Scientist Stories: Euan Ashley, The Quest to Solve Medical Mysteries

It's my great pleasure to spend the next half hour or so, plus some time for questions, talking to you about medical mysteries and the quest to solve them, something that we feel very passionately about in our group at Stanford. But many times when people hear me start my introduction, another question comes to their mind before we get started. And I thought I might give you a little bit more than I usually do about who I am and why I talk. Funny. I come from Scotland originally. This is a photograph of Scotland. Some of you might recognize on the left here one of our best single malts, Laphroaig. This is from a little island called Islay, where the best, tastiest, and smokiest single malt Scotch comes from. I didn't grow up in the Laphroaig distillery or even on Islay, but rather in the big city Glasgow, a few hundred miles to the, to the east of that. This is my family, however, on the pier here during a visit just in the last year or two. My dad came from a medical family. My dad was a GP. This is what counts for a swimming pool in summer in Scotland. That's me with my blonde hair sitting on his back. This is also summer in Scotland. That's me shivering on the pier doing some fishing on the west coast. I was a bit of a nerd growing up, to be honest. I loved puzzles and I started on the Rubik's Cube and did that in, I think, about 30 seconds, although the rose tint in my mermaid tends to obscure those records. I also started programming computers when I was a kid, about 12 years old. This was what, uh, what a home computer looked like back in the 1980s, a ZX Spectrum with rubber keys. So I used to build computer programs for initially games to try and impress my friends, and then eventually I wrote a tax program for my dad's medical practice and spent rather too many hours coding, which, which is one of the reasons that Stanford was so attractive to me. I will get to that in a moment. Here's a picture of my dad on the day he retired actually from his practice, general practitioner. He was a practitioner for about 40 years in a town in Scotland. And my mom, who's sadly not with us anymore, shown here, she's a— was a midwife for many years. And I used to go on house calls with both of them because although I was a nerd, I really from the earliest point in my memory knew I wanted to be a doctor. And really it just seemed like the thing I was put on this planet to do. Little did I know later, and it was during a PhD in Oxford that I first realized— this is a picture of me outside Christ Church College in Oxford completing my PhD— that it might be possible to combine my passions in medicine and data science, although certainly wasn't called that back then. Computer science even then was not something that was readily apparent to a teenager in the west of Scotland, but it's something that became apparent as I went to medical school and visited Stanford for the first time. Shortly after residency, I did take the enormous leap. My wife was kind enough to come with me, and I arrived at Stanford in 2002, originally for a postdoc. I was going to be here for 2 years, and 20 years later, I'm happy to say we're still here. And life is fun. Stanford has been really, really good to me, and I hope it was good to you too. But through Stanford, I've had the opportunity to meet and spend time with advising multiple governments, including, as mentioned, the Obama administration. We're in conversation with the Biden administration now too around genomics and some of the topics we'll talk about. Number 10 Downing Street, the opportunity to travel to Asia and talk to multiple countries in that region and that part of the world about personalized medicine. Of course, we remain the tech capital of the world. And so the ability to meet with the leaders of the really transformative companies in our era is something that is not far away at Stanford. And it's an incredible moment because many of these tech companies are now interested in health in a way that was not the case maybe as much as, or as little as 5 or 10 years ago. But this is how I'm normally dressed, at least at a party. I think this was my sister's wedding. And we now have 3 American children. This is one of them wearing his kilt. It's a Fraser tartan in case anyone's interested. That's my family name, although my last name is Ashley. Fraser is our clan. And this is what I look like at work where I attend in the coronary care unit at Stanford Hospital as a cardiologist and see patients in the outpatient clinic there who have inherited heart diseases. And as we're gonna talk about during Greta's talk, mystery diseases of all shapes and sizes. So the essence of what we're going to talk about today is a secret code. It's a code that's inside every single one of you. It's the same code for you that's in every cell, but that code is unique for you. Even if you're an identical twin, there is nobody else on this planet, there is nobody else that has ever been on this planet that has the same code as you. And yet that code links you to every other living organism, to the history of your family, and to the history of the human race. This is, of course, your genome. 6 billion data points, 3 billion base pairs of data, tightly wound into DNA and 23 pairs of chromosomes that if stretched end to end, the DNA from every cell in your body would reach to the moon and back thousands of times. And it contains information. It contains information about how tall you are, how well you can see. It contains information about how you live your life and contains information ultimately about how you may die. It is in essence part of our humanity. And so it's not surprising that we spent many years trying to work out how to read that code. The code is so large that if you were to stare at the screen for 24 hours a day, 7 days a week, it would take you several months before you had seen the entirety of your genome on the screen. This is a lot of data. And what we're going to talk about today is how medicine has started to think about embracing that data and the good it can potentially do for biomedical research and the application of clinical medicine. But here's a place where we're going to start. This was the day on Tuesday, June 27th, 2000, after about 10 years, that the genetic code of human life was officially cracked. At least if the New York Times said it was, then it must have been that day. But this was the day that Bill Clinton, just off to the side, and Craig Venter, shown here with Francis Collins, the recently ex-director of the National Institutes of Health, announced to the world the draft sequence of the human genome. And this was a $3 billion project that took 10 countries, 10 years, and thousands and thousands of people to sequence just one genome. In fact, the end of the day was about half a genome, uh, from, uh, just about 1.5 people, but we'll, we'll come to that in a few moments. The second genome was actually from Craig Venter, who was shown on the screen there in that New York Times headline, and his genome was considerably cheaper. $100 million. And over the course of the next 5 to 8 years, the cost of genome sequencing dropped so much that it was compared to Moore's Law. And being from Stanford, you might be familiar with that. Gordon Moore and Robert Noyce, Bob Noyce, the co-founders of Intel. Gordon Moore in particular tried to encapsulate the speed at which technology moved forward, and particularly the number of transistors that could be put in an integrated circuit, thinking that it seemed to double every 1 year. Later he said 2 years. We'll call it 18 months. Whatever, it was really, really fast. Technology was moving really fast. And as you can see, the drop in price of human genome sequencing was going just as fast until in 2008, something kind of extraordinary happened. What happened was that the cost of genome sequencing left Moore's Law in the dirt, dropping out of nowhere to the point that you can sequence a human genome today for under $500. That is a spectacular, unprecedented change that rivals and even surpasses the changes we've seen in the technology sector. A genome has gone from $3 billion to $500. And I show this slide a lot in different talks, but I was reaching one day for a slightly more tangible way of showing this rather than a graph. And I was driving back actually from Stanford campus at the time we lived in Menlo Park. And so I was driving along El Camino Real, you'll remember it well,. And I looked to my left. We were going through Atherton. So there was the Ferrari and Maserati garage. And there was this car, a little bit out of my price range as a Stanford faculty member, $400,000. But as I was doing the mental math in my head at the stoplight, I realized that if this car had dropped in price as much as human genome sequencing had dropped in price, then instead of $400, it would cost around a cent. I think initially when I did this, it was 40 cents, and then 25, and then 5, and recently as little as 1 cent. A 1-cent Ferrari is what we're talking about when we think about the human genome in medicine. It's just quite extraordinary. So what does that mean for, for medicine? What does it mean for patients? Well, to tell you about that, I wanted to tell you a story about a friend and colleague of mine. Actually, at the time, he was and not a friend yet, but a colleague. This is Steve Quake, and he was the professor of bioengineering and chair of the department of bioengineering when I went to find him in his office, and I hadn't actually met him before, in around 2008, 2009. And he had made headlines that last summer by sequencing his own genome in just a few weeks for as little as $50,000. Now, I've just told you that you can get a human genome for under $500,000, so that seems really expensive. But the one before his genome— his was the 5th in the world— the one before him was $250,000. So this was a massive reduction in cost and had also taken several teams multiple months. And he did it in 5 days with 2 people in his lab. So this was an incredible feat that made headlines around the world. And I went to talk to him about a seminar. But actually, we never got to that seminar. And the reason was, after I sneaked through the Clark Building at Stanford and found through the gratuitously complex numbering system in his office, and inside his office found him, you know, with stacks of papers all around him. He was there pecking at his keyboard, and he called me over and said, hey, before we start, take a look at this on the screen. And he, on the screen, he showed me one of those old HTML web tables, and it seemed to have a lot of Gs and Cs and As and Ts. That's what I recognized as genome data, but I wasn't used to seeing it on a screen in this format. And I said, what is that? And he said, well, that's my genome. And this was just a jaw-dropping moment because here was the fifth person in the world ever to sequence a genome. He sequenced it with a technology of his own invention, and here he was scrolling through it on the screen in front of me. For context, at the time— and I have been since that time and before— running a center for inherited cardiovascular disease at Stanford where we send genetic testing to labs for our patients. And at the time, I would typically send a panel of 8 genes, and it would take about 3 months to come back, and it would cost around $5,000 to $10,000. And he was showing me not just 8 genes, not just 8,000 genes, but all 20,000 of his genes and all the other information in between, and he'd sequenced it by himself. This was just an incredible moment. But it went further than that, 'cause he started pointing out specific genes and saying, have you heard of that one? Have you heard of that one? I'm not sure if he thought I had a photographic memory for the entire human genome, which I definitely did not. But in fact, he started pointing at genes I recognized, including one that caused a condition known as hypertrophic cardiomyopathy, an inherited heart disease that's associated with sudden death. So all of a sudden my demeanor changed and I went into doctor mode and started asking him if he had any personal history of heart disease. Thankfully he didn't. Did he have any family history of heart disease? Initially, of course, he, like most people who knows their family history, he said, "No, I don't think so." And he said, well, wait a minute, my dad has some heart rhythm thing, ventricular tachycardia, he said, describing a potentially fatal heart rhythm. I said, well, anyone else have any heart problems in your family? And he described several individuals in his family. In fact, everywhere on this family tree that you see in front of you, where you see orange is some kind of cardiovascular disease. And when I asked the question about sudden cardiac death, had anyone ever died suddenly, he mentioned his cousin's son, shown down here at the bottom, who at 19 years of age with no former medical history just did not wake up one morning. So all of a sudden the meeting changed, and instead of it being a scientific meeting, suddenly became a medical consultation. And as I invited Steve to come to my clinic to be checked out for these familial cardiovascular problems, I realized, and he realized at the same moment, that he was about to become the first patient ever to walk into his doctor's office with his genome, his genome on a hard drive. And much more amazingly, and much more worryingly, I have to say, and anxiety-inducingly, the clinic he was walking into was mine. And the medical establishment, never mind our clinic, was just not ready for data of this kind of size. We're used to dealing with a cholesterol panel that has 4 numbers on it. We're used to dealing with a biochemistry panel with 10 numbers on it. We're not used to 6 billion base pairs of data, and indeed data that could speak to any disease and every disease. So this was a new era that was beginning. And so we started to call our friends and think about how we would start to attack a genome in the context of both a single individual and in the context of potentially a short medical visit with an individual. And we put together a team. This was— these were the junior leaders at the time. Like me, they all look a little older than they did back then, 10 years ago. But they put together tools like the one you see on the screen, where we can give specific genetic information about the risk of individual diseases, many different diseases, all from one genome, all from one test. We can even integrate that with environment and with lifestyle. Wherever you see large font size on this particular graph shown here on the right, It's something to indicate that the disease is at higher risk. And every time there's an arrow, there's an environmental or lifestyle influence, like for example here, exercise, diet, coronary artery disease. But we're integrating that with genetic data for the first time. And this was kind of an amazing moment for us, and we were very excited to do it, but I don't think we quite realized the impact at the time it would have when we released this into the world. We ended up for several days getting lots and lots of calls, including to speak to NPR. This was on the Morning Edition show from a number of people different media outlets. And I think people were fascinated with the idea of personalized medicine, the idea that the genome in some way tells you something about what it means to be human. And more than that, that maybe it's time we start to think that we could get personalized care from our doctor in a way that we expect personalized care from, let's say, our hairdresser. But medicine isn't today been set up by that. And we were privileged and honored to spend time talking with the Obama administration and thinking about precision medicine as it moved forward in the, the in-place of the genome in medicine. And the work was even fortunate to be featured at the Smithsonian. This is a picture, and I would get text messages every, every few months or a few weeks as somebody would go by and be rather surprised to see some faces they recognized there. Well, to take you to the next chapter in the story, I wanted to take you to Buffalo in New York. And I don't know if any of you are from there, have visited there, have relatives who are there, I personally only knew one thing about Buffalo in New York before this moment in time, and this is it: buffalo chicken wings, which was really the only thing I knew. And I appreciate there's a lot of, much more value to Buffalo than chicken wings, but even the story behind chicken wings is really interesting, and I won't— I don't think I have time today to go into it, but if you're interested, it's both in the endnotes of my book and online if you want to Google it. They really do come from Buffalo, which was the first thing. But the reason Buffalo is important here is because it was in Buffalo that the individuals who were recruited for the Human Genome Project were, uh, were recruited. And indeed, here is the original advertisement in the Buffalo News. It ran on March 23rd, 1997, and it, it was essentially looking for volunteers to participate in the Human Genome Project. The idea was to get 20 people to then choose 10 and mix their DNA together so that nobody could be identified, and then build a human reference sequence. Turns out things moved a little more, uh, in a slightly different direction, and I certainly have more time to go into that, uh, in the book if you're interested. But the reason I mention it is because one of the two people whose DNA makes up most of the genome was Dutch, and it turns out that there's a very important variant in the genome that determines your likelihood of blood clotting in the Dutch population. So just pause with that information for a moment, and I'll tell you about this family and this phone call I got. Basically, I got a call from John West, who's shown down here at the bottom. One day, it was in the aftermath of the paper around Steve Quake, and he said, he started telling me about this problem he'd had with a blood clot in his lung and that he treated it using warfarin, a blood thinner. You might have heard of that. And that he'd gone on to find that he'd had another blood clot, even though he was taking his medications exactly as he should. And I was very interested and compassionate about this, but I wasn't sure why he was calling me 'cause this sounded like a hematology problem. And I was getting ready to refer him to one of my colleagues in the blood service. But then he said, oh, you know, we spotted what you did with SeqWakes Genome, and I've just sequenced my entire family and I've been analyzing the data with my daughter who's in high school and I wondered if you could help us. And I paused for a moment because there, right there, is a kind of Stanford moment. And I had to rethink again. I got— so wait a minute, you know, at the moment we're still talking about 8 people in the world who've had their genome sequenced. You have sequenced your entire family, and you and your daughter who's in high school have been analyzing these genomes. And I just thought for some reason that I went to the poor science teacher who was looking after Anne West's AP biology class there when she brought 4 full genomes to school to analyze as part of her AP bio project. This is a very Silicon Valley thing. Turned out John West was the CEO of a company that was sold to another company, Illumina, that had developed the principal sequencing approach that still to this day is dominant for sequencing genomes. And they decided to sequence 10 people, and in that 10 were his 4 family members. Shown here. And indeed, he and Anne had been working using Microsoft Excel, believe it or not, which at the time could only hold about 200,000 rows, which is not enough to sequence a genome. Uh, shortly after, it went to several million. Anyway, this Wall Street Journal article talks about their pioneering work to the genome. But he wanted to talk to us because he had all four family members and wondered if we could help him analyze these genomes in context. And so we got the band back together, all the people who worked on Steve Quake's genome, And we built tools like this one. How do you map a whole family's DNA? So these are the 23 pairs of chromosomes. You don't see the Y there, but all the other ones shared by 4 family members. And every place you see a color is an area that maps to a particular inheritance, uh, mode from mother or father, as shown here. So pretty colorful, but the first time we'd ever been able to map exactly within a family which piece of DNA went to which individual in the family. We were also able to do things like, like this. This is the risk of coronary artery disease of all 4 family members. You tend to think your kids are going to be somewhere between the 2 of you, but here I'll tell you that the risk of coronary artery disease, heart attacks, for the mother and father is shown in the middle of the screen, and their son is shown on the left, lower risk. Their daughter is shown on the right, higher risk. But the thing about a genome is if you can do this for coronary artery disease, well, you can, you can do it for a whole bunch of other diseases— celiac disease, neuroblastoma, pancreatic cancer. If you can do it for that, you can do it for literally every single single disease we know about because now we have the power of data in our hands. So we published this paper as well, and it was well received. We actually formed a company at the time, very Stanford thing to do, but we were getting lots of calls at the time, actually, of people who were sequencing their genomes all of a sudden and were interested in interpretation. And obviously that's not something we could do in our Stanford labs. Personalis is now focused on immuno-oncology, specifically the idea of using the genome to help with cancer treatment and cancer therapy. But at the moment is the company, I think, that has sequenced the most genomes in the U.S., which, just to put this in perspective, these were the 10 to 15 in the world. Personalis has now sequenced 150,000, and worldwide we're getting close to more than a million genomes sequenced. Well, it's not just people who call up who have access— early access to technology. That benefit. And over the course of the next few years, we realized that having the genomic information could help all sorts of patients of all sorts of shapes and sizes, including newborn babies. And the baby I'd like to introduce you to now is one who came to our attention when she was 36 weeks of gestation, so she was still in her mother's womb, and it'd be noted that her heart rate was low. Sitting watching the seminar, imagine your heart rate's 70 or 80, 90 beats per minute, A little baby swimming around in a mother's womb, it should be more like 140. But her heart rate was 70, about half what it should have been. And so we transferred her to Stanford Children's Health and delivered her by cesarean section, where we realized that this wasn't heart block, which is one explanation for low heart rates in newborns, but rather Long QT syndrome. And that's an inherited syndrome that affects the electrical activation of the heart. This is the ECG, the electrocardiogram, and it doesn't allow the heart to reset after each beat., and because it can't reset properly, it essentially, they can't restart the next beat. And so the heartbeat is overall a bit slow. The problem is not that slow heart rate though. A baby's heart is perfectly compliant to deal with that slow heart rate. The problem is it predisposes to potentially fatal heart rhythms like the one shown in the middle. This is called torsades de pointes. The French word rather belies the fatal and potentially fatal nature of the rhythm, turning of the points, as you can see on the screen. This this particular rhythm is not associated with any output from the heart. And in fact, on her first day of her life in the outside world, this little one suffered 5 cardiac arrests. 5 times we went to the bedside to try and bring her heart back. She became the youngest baby in history to have a defibrillator implanted. This is her X-ray, and the defibrillator is this large device. In adults, we put it up here by the clavicle. But in babies there's no room there, so we put it in the abdomen and put the lead up to the heart to bring her heart back. But bringing her heart back is not an answer to the problem. And what is the problem? Well, we knew it was likely genetic, but at the time we're still talking 3 months to get a genetic test result back. We're still talking $5,000 to $10,000. But we knew there was a different way. We knew it was also possible to sequence genomes much quicker over the course of a few days., and if you could speed up your computational algorithms, you could potentially also make diagnoses over that period of time. And so we did that. We sent off her genome, and 3 days later, we got the data back in this packet that I took a picture of. We, over the weekend, speeded up our computational algorithm so that we could make a diagnosis as quickly as possible. And after we got the raw data back, it took us only about 18 hours to get to the answer, which was that this potassium channel here was responsible, along with this other channel shown here for the Long QT syndrome. What that meant was that we could then offer precision therapy. We could tell the bedside team, this is— use this medicine, not that one, because many of the medicines, especially little babies, have a very narrow window between side effects and therapeutic benefit. And so we were, for the first time, over the course of just a few days, able, using whole genome, to deliver the result to the bedside. I'm happy to say that she did really well. This is a picture of her 6 months later. I actually have a picture of her now 5 years on, celebrating her 5th birthday, and she is doing incredibly well. I think still with only a fairly minimal understanding of the extent to which technology helped her through those first few hours. Well, sometimes the answer is obvious and the question is what underlies that, and the genome can help. But other times the answer is not obvious and the disease is a true mystery. And so I'd like to tell you a story of, of one family, uh, that we got to meet through our Undiagnosed Diseases Network. This is a national, actually international network of doctors, medical detectives who attempt to use the genome and really any technology possible to solve medical mysteries. And we not uncommonly get patients who show up with stacks of medical notes this high. They've been to doctor to doctor to doctor over many years. With no answers. And this is called a medical odyssey. And it's called a medical odyssey over this famous story from Homer, The Odyssey, with a 10-year journey back to his betrothed after multiple shipwrecks and evading giants with one eye who were trying to eat his crew and strange animals and siren songs. There's a reason that we label what our patients go through medical The Medical Odyssey. I also like this picture. I put it on the slide because it says one of the 10 best books of 1996, which I thought was quite amusing. But then also even better on Amazon, you know, if you, if you're interested, you can follow the author, Homer, just in case Homer decides to release any more books. Seems a tad unlikely to me, but I wanted to tell you about these two boys, Carson and Chase. And as you can see, beautiful little boys, blonde hair, great sense of humor. Carson was born first, and essentially his dad realized that after a year he started to slip back in his developmental milestones and wasn't quite hitting them. And his dad realized that something was up, but it was really unclear and no doctor could really help. They went to a pediatrician first, of course, and then a regional pediatrician, a developmental specialist, and others. And when his little brother ended up going through exactly the same thing just a year or two later, they realized this was likely a genetic condition. And his little brother went through exactly the same thing, and they went to see then multiple neurologists because what happened was that they were unable to feed themselves anymore, they couldn't walk anymore. Their intellectual development was normal, but their motor development was impaired, and scans of their brain showed that the motor center was the one that was the focus of the problem. But none of those individuals unfortunately was able to help them. They'd been at that point to multiple doctors. And he would, Danny would spend time with their dad going to work during the day. The mom, Nikki, would take over the kids during the day. They would swap when he came back. After the kids went to bed, he would get on the internet for 4 more hours just trying to think somewhere on the internet, there must be somewhere he could click that could find an answer. He had no idea. Did this mean his boys were going to survive longer than the next year? They seemed to be deteriorating. Where, where would he go and what would they do and what was life going to look like? This is the torture of a medical odyssey and one of the reasons why we have this network of detectives. And he heard about us, the Undiagnosed Disease Network, and so he was referred to us after multiple years of really trying to find an answer. Just like I described for John West's family, now standard, we sequenced the entire family, the two boys and the mom and the dad. And this is the actual shot from the genome. Because as soon as we sequenced the 4 genomes— and he'd had studies quite similar to whole genomes in the past, but until we were able to do the whole genome on all 4 individuals, it wasn't clear what the answer was. But the answer became clear almost immediately we looked at the genome data. Here is a segment in this gene called MEKKR inherited from Nikki, his— the mom, to both boys shown here, an extra T in the genome, one letter change. But by itself in Nikki, no problem. She was— she had developed normally and was living a normal life with no problems. And exactly the same gene, a variant from her, the dad, from Danny, and had been inherited by both boys. Each of them had two hits in each copy. We all have two copies of each gene, one from mom, one from dad. In this case, both of these variants that each individual, the mother and father, had were not destructive in and of themselves because the other version was okay. But in both boys, they'd had knocked out this gene in both individuals, and it turned out this gene was really important for motor function in the basal ganglia, the part of the brain that deals with motor. And I'm not going to give you a quiz on this particular biochemistry pathway later, but think of this like a factory floor, and each one of these components is being passed along the factory floor. These are the workers who are here in gray —and then they get passed around this factory floor. They get to this point, and this is the gene that was hit, MEKKR. You get to this point in the factory, and of course you get a buildup of everything here because this gene doesn't work. And then downstream, you're missing all these really important things. And so Danny now had the information that he needed to, to be a bit more specific in his Googling. And so he started Googling for the gene and looking for the things that were on this pathway, just in case there was something, hoping that there might be something he could find. And in particular, one of the things that really is important that is missing when you miss this gene is something called lipoic acid. He thought to himself, well, if only we could replace that, then maybe the boys could do better. And if you're thinking that lipoic acid is potentially something really expensive, something that maybe drug companies charge half a million dollars for that you read about in the press for gene diseases, well, you'd be wrong because as Danny was checking the internet that night after we gave him the answer, he realized that alpha lipoic acid was something you could buy on Amazon for $16 a bottle. Part of the answer to the medical mystery of his boys' developmental problems was quite literally one click away on the internet. He could get it essentially delivered the next day, and this supplement came back the next day, and he started giving it to his boys. And they started improving. Now, this wasn't the cure. This can't reverse everything about what they do, but this is just to try to demonstrate to you the power of the genome. He's been giving them the supplement, obviously in collaboration with the neurologists, and they've been improving again. Their motor function started to improve slower than the normal, of course, but started to improve. And he was lucky. He was kind enough to share his story with CBS Morning News, which also featured the work of our entire network. In fact, this is just one case. We identified 31 new syndromes in the first 2 years of this network, solved 132 cases just like those ones, and we've continued to do that work over the last 4 to 5 years. In fact, one of my proudest moments, I think, was being featured in the last few years. There's been a lot of bad news, but we were featured at the top of This Week in Good News in the New York Times. And this is something that is very true. We, we answer about half of the We can solve about half of these referrals that are sent to us, but the other half, unfortunately, we still have to keep working on, and we never give up. So for the last few minutes, I wanted to talk about a different genome. This is actually the entire genome on the screen. These are the amino acid letters, and it's a bit small, smaller than the one I just mentioned. You have to spend months looking at the human genome, but this one's only 30,000 letters. In length. But don't let the size of this genome fool you, because size can, can fool you a lot. This is the genome of SARS-CoV-2, the virus that has wreaked havoc around the world, and that none of you on this call would have escaped, because it's everywhere. And as the pandemic began, we realized that, that we were in a very difficult spot. In fact, when we look back at what happened in the 1918 flu pandemic, many of the things are very similar. I mean, abs— you, you could sort of cookie cutter the things that were happening then to do with people not trusting science, people not wanting to mask, people not wanting to lock down. Very, very similar. One big difference: 100 years later, we have genomic technology. And so within a few days of the release of the genome of SARS-CoV-2 from the Chinese Health Authority, there were diagnostic tests being made around the world. It's the reason that we've all had things pushed up our nose multiple times over the last few years. But the power is the power of the genome, and the ability to sequence the human genome is exactly the same as sequencing— in fact, much easier to sequence the virus genome. This is the Stanford lab that produced one of the earliest genomic tests for SARS-CoV-2. Of course, now we use it even in wastewater to try and look to see where the infections are coming next. We in our lab at Stanford started to look at the genomes of both the virus and the individuals who were coming to the hospital, we realized that those traditionally underrepresented in the world are overrepresented in those who were coming to the hospital. It's what's shown here on the screen. But the age of genomic vaccines is what really blew everything away. And you're very familiar, of course, and you likely have received vaccines from Pfizer and Moderna. In the rest of the world, just as important, the AstraZeneca vaccine. And these— all these vaccines have had ups and downs, but we should not forget that none of this was preordained. The idea that we'd be able to, to get a vaccine approved in under a year was just something that nobody could think of. And the idea that that vaccine might be effective at the 95% level for any cold symptoms— this is not for death or dying, this is 95% for any cold symptoms in the trial, 100% for death or dying— we just don't have drugs that good. And although there's been a lot of bad news The idea that multiple vaccines were approved within a year at that level of efficacy is, is quite remarkable. And it's all because of the power of the genome, understanding the genome. These are RNA vaccines, as you know well, and the code that we understand for the virus allows us to, to move quickly. And the new Omicron variant vaccines are coming shortly. Even preventing the next one, also important for biosurveillance. Based on genomics. All right, the last little story I wanted to tell you just before I stop talking is we've talked about cost, but sometimes what you need is not cost or accuracy. I haven't talked much about that, but accuracy improved a lot in the last 10 years. Sometimes what you need is speed, and that was the case for the next patient I'll tell you about, Matthew. Matthew started to have a cough and cold symptoms in the first winter of COVID His parents and he obviously thought he had COVID, and so he went to his local hospital with fatigue and cough. But he realized quite soon after when we did this kind of scan of his heart shown here that actually he was suffering from heart failure. Your heart should normally contract significantly with each beat. And as you can see, this one is really just wobbling a little. This blue shown here is a valve in his heart leaking significantly and sending blood backwards instead of forwards. He was airlifted to Stanford Children's Hospital where he rapidly deteriorated into cardiogenic shock. That's a, a unsustainable disease and unsustainable moment in time. So he was crashed onto a heart-lung machine called ECMO, and finally, in fact, this surgical device was implanted into his chest where it is here in this photo. The battery for it he is wearing around his shoulder. So the question at that point, having stabilized him and have saved his life, was, well, what to do next? Because If he had a potentially reversible cause, like something called myocarditis, an inflammation of his heart, then he could potentially have that device removed and go back to having a completely normal life. But if he had a genetic heart disease, then he needed a heart transplant. So what could we do? How to make the transplant decision? Well, we could do a genome, but if we were to do that, it would likely take some time. Certainly back in the day, as I mentioned already, There was a 50-hour genome. We did a 42-hour genome with the little baby that I showed you earlier. And in the interim, Stephen Kingsmore, a colleague of mine from San Diego, had essentially won a Guinness World Record for sequencing a genome and making a diagnosis in as little as 19.5 hours. That sounded more like it. But we had even wondered if we could go faster. And the problem was that if you sent a genome to a standard commercial organization, It could still take 56 days. Even the rapid ones were 7 days or 2 days. We wanted to go faster. So I pitched this to some of my students. This has just been the last year now, and they set to work. And the key to it was this machine called Prometheon, a way of sequencing an entire genome potentially as little as 1 to 2 hours. So we talked to the company Oxford Nanopore. We also pulled together folks from Google Health, UC Santa Cruz, Baylor College of Medicine, and NVIDIA to see if we could do this faster than anyone else had. We called it ultra-rapid. So we extracted the DNA, we sequenced on this Nanopore machine, we did all this computational work you'll see in this box, we prioritized variants and tried to make a diagnosis. Well, Matthew was the second call that we got. So let's just take you through what that took. Here's the initial part in the wet lab, the green part. That took 4.2 hours. We'll put that on the timer against the world record of 19.5. We move to the sequencing part. Let's add that on, 6.9 hours. Now we have an entire genome sequenced and the raw data ready. Well, then we need to turn it into A's, T's, G's, and C's and start to work out how to put the jigsaw together of his genome. 3.6 hours. We're now up to 10.5 hours, and we have the variant call file. The problem with the variant file is it's 4.5 million lines long. We have to get that down to one letter change that might be causing this disease. Well, our filtering worked fairly quickly, and in 46 minutes later, we had narrowed that 4.5 million variant list down to one variant in a gene called troponin T, a critical part of the heart muscle motor that turned out to be the clear cause of his condition. And in, but in doing that, we made the diagnosis in just over 11 hours, breaking the world record by about 8 hours. As a result of this information, he was urgently listed for transplant and received a new heart 21 days later. 6 months after that, this is him in front of the machine that sequenced his genome, smiling and doing incredibly well, now almost a year later. So we can add our new approach here, Stanford Ultra Rapid Genome, and most of the genomes in the end were done in under 8 hours from blood draw to diagnosis, faster than we'd ever seen. And to give you a graph like the one I showed at the beginning, Steve Quakes genome in the middle, 5 days at the time was as fast as it could go, and here are new ones down here in under 8 hours. We were recognized by Guinness, it turns out there's a Guinness World Record for speedy genome sequencing, and this is the first time the entire team got together in 3 dimensions for the Guinness World Record ceremony. Well, to do that you have to do a number of things. I came into my office one day and I saw this, I wondered what's that pink note there, and it turned out that it was John's sleeping bag, because to be able to do genomes this fast, sometimes you have to sleep in the lab. And in case you're wondering, this is how good it looks to make a diagnosis at 10 or 11 PM at night, as fast as anyone ever has in the world before. So maybe we need to update the Ferrari slide. If we do that, then the top speed is 199 miles per hour. So you could get a 1-cent Ferrari, just how fast would it go? If we speeded up this Ferrari as much as we speeded up genome sequencing? Well, 32 million miles per hour. I was rather hopeful that would be warp speed. Unfortunately, we're still a little short of warp speed, so we still have some room to go. If you like these kind of stories, I mentioned the book earlier. The book is really full of these kind of patient stories and the sorts of difference we try to make to our patients' lives using the genome. They let me read the audiobook. Believe it or not, I did have to do an audition to make sure they could understand my accent. I hope you've been able to understand me talking today. I passed the audition, and so if you're interested in audiobooks, it's there. In fact, the paperback, which will look like this, is out early next year. Thank you so much for joining me today, and I'll be very happy to answer some questions in our remaining time.

Thank you, Dr. Ashish, for such a fantastic and hopeful presentation. I'm going to jump into questions from alumni. From Alicia, it seems like personalized medicine has gotten more complicated. The epigenome and the transcriptome, can you talk about how those factor into precision medicine?

Absolutely. So we're in a world of multi-omics. So omics is something that will be familiar to some of you, obviously to the question writer there. There are multiple omics and we call them omics because the end of the word is omics. So genome, transcript, omic, the transcript is the RNA, the message, uh, proteomic for proteins, epigenomic for the parts of, uh, right, that are around the DNA that determine whether it's read or not. So we have lots of omics now. Metabolomics is another one, the metabolism. And it really is just a way of saying that things that we used to measure in ones and twos, one gene, one metabolic factor, we can now measure hundreds, thousands, millions, or even billions of them. And you're right that the world of precision medicine has gotten significantly more complex because of multi-omics. And it's exciting. And I think the key though for us is, are we making a difference to patients' lives? Are we making diagnosis and then getting treatments to patients who didn't formally have them? And I'm happy to say the answer is yes. Are we costing the healthcare system a lot more money? I didn't show you this data here, but no, this is an example of a technology that not only saves lives, but saves money. This is— and you can imagine that could happen because the tests that we normally do are very expensive. This genome is now, you know, $500 for the reagents, basically just a few thousand dollars for the entire diagnosis. And if you can interrupt a diagnostic odyssey— in fact, we did the calculations— 95% of healthcare costs could be saved. So this is something that's highly cost-effective. And so even though it's highly complex and only, as you point out, getting more complex, we are up to the challenge of integrating all that data together, helping our patients and saving lives and money.

Fantastic. Here's a question from Nelson. Since price for genome sequencing has decreased so dramatically, do you expect in 5 to 10 years it'll be done to all babies born in the US and in the world to detect life-threatening diseases?

Yeah, I think that's a really great question. And I should say, in fact, that the genome price had dropped further. In fact, that $0.01 price I showed you on that slide is the most up-to-date version. In fact, just in the last few months, a couple of companies have announced $100 genomes. And so the $100 genome is the $0.01 Ferrari. Pardon me, we don't have that ready. If you walk into Stanford Hospital today, you won't get a $100 genome, but it's coming. And if you think about this amount of information and if you think about the maturity of our computational pipelines to interpret that data, then we are talking about a technology that I believe should be available to anybody who wants it. Certainly true that the technology used for our neonatal screening, which is almost universal in the US, is technology that is decades old, and we're talking with the genome about a technology that's much more powerful. I think it's likely that that will happen and that will be made available, and that it may even be a direct-to-consumer sort of situation. But of course, I currently believe that most of the interpretation of the genome is that it's best to happen within the context of your healthcare system, because this is complicated data. And especially when we think about doing that on newborns who are not old enough, of course, to give consent, that, that requires some careful thought, because there's information in the genome about diseases that might be relevant 40 years from that moment. Diseases of late onset, Alzheimer's disease, for example, we can look at genetic variants for that But if our aim in screening, especially newborn screening, is to avoid really big significant diseases and to do that with the most power that we can, then I do believe it should be made available and people should have that choice.

Here's a question from Todd. What do you think of proactive value of genome sequencing? Since we're discovering and validating a rapidly increasing number of genetic contributors to disease, monogenic, PRS, etc., and many of these have established interventions, Won't this become one of the most important ways to improve human health?

Yeah, thanks for that question, and I do agree with you, actually. So I focused on the last 10 years, basically, of history of our use of the genome, which by nature has focused on those diseases with the highest impact on an individual basis. And that wasn't necessarily something I would have guessed was going to happen. When we did the original Steve Quake genome, we gave 3 examples. One was, of course, the ruling out of Steve inheriting the variant that caused sudden death in his cousin, or rather that his genome didn't have obvious variants for that, um, cousin's son. But the other 2 things we did, and the ones I showed on the screen, were actually more about preventive genomics. And in fact, over the course of the next 10 years, they have not become mainstream, but they're getting there. You mentioned PRS, which for everyone else stands for polygenic risk score. And I did give some examples of that with the John West family situation, where you can use the genome to basically give a modified risk of, of every single disease, because every disease has some genetic component. And so I do think that the future of preventive care will have genomics absolutely at the center. In fact, just this year at Stanford— in fact, in the next few months— we're about to introduce an integrated test for the risk of heart attack that includes not just the things you're used to your doctor asking about, blood pressure and diabetes and your family history, but includes actually your genetic risk and includes that integrated together to give a much more precise estimate of your risk of having a heart attack, which allows us much more precision in deciding whether you should have a cholesterol medicine. And so I completely agree, whether it's for cancer, cardiovascular disease, or another 3,000 diseases, the genome will be at the center of that as we move forward.

This leads into our question from Keith. So genome can be used to detect disease, but how about it to use to detect good health? Now, some people do not get affected with HIV or COVID, but what can we learn from their genome?

Yeah, and I touched just quickly on that towards the end, the work that we've done looking at people affected with COVID and we are part of a global consortium. One of the amazing things about COVID— and there are plenty of silver linings as well as the challenges we've all been through— one of the silver linings is that there was a global response in the scientific community, the genomics community, the vaccine community. I mean, 2 billion of those doses that came from AstraZeneca were delivered at cost, and I should mention I have a relationship there with AstraZeneca, but But that's not something that's business as usual. This was a situation where everybody was all hands on deck. And to be able to say 2 billion doses to the world, 2 billion doses of any medicine getting out to the world is pretty incredible to do that. The cost is also great. So there were some first things that were happening that were really welcome. One of them in our domain was the sharing of genomic data about predisposition to COVID. And as the questioner mentions, the predisposition to multiple diseases, including infectious diseases, is coded in our genome. But until we get large enough amounts of data— and we contributed a few thousand data points to that, but the overall numbers were much more like hundreds of thousands and probably getting close to millions now, of course, for the genomic study— we start to see signals that we wouldn't see in the smaller data set. So that kind of global collaboration, if not new, was certainly scaled up during COVID And so I think it's going to be part of our lives going forward, and our medical record systems are going to have to work out how to integrate that information. At the moment, they're very good at— well, they're very good at billing, they're very good at following up information that's in a data format, but they're not really built for genomic-level information. But what should happen is when you go to your doctor for your next prescription, you know, that your medical record should look up your genome and advise the doctor what the best medicine is for you. And that's truly the future that we're working towards, and we have some movement towards that also at Stanford, I hope in the next year or two.

We have a question from Jerry: How private is the gene sequencing? Is it kept from insurance or long-term healthcare insurance companies?

Yeah, really important question. Data privacy is something that we've been focused on from the beginning. In fact, when we published the first paper about Steve Quakes, you know, we worked very closely with and Grevely, of course, a world-famous ethicist in the biomedical sphere from Stanford. And he was involved, in fact, in the planning even before we released any information back to Steve and really helped us think through how to do that, both from a consent perspective, an ethics perspective, but also as we began at that moment to think data privacy. On the ethics side, of course, whenever you sequence your genome, you are by definition sequencing half the genome of any of your first-degree relatives. You don't know which half, and they don't know which half, but you share 50% approximately of your genome with any first-degree relative—your mother, father, brothers, kids—and so it's a very interesting situation. And so that's an ethical dilemma. Another is that if you're being sequenced, let's say, for a cardiovascular disease question, well, that—you're still going to get information in it about neurological disease, about cancer, about all sorts of other diseases, and How do we deal with that? So we built some tools to help people decide what they wanted to know, what they didn't want to know. And then finally, on data privacy, that's moved along significantly. I think back to the early days of the internet when really people were hesitant to even put their credit cards into the internet. Some people still are, and maybe that's reasonable, but we've definitely moved to the point where we have much better processes. And we think of your genomic data as absolutely as secure and as private as any of your medical record data. It's, of course, identifiable in a way that some are not, although someone would have to really connect dots from a whole bunch of places to really get down to identifying it as you. But it is not a number like a cholesterol number is not identifiable because lots of people could have that number. A genome is ultimately identifiable. And so we're very, very careful with data and privacy. Thanks for the question.

A question from Karen. I was fascinated by the recent article in Stanford Magazine about the work of Michael Snyder and his biomedical tracking and comparing it to the genome. Are you guys collaborating on that work?

Yeah, in fact, we sequenced him. He wanted to be sequenced and he probably, if you read the article, he had sequenced himself. We helped analyze his genome and he talks a lot about the impact of the genome sequencing on his medical care, in particular his diagnosis of diabetes. We have a few other things that we also want him to focus on as a result of the genome analysis. We and he also both work on digital data, so we collaborate a lot. Mike is a good friend, and we both live here on Stanford campus, so we see each other all around all the time. Yeah, I think I didn't talk about it today, but the integration of all these different high-throughput data sources. So just to give you one example that we've been working on, you know, your watch now, whether it's whichever, from whoever you get it, Fitbit or Apple, or it could be any number of others, Garmin, there's plenty out there. Increasingly is focused on your health, whether it's diagnosing a rhythm like atrial fibrillation, where the top of your heart just wobbles instead of beating properly. Your watch can potentially diagnose it for you. And at Stanford, of course, we've worked closely with Apple in helping them develop that technology and giving them medical advice about it. But I like to think about, well, what if we integrate your risk of atrial fibrillation from your genome with the algorithm that's actually on your watch? Because if you have a low risk of atrial of atrial fibrillation, then maybe it shouldn't be tuned to be quite sensitive. If you have a high risk, maybe it should be tuned a little bit more sensitively. So when we think about that personalized medicine, a day of personalized medicine where these different technologies are integrated together and whether it's in a secure cloud or actually on your wrist, then that information comes together to give you the best possible diagnostic to prevent these situations occurring before they do.

We have time for 2 more questions. From Kathy Shaw, has any progress been made on the causes and treatment of non-genetic ataxia?

Well, ataxia, which means the inability to walk in a coordinated manner, is something that is one of the most common presenting symptoms for our patients with mystery diseases. And it's certainly true that, in fact, neurological symptoms of which ataxia is one, are among the most common symptoms in that group. So in fact, the chief of pediatric neurology at Stanford, Paul Fisher, is a part, very significant part of our undiagnosed diseases network because so many of the little ones present with seizure disorder with other neurological challenges. And we've been very successful in that group in diagnosing them. Now the question was asking about non-genetic ataxia. I certainly would have to refer you to neurologist to talk about that. The genetic aspects, certainly we spend a lot of time thinking about. Acquired causes of ataxia are multiple, but they're very different. And I think many particular, I mean, every disease has some genetic components. So even acquired diseases that we think of, like we just talked about infectious disease, that's by definition acquired in your life, but your predisposition still has a genetic component to it. I think that's probably all I would want to say about non-genetic causes of ataxia here.

We have time for one quick question. What is the easiest way to find who can sequence my genome based upon location zip code? Some of us don't live near big cities.

Yeah, and so it's a great question. It's certainly one that I've been answering, of course, since we first started doing this work. Over the course of the decade, it has been possible at various times to just basically get on the internet to a website and order up your genome and they'll send it back to you. None of those companies actually have survived, and it turns out there's not enough people willing to pay enough money to have their genome sequenced for that to be the case. So genomics has really had its place within the healthcare system, triggered at the moment really by a presenting diagnosis. The program I mentioned at Stanford that's coming up where we integrate your polygenic risk, so your multiple risk across the genome, with your cardiovascular risk for heart attack, that's one of those situations where you might go to a primary care doctor or your cardiologist and they'll send a test. You don't need to have a rare or mysterious disease. But there aren't many examples of that. There's a clinic like that in Boston, but not too many others. There's a company called Invitae that works out of San Francisco that does this kind of thing, but again, not yet direct-to-consumer. You might have heard of 23andMe, another local company, and they do some degree of genetic testing through a much more limited platform. To give you an idea, they have a few million spots in your genome. If you remember, I said there's 6 billion total. That's what you get with the genome, and they have a few million. So it's— but it can tell you some things about ancestry in particular and some medical conditions. Yeah, at the moment, there isn't an easy way to do that. But if, but through the healthcare system, I think is the most likely way.

I'd certainly chat to your medical professionals about it. Well, Euan, thank you for such a fantastic presentation today. Thank you for having me.