the National Research Council. And in it, if you look in the appendix at the back, it actually tells you why they decided to use the term precision medicine instead of personalized medicine. They felt that there was some misinterpretation with personalized medicine, that maybe we were— after seeing a patient, we were going to go into the back room and make up a medicine specifically for that individual. And they thought that was somewhat confusing and felt that this idea of precision And the whole concept of precision, doing something precisely, you can't argue with that, is a good thing. Of course, the implication that we used to do medicine imprecisely is not one that we're exactly excited by. But I think doing something precisely, no one can argue with that. Well, here's the interesting thing you might not have known. Here's a much younger looking Senator Barack Obama from 2006. He introduced a personalized medicine bill in 2006 when he was a senator. If you read the bottom, it's a little small there, but it says, "A bill to improve access and appropriate utilization of valid, reliable, and accurate molecular genetic tests by all populations, helping secure the promise of personalized medicine for all Americans." This is not a new thing for our president. It turns out that Barack Obama has had a longstanding interest in the idea of personalized medicine, the idea of using genetic testing and other kinds of testing to personalize or individualize the practice of medicine. This bill was not successful, otherwise you'd have heard about it. There were hundreds of millions of dollars they were looking for. But now he's the president, and in his last 2 years, this came from him. And over the last few months, it's become clear what process led to that appearing in the State of the Union. And really, a word came down from the White House in July of last year to Francis Collins and folks at the NIH. And it really did come from Obama's desk. And those who have worked closely with him have been asked, is this something he really cares about, or is it just something that he's doing? And they said, no, he pushes aside other things on his desk and says, let's talk about this. So I think right at the top level of our government, we have someone who sees the inspiring potential of personalizing therapy. And so over the next few minutes, I'd like to tell you a little bit more about how we got to that point and what it really means for the future. So how did we get there? Well, it kind of all started here. Somewhere in the '90s is when we started the Human Genome Project. Somewhere in the early 2000s is when the draft released both from the public and the private partnership came out together. And this is the New York Times from the front when the draft partnership was announced. Bill Clinton is just off to the side there. And here you see Francis Collins and Craig Venter. And we'll come back to them at the very end of the talk. But this was the beginning. The Human Genome Project was funded for $3 billion. They didn't spend all of it on sequencing by any means. $3 billion was a significant amount of money. And probably most of you in the room have heard that the cost of sequencing is no longer $3 billion. In fact, many of you might have been at talks where people like me, and maybe even me, have shown you photos of graphs with a very steep curve of the decreasing cost of sequencing. And $3 billion, that's a lot. And by the second genome, it actually had dropped significantly to about $100 million. In 2008, it cost about $3 million. When we first got into this game about 5 years ago in 2009, probably about $50,000. You can now sequence a human genome for $1,000. We've heard about for years the idea of a $1,000 genome. Well, Illumina announced a machine just in the last year that can, with the amortization cost actually included, sequence human genomes at $1,000. This is truly transformative for medicine and truly transformative for our understanding of multiple diseases. And I've shown that graph multiple times. And I find that in Silicon Valley, I have to use a different kind of currency. What we understand in Silicon Valley is Ferraris. And so if you want to get a scale of just how much a millionfold drop in the cost of sequencing is, you should compare it to how much this $400,000 Ferrari would cost if it had dropped in cost the same amount as sequencing. Sequencing. I don't know if the dean is listening, but I don't get paid enough to afford the $400,000 Ferrari. You should know that. If it had dropped in cost as much as sequencing, it would be available for $0.10. $0.10, this Ferrari. That's a Ferrari for everyone in the room. We can all be driving rapidly away from here at the end of time. And I was also trying to think, what else is expensive? I thought, well, housing in Palo Alto is expensive. But I thought, well, you know— We're all here, and we all sort of know that. What else can you get for $2.5 million? Well, it turns out you can buy an island for $2.5 million. And if this island in Fiji, which is available, I'll tell you, I picked this off the web yesterday, if it had dropped down in price as much as sequencing had dropped in price, it would be available for $2.85. And if you're interested in this island or any others, I'll direct you to Private Islands Online, which is a real website actually available. And that Fijian island is available. So if you happen to have a couple million dollars available and you fancy it, feel free. We will all come and visit you there when you arrive. The question is then, if it has dropped in cost a lot, which is pretty amazing— now $1,000 for a genome— what are we going to do with all that information? And this occurred to a few of us here at Stanford a few years ago. And this was a paper that came from that idea essentially. And the idea of the paper is described on the front cover here, which is essentially that the data is no longer the problem. We are now at the situation where it's not getting the data that's going to be the issue. It's parsing the data and understanding the data and making the data meaningful for our patients. And that was pretty much the situation that we were in when I was in my colleague Steve Quake's office just a few years ago. He was only the fifth person on the planet to have sequenced a genome. And he sequenced his own. So he was the first person really to have done that. And he was describing his family history. In particular, he talked about his cousin's son who had died suddenly and unexplainably one morning after having had no illnesses in the rest of his life. He also described a very significant family history of cardiovascular disease. And as I was inviting him to my cardiology clinic to be screened for those various diseases, it occurred to me that really he was going to be the first patient ever to walk into a clinic with his genome. And that's a pretty remarkable thing because it was $3 billion just a decade ago. And it was $3 million just a couple of years ago. And now here was someone walking into actually our clinic just over there with his genome kind of on a USB drive. And so what do you do with that other than panic? I mean, we're not used to that. We're used to one gene or two genes. And it's only the coding part of each gene. We're used to like a tiny percentage of that information. Suddenly it's all there for us. What do we do about that? And so what we do and what I did at the time was call my friends. And some of them you recognize from this slide. One of them is actually up on stage later. And we broke the problem into three major parts. My group was interested in Mendelian disease. Obviously, particularly cardiovascular, but really across the spectrum. We could look at the whole genome now. So we can look at potentially the genetic causes of every disease. Pharmacogenomics is the idea of personalizing therapy according to the genome. Russ Altman, Teri Klein, and the whole group here have been building a database for years for that. Anatul Bute, who's on stage later, had been building a database to do predictive analytics for common variants. So we had all the right kind of people. And we had the data. We had the patients. So we just had to put them together. A great deal of fun over the course of 6 months really just trying to build the first tools really to interpret genomes. And here's a picture of Steve. And here's one of the questions. He had a high cholesterol at the end of this, a family history of heart disease. He said, can the genome help me decide whether I should take a cholesterol-lowering medication? In much too small a font, you can see on the screen there the guidelines at the time. They've been subsequently overwritten. But if you follow those guidelines through for a 42-year-old man, all the way through, there's 5 different steps. You plug in his age. You plug in his cholesterol. You plug in his family history. At the end, it tells you, physician choice. That's helpful, isn't it? Physician choice. You decide at the end. He fell right in the middle of the whole thing. So there was no guideline to help us. So we said, well, maybe his genome can help us. And so we looked in his genome for rare variants. And he had rare variants that predisposed to coronary artery disease. According to a tool's predictive analytics, he had common variants that increased his risk of cardiovascular disease. He didn't just have a family history. He actually had the genetic variants. In his own genome that predisposed to a history. He had the right variants that would give him a positive response to the drug we were talking about. So he had a pharmacogenomic profile that was suitable for this drug. And even better than that, he had a pharmacogenomic profile that meant he was unlikely to get the side effects from the drug. Putting all that together, we were able to make a personalized, predictive, individualized decision for him on the basis of a genome, which was just, I think, because we were lucky enough to be in the right time and place, the first time that had happened. But something we really looked forward to being the future of medicine. Because if it's $1,000 for a genome and it's $3,000 for an MRI, it's time that we started getting this kind of information available to a lot more people. And so I'm going to tell you a little bit more about exactly how that has happened in the last few years. This was 2010. Shortly after that, we had this kind of a headline. This is from Matthew Harper in Forbes. And this little boy is called Nicholas Volker. And from the age of 2 to the age of about 6, he had suffered incredible problems with inflammatory bowel disease. He had had multiple surgeries. In fact, they counted in this article 100 surgeries that he had. He was in and out. He was in hospital more than he was out of hospital. And nobody could explain his condition. It didn't seem to be any of the common causes. It didn't seem to fit the pattern. It was a very severe Crohn's-like illness, but nobody could decide what it was. The team were pretty desperate about this. And as you can read in the article here described by Matthew Harper at Forbes, They decided to do something very desperate at the time, which is to sequence his genome. In this case, actually, the exome, which is the 2% that is responsible for the protein coding genes, the genes, the proteins that do the work of the cell. And what they found was something that they weren't expecting because they found a mutation in a gene called XIAP, X-I-A-P. And this was a gene that caused a lymphoproliferative disorder, inherited disorder, that on rare occasions is associated with some degree of inflammatory bowel disease, but nothing like this Crohn's disease-like illness. They even did some cell studies. They said, well, maybe this is the cause of his disease. And they went and did some studies as shown here in the lab. They actually had a cellular model. They essentially stimulated his cells, and they did not respond appropriately with IL-8 here. Or the apoptosis pathway was shown to be increased in these cells that had this mutation and not in the ones that didn't. And so then they thought, well, if this really is an immunocompromised disorder, if it's a lymphoproliferative disorder, it could potentially be cured with a bone marrow transplant. But this had never been done before. In fact, no patient in this condition had ever had this kind of sequencing before. They had a lab-based study to try and demonstrate that this was real. So what they did was they went to their IRB. They went to the ethical review board and said, we have this information. We think this might be it. We'd like to do a bone marrow transplant because there is nothing left for this kid that has had 100 surgeries in the last 4 years. And the IRB were kind of a little scared of that. And they said, well, let's talk to another institution. So they went and talked to another institution's IRB. And then they talked to another institution's IRB. And finally, together they all came to the idea that yes, they should go ahead and do this because there was really nowhere else to turn. And an absolutely remarkable thing happened. They did the bone marrow transplant. Immediately the kid did much better. And after— this is now 3 or 4 years later— I do not believe he spent a day in hospital since. Absolutely remarkable story of the power of genome sequencing to find the answer to really these diseases that we had no clue about in the past. And I think this is very much a poster child. And Nick Volker is in many ways the poster child for what this can do in pediatric medicine. In fact, the Milwaukee Sentinel Journal— a team of journalists there won the Pulitzer Prize for telling this story. And if you're interested in more details, that would be a good place to go first. And the paper published in Genetic Medicine is shown here. Well, we talked about having a lot of information. And then we also talked about having it be cheap. But sometimes we need to move fast. And I've already shown you one Ferrari. Actually, sometimes I give these kinds of talks and people think I'm obsessed with Ferraris. I'm really not obsessed with Ferraris. But what I wanted to talk to you about next was doing this faster. And it seemed like a Ferrari was the right metaphor for that. Because sometimes we do need to move fast. And one of those places we need to move fast is the neonatal intensive care unit. Sometimes having the answer in hours rather than days, weeks, or months is critical to save lives and to improve therapy. And I wanted to tell you a couple of stories now that come from here and from team members. I think I even see some in the audience that were involved in these little babies' lives and stories. This first one was a little baby who was transferred in at— rather, the mother had been transferred in because it was felt that her baby at 36 weeks of gestation had heart block. The heart rate was found to be low. And it wasn't very clear at that stage what the cause was, but when the little baby was born, it became very clear looking at the ECG that this was Long QT syndrome. So this is not a mystery disease now. This is something where we actually can make the diagnosis. And the problem, as many of you know, with Long QT syndrome, in particular in a situation of high stress and when the QT is particularly long, is it can predispose to dangerous rhythms. And the dangerous rhythm shown on the bottom left of the slide here is torsades de pointes., and is a rhythm that is not associated with an adequate cardiac output. It's associated with cardiac arrest. And in the first few days of life, this little baby had multiple cardiac arrests and was saved on multiple occasions by the team. Eventually, she underwent implantation of a defibrillator, shown here in her abdomen, and had a bilateral stellate ganglionectomy. But Long QT is a syndrome, so we understand what it is. And over the last few years, we actually know for Long QT what the cause is. And if you know what the cause is, and the genetic cause, you can direct the therapy. Now we're not great at that yet. But we do have some therapies already that are a little more directed towards the underlying cause. In particular, of the 3 common causes of Long QT, the first 2 are caused by potassium channels. But the third is caused by sodium channel, completely different mechanism. If it was possible to know the answer, to know what gene was causing it, and to know quickly, that could make a significant difference. And in particular, for a little baby who's had multiple cardiac arrests in the first 2 days of life, it could make a very significant difference. So if you wanted to know the genetic basis of Long QT syndrome today, you'd send off a panel. It would cost $5,000, and you'd get about 5 to 8 genes back. And it would take 3 months. That is the world that currently exists simultaneously with the one I just mentioned to you, where you can sequence a genome for $1,000. This is the crazy world we live in. This has moved so fast that the standard test takes 3 months, and you get 5 genes, and costs $5,000. But the one we sent on this little baby took 2 days and cost $1,000. And we had the answer basically over the course of a weekend. We sent this on Friday. By Monday, we had the answer. In this particular case, KCNH2 is a long QT2. And we even found another variant within the rest of the genome that likely explained the neonatal presentation. So this is something that we can now do fast. And it's something that we should also be doing. And there are a number— ours is one, but there are a number of centers around the country that have begun to put sequencing machines close to their NICUs so that they can make these kinds of diagnoses early and fast. And this is the work of James Priest who did this in my lab. And particularly he built— oh, and a little picture of Jaslyn who's doing extremely well several months later. And a picture actually of a Circos plot of her genome right next to her there. So a happy ending to the story. One of the things that James developed for this was rapid turnaround analysis. And because it's not just the data, as I mentioned, getting the data is one thing. We can get that in about 28 hours. Actually, we can probably do that in 18 now. The rest takes a little bit of time. And these numbers are significantly faster than we've been able to do recently. In fact, we can probably knock another 5 hours off that second step and do this in just over 24 hours now. So that really is a game changer if we can do this fast and do it well and do it cheaply. And it's really a testament to how fast this world has moved that both these worlds are kind of living next to each other right now. Well, I wanted also to tell you a little bit about another case, a very similar case, actually. And you'll see why I put this particular quote up in a moment. There was another little baby who remarkably came just a few weeks after the last one, a very similar kind of case, with a mother who had a baby at late gestation with what seemed like it may be heart block, but a low heart rate. And then when the baby was born, it was again clear this was a presentation of neonatal long QT, with an even this time even longer QTC, and again a recurrence torsades de pointes and cardiac arrest. And this time we did the same thing. We'd done it once, we thought we can do it again. So let's just, you know, call the cavalry. We did it, we sent it off, we got the result back, and it seemed pretty clear. Except that before we really deliver the results back to a treating team, we always use the old-fashioned method too, just to make sure we have it right. And in this case, when we did the old-fashioned method to test the result, We found the result was a mutation, in this case, in the sodium channel gene shown here, SCN5A long QT3. But when we did the old-fashioned method, when we looked by Sanger sequencing, which has been around for decades, we didn't find it. And that seemed curious. So the new method found it, although there was a little bit of uncertainty within some of the— some of our algorithms called it, some of them didn't. It was very confusing and very unclear what it was. And then we didn't find it by Sanger. And yet it seemed to be the perfect answer, because right at that position, a previous kid had been diagnosed with long QT in the neonatal period. Exactly the right answer. So what was the explanation? We find it with the new technology, some of the algorithms and some not. So we moved on to think about what this could be. And the first thing we do is always look at the raw data. And just so that you can see, this is some of the raw data from the genome sequencing for this little baby. And you see all the reads. It's a little bit small for you to see. But where that yellow box is, that's where the mutation is. And where it shows you those A's, those are the mutant reads. And we normally get 30 reads per position. And in this case, 5 of them were variant. But 5 is not quite enough. And that's what confused some of the algorithms. So we did another thing, which is go to a local company— disclosure, I'm a founder of this company— and they did exome sequencing for us. And this is much deeper sequencing. So they did 210 reads. They also did this over the weekend. We called them and said, there's a little baby in the NICU. We need this result. And they came in at the weekend. And they ran this exome sequencing to try and help us explain. And they found the same smaller percentage. Within this baby's genome at the same kind of percent. Now if you're familiar with genetics— and I forgive you if you're not. Not everyone is. But if you're familiar with genetics, you'll realize that in general you will have all A's at one position. Or maybe you'll have A's and T's because you'll have one from your mom and one from your dad. So we're looking for ratios of 50% when we look down here. And that's what the algorithms are looking for, looking for 50%. So do you have all one? Or do you have one from mom and one from dad that are different? They don't know what to do, these algorithms, when we find 10% of the reads. That doesn't make sense. You don't have 10 parents and one is slightly different. That was very confusing. But this number was pretty stable at 8%. So we went through a number of different things. This was very confusing. Is there a duplication in the genome? Is it a very small duplication region? Is it that the reading, the map, the little individual reads are not mapping correctly to the right place? That happens with channels. We really pulled our hair out. And then the thought occurred to us that maybe this was something not unlike what we see in cancer, something called mosaicism, where not all the cells are genetically identical. Turns out that we actually are all mosaics. We think all our cells are the same, or at least have the same genetic code. But it turns out when you look carefully with the new technology, that's not likely to be the case. Well, we said, is this variant capable of causing this disease? So we expressed the channel in a cell. This is a cell line, a heart cell line. It had a significantly reduced function. In exactly the way that would cause long QT. So this was highly suspicious. So the question was, could it be cardiac mosaicism? But we didn't have a sample of the little baby's heart. And so we took advantage of the fact that the blood cells come from the same developmental origin as the heart cells. And we used a technique from my colleague who was mentioned earlier, whose genome we had interpreted, Steve Quake, who developed a microfluidic approach to single-cell genotyping. So if you can separate single cells and flow them into a microfluidics device shown here— and it's commercialized by Fluidigm. Fluodyne, you can put a single cell in each well. And you can take that single cell and even sequence the genome of it or look at the gene expression of it. In this case, we were really focused on this one particular spot. So using this technology, we were able to focus on that spot. And what we did was find the same percentage. So there indeed were 10% of the cells that had the variant, and 90% of the cells did not. So this really did seem to be mosaicism. What we didn't know, of course, is could that be cardiac mosaicism? Could that still be the cause? And what happened actually later with this individual— before we actually got to that point, before we got to that point, we were interested to see at what point in the developmental lineage that this came across. And we were able to get— James Priest, who also did this work, was able to get hold in conjunction with the team who were here, the clinical team looking after the baby, of some urine and some saliva and some blood that came from the different developmental origins and found, as you see here in red, that every single one of those had this mosaicism shown.. But for the mother and father shown at the left, it was not there. So this was not something inherited. This was something that arose during the development of this little baby. But the final question then was, is this in the heart? And as it turned out, that particular gene not only causes Long QT but can cause dilated cardiomyopathy. And the little baby went for a Berlin Heart and then finally a heart transplant. And at that time, we were able to get a little piece of the heart tissue. You can see her little piece of her heart here and demonstrate that not only was this shown in the heart, the gene and the variant was expressed. And so we were able to demonstrate cardiac mosaicism for Long QT for the first time and demonstrate that maybe there's a great deal that we're missing. Remember, the Sanger sequencing, what we regard as the gold standard, didn't pick this up at all. So like many things, when we have a better tool, a molecular microscope, we can often delve deeper and find things that we didn't ever see before. Well, let me take you back to the president. I have just a couple of last stories for you here. And one of the other things he said and one of the other things he pointed out in his talk his talk, State of the Union, a small talk, kind of like the ones we're doing this morning. Nobody really listening. To the president, that is. One of the amazing things that happened— one of the amazing things that happened, you can see shown on this slide here. During the State of the Union, when they announced the Precision Medicine Initiative, both the people behind him stood up. That does not happen very often. When you're looking on Google Images for a picture of the president giving the State of the Union, It's almost always that the two people behind him, one is sitting and the other is standing. And usually the guy on the left is the guy clapping and the one on the other right, who doesn't look that happy, but he is standing up. OK? That's important to note. He is standing up. And the president at that point was saying, in some patients with cystic fibrosis, the approach has reversed a disease that once we thought unstoppable. So precision medicine in relation to cystic fibrosis has also had a very interesting path forward. And I'd just like to tell you about that quickly. He actually pointed out— and visiting with him there for the State of the Union was was this guy, Bill Elder, who'd been a cystic fibrosis patient for many years and is now a medical student and was also at the press conference where they talked more about precision medicine. And he suffered from cystic fibrosis and had this particular variant that had been targeted by a company who developed a drug that could make the chloride channel, the one that's responsible, the one that's mutated in cystic fibrosis, can make that channel activate and be much more likely to be open. Well, and that is useful if you have a cell, an epithelial cell, the lining of your airway, where that channel is expressed on the surface. And there are different forms. When we dig deep into cystic fibrosis using the genetic technology, we find that some forms of it, the channel sits in the cytoplasm. It doesn't make it to the surface. Other forms of cystic fibrosis, it makes it to the surface. It just doesn't open enough. So if you have the subtype of cystic fibrosis where the channel gets to the surface, then this drug is going to work really well for you. But if you don't, it's not going to work at all. So it makes much more sense to learn about the genetic basis of the disease in these individuals and then target it to the people in whom it's actually going to work, which is very different from the way we do business right now. We take a drug and we try to give it to the most people possible because we're trying to make as much money as possible. But it's not a personalized or individualized therapy. And you can see in this follow-up data from last year that this is a strikingly effective agent. If you give the right drug to the right person, just as the president wants us to do more of, you see that here on the FEV1— this is the forced expiratory volume in 1 second, so this is a measure of the function of the airways— there is a very significant increase in blue shown for those who got the drug at the beginning, 10% increase towards the predicted value. And those who were delayed, you see there was no change, if anything, a small decline in this measure. But once they got the drug, a dramatic increase. And this is the follow-up safety data that was published just last year. So more than a year's follow-up. This is a remarkable thing. And the story's remarkable for one more thing. The Cystic Fibrosis Foundation, the patient foundation, actually invested I think $150 million in this company, which you'd think was a very controversial thing to do. But the company did rather well. And on that investment, they sold their royalty share last year for $3.3 billion. So this is a new thing, a not-for-profit investing in a drug development to help those patients, and then reinvesting the profits from their royalties when this drug did so well. So a very fascinating story, and the one that the President pulled out. And so I'd like to finish then today by just telling you one last thing. And Julian mentioned it graciously at the beginning. This is, I think, another inspiring initiative that came from government initially and is now spreading around the country. This gentleman here is Bill Gahl. '08 and the years before that, he went to the leadership at the NIH and said, I think that we should form a center for undiagnosed diseases. Because there are patients out there who go to doctor after doctor after doctor, center after center. They don't find the answer. And there is new technology, very much like the technology I've been describing in the last 20 minutes to you, that can help solve these cases. And they listened to him. And he was very persuasive. He's a very persuasive man. And they funded an initiative at the NIH called the Undiagnosed Diseases Program. And over the course of 5 years, this has been a tremendous success. And there's a lot of detail on this slide, but I think it's worth showing you the detail because it has really been an incredible success. They got 3,000 inquiries. And from that came up with about 1,000 medical records, a little bit over 1,000. And they reviewed them all. And these medical records are this high. Or now maybe they're online, but at that time they were still all this high. They were kids. They were adults. They were older people. They were men. They were women. There were all sorts of diseases across the spectrum of subspecialties. The ones I've highlighted— you probably can't see, it's a little too small— show that actually neurological diseases, and particularly child development, child neurological development, was a major focus of the undiagnosed diseases. This is an area where we do not have enough good tools, and we really need to work harder. They described new disorders as part of this. They found disorders that had previously only been described once that they were able to underline and show that they were real. It was really a remarkable thing, such that Even more remarkable, when they went back to Congress to ask if they could be renewed— like, are we doing good work? Could we have some more money? Congress did a very unusual thing, which is said, have more money. Go do this more. Spread around the country. Set up networks around the country and solve more of these undiagnosed diseases in more geographic locations. So that's what happened. And the NIH actually expanded. There's now 7 sites, 2 sequencing centers, a coordinating center at Harvard, and where Lucky enough to be one of those sites shown there, as you can see, on the West Coast with a group of folks including Paul Fisher, John Bernstein, and Matthew Wheeler. We have a center here at Stanford. And it's taken a long time to get the paperwork for the group together. We've been working for the last 6 to 9 months to be able to open the gateway so that we can open the network to the whole country. But we are inching there. And I think within the next month, we should be ready to formally open this. Here's a picture of us together at Stanford just a month or so ago coordinating and moving towards being able to open. The gateway. Well, in summary then, I think— I hope I've managed to persuade you that the new technology has been transformative already. Genomic medicine is here. We're doing it. We're sequencing genomes. We're sequencing exomes. It's changing the lives of our patients. It's changing the lives of the children who come to see us. But we can do more. We can be more precise. We can develop more drugs that are precisely targeted to this underlying basis of disease that we're able to describe in more detail than we ever were. We've gone from this situation where Francis Collins and Craig Venter were kind of vying to sequence the first genome to a situation where now Francis Collins and Craig Venter are vying to sequence 1 million genomes. Who can make it? This is the announcement from Reuters of the Precision Medicine Project to sequence 1 million people, Francis Collins. Roughly the same day, Craig Venter is tweeting that he's going to get to 1 million genomes in 2020 with his company, Longevity Incorporated. So not much changes. But I'm going to leave you with some words, I think from Aaron Sorkin actually. In fact, I think it's Justin Timberlake aping Sean Parker from this movie, if you're familiar, where he said, a million dollars isn't cool. You know what's cool? A billion dollars. But I would say a million genomes is pretty cool. But much more cool than that is when we can sequence the genome of everyone on the planet. And a few billion genomes would help us greatly. Thank you so much for having me. It's been a pleasure. I'd be happy to take questions in the forum later.