Dr. Wendy K. Chung - Putting the Pieces Together from Head to Toe: Planning for Pediatric Genomic Medicine

All right, good morning, everyone. Um, excuse me. Today I have the distinct honor of introducing Doctor Wendy Chung. Doctor Chung is a clinical and molecular geneticist and the chief of the department of pediatrics here at Boston Children's and at Harvard Medical School. She, uh, received her degree in biochemistry from Cornell, then went on to earn her PhD at in genetics at Rockefeller University, and her medical degree at Cornell University. She did pediatrics training at Columbia Presbyterian Medical Center, as well as her fellowships in clinical and molecular genetics. She stayed on after those fellowships as faculty at Columbia, where she built a truly impressive career, uh, prior to coming to join us recently here in Boston. Doctor Chung directs NIH funded research programs in human genetics of pulmonary hypertension, breast cancer, obesity, diabetes, autism, and birth defects, including a lot of the conditions that we treat, including congenital diaphragmatic hernia, uh, esophageal atresia, and congenital heart disease. She is a national leader in the ethical, legal, and social implications of genomics. She has won countless awards in recognition for her research, including the Rare Impact Award from the National Organization of Rare Disorders, and she's a member of the National Academy of Medicine and the Association of American Physicians. In case her research achievements were not impressive enough, her commitment to education and her reputation as an educator is truly second to none. She's organized and directed, uh, lectures and courses for trainees of all levels, and notably, she directed Columbia's core pre-clinical course in genetics for over 20 years, uh, which is where I had the great honor of taking that course and meeting Doctor Cheng, which I'm sure Doctor Cheng remembers vividly from many years ago. Um, It was evident to me in that course why she has won so many awards for her teaching. I'm truly engaging, uh, and passionate as an educator. Her many awards include 7 lecturer or Teacher of the Year awards voted on by medical students at Columbia. And in fact, the Department of Pediatrics at Columbia gave her the award for the best grand rounds of the year. I didn't realize that when I invited her, but it has me even more excited for what we're going to listen to for the next hour. So, please join me in welcoming Doctor Wendy Chung. Thank you. I'm thrilled to be here this morning and uh as I go through, uh, I wanna extend an open invitation to either you or your friends, uh, and as we talk about some of the things that I've been doing and what we're doing here at Children's and being able to be part of this in some way. Um, very open invitation in the Department of Pediatrics. Specifically, we've been doing a lot of strategic planning for what we're gonna think about the future of precision medicine, genomic medicine. And although it started in the Department of Pediatrics, I hope it'll be extending across the hospital as we do things and, and we very much wanna have input from everyone and, and, uh, are open to ideas. So, um, I started this journey, uh, as I'll tell you about 20 years ago and thinking about one specific, uh, congenital anomaly, uh, diaphragmatic hernia, and I grew and I learned and, in part, the fields grew and I'm gonna sort of share that experience and that journey with you, but I've realized that the pieces are all connected, um, and perhaps that's not surprising, but it took me a, a while to realize that and I'll tell you about some of the fallacies I learned as a resident and how I'm now just undoing some of what I learned then. So, uh, many of you know congenital diaphragmatic hernia from the surgical side of things, and I'll try and share as I think about things. So, this anomaly is, uh, relatively common and increasingly, we see it obviously prenatally, um, and we had a program at Columbia called Prenatal Pediatrics, and we would sit around the room and measure the LHR and try and make predictions based on what we saw with compression of the lungs about how that baby would do after. Birth. And then we'd come back after birth and see how right we were or more often, how wrong we were, uh, in terms of those predictions. And, um, within this, although we thought we got better, we still had a relatively high mortality rate, um, despite ECMO and incredible surgeons and great pulmonary hypertension docs. And so, from the time I was a resident, I tried to understand what was the difference in terms of who was gonna do well and who wasn't gonna do well. And as I said, The markers we had in terms of measurements or anything else we could tell based on anatomy just wasn't doing it for us. Um, in doing this, uh, many of the prenatal patients would come to me and say, uh, we see this, uh, anomaly in ultrasound or an MRI. Um, Doc, tell me what my child's gonna be like in 20 years. Are they gonna live independently? Are they gonna go to college? Are they gonna grow up and have a family of their own? And with this, it was very hard to tell them. Um, so not only based on whether they'd have pulmonary hypertension or live, but what they'd be from a neurocognitive point of view. And I got frustrated with this. They were frustrated with it, I was frustrated with it, and so that's been part of my goal is to try and understand that better and to do better. I still remember the first time I was on rounds in the NICU and my attending said, uh, was explaining to me the physiology, uh, associated with this, and they had a very, uh, what seemed to be very logical explanation. In utero, the uh intestinal contents, abdominal contents are simply pushed up into the thorax compressing the lungs and developmentally, the lungs can't mature. They can't develop fully, uh, just because of the compressive effect. I thought, OK, that makes sense. Uh, cer certainly, if we could do something in terms of decompressing, maybe that would, uh, be the answer. And so within that, many have thought about even fetal or prenatal surgery to be able to, uh, decompress the, the chest or other ways, fetal now in terms of being able to allow for lung expansion. Um, but I'll sort of cut to the punchline, which is that I don't think the anatomy is going to show us everything and I don't think the genomics yet shows us everything, but it is showing us a new dimension of information. So, as I said, uh, CDH in the middle of this circle, there are many things that can be associated, not all things in all people, uh, and I would argue you could substitute almost any other birth defect or congenital anomaly in the middle of this. So whether it's esophageal atresia, congenital heart disease, uh, genital urinary mal. Within this, there's just a wide spectrum of associated features. And in trying to understand the future, we really need greater specificity. More data is one of the things that I'm going to say and, and within this, I do think we have some exciting opportunities for more data to be able to predict the prognosis. So the hypothesis I had uh after the first few cases of understanding this was maybe, just maybe, as we think about the primary etiology of what's causing the developmental diaphragm disorder, maybe there's a gene driving some of the cases. Maybe it's not all of the cases, but some gene driving some of the cases, and that gene may affect not just the development of the diaphragm, but may affect development of other parts of the body. And I will say, and I'll challenge all of us, that when we think about how we visualize the anatomy of the body, I don't know that we actually visualize everything and there may be subtle anatomical differences that we're not picking up. And with increasing resolution, with uh different imaging modalities, we may be able to do that, or some of them may be so small, they may be at the cellular level. We may Never visualize them with MRI, but we may be able to see these at a functional level. And so, maybe there's one gene, but it's driving or affecting multiple things in the body. So, we have a term for that, pleiotropy, but the idea being that with that, there may be individuals who have other related conditions where it's one gene, many manifestations. Many of those manifestations may affect the outcome. So as we started this, uh, it was at Columbia that I started something called the Dream Study. Um, in parallel, Pat Donahoe, as, uh, many of you have worked with Pat over the years, had been studying CDH. Pat and I, and Pat, I don't know if you're out there, but Pat and I have had a great collaboration, uh, and have joined forces, and especially now that I'm in Boston, have been joining forces to put this together. Um, one of the wonderful things about The dreams cohort is that we've continued to follow these children over time and so, uh, thankfully I'm getting older and the kids are getting older and so we're now just seeing our first, uh, little ones now becoming adults and being able to understand their outcome and that ends up being really critical to understand not just outcome in terms of time to discharge or whether or not really. Over the life course, how they do. So this requires a large team effort. Um, the sites are here. If any of you have other ideas, uh, in terms of how to expand the network, uh, we truly are limited by numbers at this point. So, let me explain one of the concepts that go with this. Um, so when we think about the era before you do the amazing things that you do, individuals that were born with CDH didn't make it. They, they really didn't live beyond the neonatal period. Without early diagnosis and early treatment, they died. And so when you think about the genetics for those individuals, they didn't live long enough to pass down their genes to the next generation. So from a population point of view, how do the, where do those genes come from? If they die at every generation essentially. Uh, they have to be born at each generation to be seen. And so one of the things we realized is that these are what we call de novo mutations. That is, they're new in the child. They're not passed down in many cases. I will get back to this. They are in some cases, but in many cases, they're not passed down. And in fact, the ones with the most severe mortality are those de novo mutations because they arise new each generation because they don't make it to be transmitted. So once we had the technology to be able to look at the genome in a comprehensive way, we could do a relatively simple experiment. We could take a sample from mom, dad, and then the baby with CDH and do a direct comparison, nucleotide by nucleotide across the 3 billion alphabet letters, and we could see in the baby which one is different from mom and dad. And that's a very limited way to focus your information. Out of 3 billion letters, each of us, and I've done the experiment myself on my own kids, each of us have about, in our children, about 100 to 200 nucleotides that are different in a child. And of that, on average, 1 to 3 affects the sequence of a gene that affects the protein. So now you go from a search base of 3 billion letters down to a very narrow search base of a few different letters, and that becomes something that at least my head can compute on. So we started doing that experiment within the dreams cohort and realized that in fact there were individuals. Some of them had large genetic changes. These are things like whole chromosomes. Some of them had medium-sized genetic changes on the order of a few megabases, and some of them had single nucleotide, one single base pair out of 3 billion letters that was responsible for their CDH. And we could now start to inventory those. Another thing that you may or may not think very much about if you think through your own patients with CDH is it's not an equal opportunity condition. And this is true actually of most congenital anomalies. We see a few more males than females. It's not that it's, you know, 4 to 1 like it is with autism, but we do see more males than females. And in general, the idea is that there's some protective. Factor in, in females. I can't tell you exactly what it is, but that there may be something protecting those females. And so, with that protection, we actually see that females with some of these anomalies are more likely to have a stronger genetic factor because they're protected, it takes something more to push them over the edge. So I'll be coming back to this in a second. So, let me just take you through the journey as I was seeing things over time. So, as I said, uh, you can see big whole chromosomes in some cases, trisomy 13 or 18, and then in some cases, we could see sort of medium-sized portions of chromosomes. And one of the chromosomal regions that we see most commonly is chromosome 8. Uh, the short arm called 2-3. And within this, we could map out those deletions. These were parts of chromosome 8 that were missing and we could map out the deletions across different people. Not everyone had the same deletion, but you could map the part that was con uh conserved or consistent across everyone, and then you could inventory the genes that are within that conserved region. And in fact, they're true transcription factors, SOC 7 and GATA 4 that were consistently in all individuals with those deletions. And so that gave us some idea that those would be important. We then drilled into GAA 41 of those transcription factors. The GAA genes are very popular here in hematology. We have a long history of studying those. Uh, and for this one in particular, we knew very precisely what we call the DNA binding regions. The region of that transcription factor that grabs onto the DNA that's very highly conserved and very specific for function. And so we could map all of those out and we could inventory then and start to identify nucleotide by nucleotide in that gene, what were the individuals that had these de novo variants within that one gene and in fact found several of them. Within this, this transcription factor, we could look at mice, and I'll get back to the mice in a second, but we could look at them and we could knock out that gene in the mouse and prove to ourselves that in fact, that was, was really causative. It wasn't just coincidence because we could see the same diaphragmatic defect in the mouse. As we were doing that, I started looking at some of our familial cases. We didn't have very many, but we did have some familial cases, and this was um one of the families that always stuck in my head. Um, down here on the bottom was the baby that we had been taking care of, and his family history was having had a paternal uncle who had CDH. He, he died and we didn't have a sample from him, but we knew that it was running in the family. But interesting that it kind of skipped a generation. As we were looking at GAA 4 across all of our individuals, we saw one of these genetic variants, again, at one particular amino acid, and it was one of these that is shown right up here. So, it was one of these that was highly conserved that I predicted it should have been associated with CDH. But it struck me as very odd that it skipped a generation because I was, again, growing up in this thing. CDH is pretty bad. This is a diagnosis you're not gonna miss. How could it be that it skips a generation? We went through the family, very carefully, looked at everyone in the family, and within this, realized that it was coming from the right side of the family. His father had the same genetic variant, his grandfather also had the same genetic variant. But as I said, it didn't really seem to compute. So I'll never forget this. Um, this is the good thing about going to Disney World in the long lines. I'm sitting in Disney World with my kids, sitting in the line for Pirates of the Caribbean, uh, interminably waiting to get on the ride, and I'm daydreaming about this family, and it just hits me. Did I ever really look for a diaphragm defect in the father or the grandfather? Maybe there was something there and we just didn't really look the right way and look hard enough. So I immediately call up. Uh, my lab and say we've got to look at the parents. So we go, uh, and arrange to look at the parent, at the father and the grandfather now, and again, asymptomatic and it's subtle, but on MRI you can see a small defect here, the same, uh, see at the same location here and the grandfather's small here and small here. But again, it's subclinical. They've never been symptomatic, they've never needed surgery, but in fact, it was there. Now that we've gone back and in other similar families realize this, we see all sorts of diaphragmatic defects that are subclinical. They never come to clinical attention. They may have not actually defects in the sense of having a hole, but we can see places where the muscle is thinner, we can see waviness of the diaphragm. But other things that if you don't look, you won't see. So, the reason I plant the seed is I'm convinced that there are other things with the way the body develops that are all over the place that we simply never come to clinical attention. But if one wanted to and start putting people through imaging to be able to see it, there are all sorts of things that we would see. So within this now comes, uh, just to account for you, de novo genetic variants accounting for some, inherited variants counting for some others. And as we were looking for this, there was always this idea that there might be both of those rare, by the way, so all of these genetic variants we rarely see in the general population. But this other thinking that maybe there were also common variants that might determine some of that probability. And so as we did this, started powering up and uh right now, we've got about 1500 cases of CDH fully genotype to do this, to be able to understand this. So we took the entire cohort, started looking at all these genetic variants, and also now start to looking, look at clinical outcomes. So now that we can account for some of these genes, can we now start to understand why we have the difference in how those kids do over time. So within this, just a little bit of terminology, these are all these de novo variants. We try and, uh, we can't recognize every single gene, but we start recognizing some genes that clearly disrupt the genes. We call these likely gene disruptive. There are some other ones that are much subtler. They change single amino acids, and those are called missense variants, but we can now start to see some of the signal from the noise. And still, I'll point out that these are modest numbers. The population attributable risk, still only 13% of all of CDH for one class, 7% for the other class, so we're still not accounting for the majority, but we're trying to see enough of the signal without getting too much noise in the background. So we're being very, very sensitive in terms of who we're letting into the club. OK. When we look at this and we look at all of CDH in terms of how it contributes, you can see for some, uh, in red are the most severe types of variants, purples, the sort of intermediate, and as a control, what we think of as silent genetic variants, uh, just as a sanity check to make sure it works. So, the first comparison on the left, all of CDH. The next comparison is now CDH that's associated with something else. So that something else could be a second congenital anomaly, it could be over time, neurodevelopmental issues, it could be growth issues. Now, we start to see some of the signal come out, those likely gene disruptive variants enriched in the cases that are having other associated features and worse outcomes. We can now start to take isolated in separation and we still see some of the signal there, the ones that look isolated and some of these look isolated but over time grow into their syndromes and we realized they actually weren't isolated. They may start to have issues once they get into school, problems with attention, problems with autism, and so these emerge over time. So this cohort constantly evolving as we're doing it. Um, and as we do it in terms of population attributable risk, again, we're now accounting for about 15, 20% of all cases as we're able to do this. Ultimately though, it's not enough to look at that collapse together. You wanna be able to look at individual genes and be able to then provide an individual prognosis to an individual family as we're doing this. So, we start to identify individual genes associated with this. MYRF is one of the genes we saw initially. Um, I'll just tell you that in1P1, 2 of the genes we've studied most. Deeply, unfortunately, very, very bad prognosis with both of those. And so when we see that, and increasingly we now see that prenatally through studies that we call prenatal seek, we unfortunately know that those are gonna be bad actors, bad prognosis, and we're always right when we see those. Um, as we go through and do that, we can, uh, also, uh, um, as I said, be able to think. About what it is we know and what we don't know. So, we do an accounting to see of the CDH genes that we know over here, how much of the signal are we accounting for, sort of a tally, how much work has to be done. Uh, what you can see here is we know very little. There's very little in terms of the known CDH genes that account for the signal we're seeing. We have a lot of work to do. I can't retire yet. Um, when we go through and we add some of the genes that are neurodevelopmental genes, genes that people have characteristically thought were not associated with CDH necessarily, but associated mostly with brain development, interestingly enough, there's more of the signal that's accountable for brain developmental genes than what we knew to be diaphragm developmental genes, and I'll come back to that. But that still accounts for some of the signal but still not enough. So we're still not there, there's still more work to be done. Interestingly enough, and I have to say I was a naysayer of myself, I assumed that to look at common genetic variants, we would need, I don't know, 50,000 cases. Based on a lot of other, what we call genome-wide association studies. I thought it was gonna be hopeless that we'd ever start to see any of that signal. But we went ahead and we did it, uh, based on some of my younger faculty, uh, sort of spurring me on. And as we did it, in fact, we have now identified, uh, just recently published. Two common genetic variants, which in an interesting way don't in and of themselves determine whether someone has CDH but they add to the picture. So, you'll remember back to that G family that I told you about. It's probably not just one genetic factor that's determining this. It's also this background of other contributing factors as well, and this is part of this choir in the background that's contributing to whether or not someone has this signal. So this is a complicated way of looking at this, but we wanted to understand now for the constellation of common genetic variants contributing, what types of CDH cases was it contributing to? Those that were isolated, those that were complex, those that were male, those that were female, those that had the single genetic factors. And this is a violin plot and these are the controls, and this is the distribution and the, the thing to look at is this dotted red line. That's sort of what the general population is. And what you can see is shifted up for every single one of the violin plots by an equal amount are these common genetic factors. So these common genetic factors are in fact contributing to everyone, not to simply the isolated versus the complex, not just males versus females, not just these large genetic factors. For everyone, it's actually increasing or it's decreasing the threshold to liability, making it more probable that someone will develop CDH. So within this, the analogy I like to use is it's like thinking about you've got a cup with ice and water in it, and at some point, the cup overflows. And how can that cup overflow? You can have large ice cubes that are sort of rising, uh, increasing the water level. Those can be those strong de novo genetic factors. You can have some medium-sized ice cubes. Those in fact are the ones that are inherited, those inherited rare variants, they aren't quite as strong, or you can have these common genetic factors, small ice cubes, and it's the combination of all of those together in any one person determining that risk or that probability of that congenital anomaly. So, with this, we wanted to understand, as I said, ultimately, the clinical outcomes. So we've done a series of papers looking at that. The first one that's easy enough to tell is just mortality. So, associated with mortality and we can account for many other factors. So, as we do these types of analysis, we include the genes, we include the size of the diaphragm, whether it's on the left or the right, uh, we We can look at, uh, uh, uh, sort of, um, gestational age. Um, we can look at social determinants of health. We can put all the factors in together that we know to account for this. When we do this and we pull the genetic factors out, so we control for the other factors and look at the genetic factors, we always see that complex cases have worse outcomes. That's not surprising. But if you look at the differential between those that collectively look like they have something genetic, some strong genetic factor going on, we see a higher frequency in those with the genetic factor amongst complex cases, but still, all complex cases, they have a worse outcome. What's most interesting to me is among the cases that we think are isolated, that's where we see most of the differential. That what's predicting for us those genetic factors is to me, those are probably not as isolated as we think they're going to be. It's simply that at the time we're looking at that child, we're missing something. We're missing a piece of the puzzle and it's likely that that child is gonna grow into it over time if they live to do that. So, we can do this and we can look at other factors that also contribute to mortality. Pulmonary hypertension is a favorite of mine in particular, and again, when we look at this, complex cases have greater difficulty. But again, the isolated cases, the differential that we see in terms of pulmonary hypertension, and this is with an echo core looking at all the echoes of these kiddos, uh, over time, the isolated cases are where we see the genetic signal pushing in terms of the pulmonary hypertension. Also, because we look at these kids over time, we can look at their Bailey's, their Vinelands, their developmental assessments to look how they're doing in terms of how their brain is functioning over time and again, accounting for many other factors in that analysis, we can see that the genetics, again, pushing in terms of what that outcome is and it's not simply um factors in terms of how long they were in the hospital, whether or not they were on ECMO. Some of those factors are important, but genetics also an important factor in terms of determining that outcome. So, as we're doing that, and, and the next question I wanna show you is that I wanna get back to that, what I've learned as a resident and uh we all have to relearn things. So, we came back to some of these genes and understanding these genes and the next thing we thought we would do is trying to understand real underlying mechanisms, specifically of the pulmonary outcomes, pulmonary hypertension, pulmonary insufficiency. So, we took one of uh what actually ended up being and still is the most common gene that we see, a gene called LAP1. I have to admit it's a complicated gene, uh, for lots of reasons I won't go into. Um, this is a plot of the gene with each person being a different patient and showing the different variations that we see. The reason that I say it's complicated, uh, for those of you who are really interested, um, this one single gene looks like it maps to three different Different diseases, one of which is CDH. Another one is a condition called CODIS syndrome, and another one is an isolated neurodevelopmental problem due to mitochondrial function. So, depending on where your mutation is, you actually have a different disease and we have to again parse those out as we understand things. We now start, can start parsing things out, um, due to AI, uh, not my own work in terms of doing this, but now being able to predict the three-dimensional structure of proteins. We can map out where many of these different missense variants are located and based on the location, the predicted location of these, we now start to see that they Clusters. So, for instance, in red are the conditions that are, or the variants that are associated with CDH. In blue are the genetic variants associated with this condition called C CODIS, particularly common in the Amish. So we can now see that in terms of just structure and function, uh, we now understand a little bit better why this happens. So, we started looking at, uh, these cases with the LA P1 variants and understanding them compared to the rest of the DREAMs cohort that was not associated with lawn P1. And for every variable you could look at, um, the prognosis was unfortunately worse. So, when we look overall at mortality, at pulmonary hypertension, and need for ECMO, uh, these kiddos were just doing worse in every single dimension, and the mortality especially high. And interestingly enough, um, and by high I mean 69%, uh, even for some very good CDH centers, uh, and this was simply at the time of discharge from the NICU. So, in looking at this, we now had the ability to uh use some of our furry friends uh to be able to help us understand whether or not this truly was, 1, a CDH gene and #2, why it was we had such bad outcomes. So, the first was to be able to knock this out in a mouse, uh, in the diaphragm specifically, and when you do that, you can see that you don't get a normal development of a diaphragm in a mouse. So we were convinced we were on the right track. This really was a CDH gene. But now you could do an interesting experiment which is to disarticulate the diaphragm defect from the pulmonary problem. And so now, what we've been able to do, and we're doing this now for a series of genes is take that gene out in the lung but keep the diaphragm intact and keep the rest of the body intact. So now you can isolate what are the lung problems from the rest of what's going on. So, those mice develop completely normally, no compression in the chest, no problems in terms of any of that mass effect. And you can see both in terms of grossly as well as in terms of the microscopy, how malformed those lungs are. And so there is a primary developmental disorder of the lung within this condition. And the reason why we have such bad outcomes is not at all because of the compressive effects. It has nothing to do with the LHR. It has nothing to do with any of that. It has to do, there's a primary problem with the lung in terms of this. We're now starting to get even more sophisticated and disarticulate within the lung, what things are due. Uh, for, for instance, the pulmonary vasculature and pulmonary hypertension. What are problems with alveologenesis and being able to understand the different mechanisms, the different levers we can pull in terms of treating that. Are we gonna treat pulmonary hypertension or are we gonna have other problems with aeration just because of what we've got with the alveoli and gas exchange? So in doing this, we now have both better prognosis. In some cases, it may not be good news, but at least it's realistic news. And in some cases, we hope we're gonna be able to get to, and I won't have time to talk about it today, ways of being even able to reverse that and I'll challenge us that the earlier we can reverse it, the earlier we can identify the issue, the earlier in development and even prenatally, we may have opportunities to be able to develop, to reverse that. So, just briefly to show, uh, and to show you how humble, uh, the genetics has made me, I'll just show you very briefly. Uh, the other gene I mentioned, which is we have more experience with but unfortunately is a bad prognosis is a gene called MYRF. Uh, we often see this associated not just with CDH, uh, but congenital heart disease, genitourinary malformations, sex reversal in some cases. And again, within the Dream series, uh, the majority of children died, uh, who had this. What humbles me is the following, I don't know if you can see them here, um, but a set of identical twins that we had. So this set of identical twins, uh, genetically we proved that they were in fact identical twins, but the thing that humbled me is that one of the twins, both of them, unfortunately died, um, and this was a de novo variant. Uh, one of the twins had CDH, a left CDH. The other twin had hypoplastic left heart syndrome. These are, again, twins that shared the same intrauterine environment, genetically, absolutely identically the same. Uh, had both associated features with the condition, but different manifestations. And so this also tells me there are stochastic processes in terms of cell migration that are overall sort of organized by genes like this, but there's still things that on an individual basis are literally random chance in terms of what happens and what doesn't happen. So the reason I say this is because you can imagine, I call that noise. That's just the noise in the background that we have to deal with and why the numbers become so important to be able to see that signal from the noise. Again, just to show you, lawn P1 is not an isolated case. We can do the same mouse experiments that we did in lawn P1 with MYRF. We see in a similar way that although this is called a myelin regulatory factor, in other words, a people thought had to do with brain development initially. Uh, it isn't just about the brain. Uh, we see similar problems in terms of lung development, both macroscopically and microscopically, and again, this is in part why we have some of the bad outcomes. Uh, this is a gene that has, um, a sort of major role in terms of conducting the orchestra of human development and why we see such major problems. Um, to switch gears for a second, CDH is not unique. Uh, so I'll just very quickly go over that this, in fact, we've seen now with every single congenital anomaly that we've looked at. Um, the next anomaly that we have the most experience with is congenital heart disease, uh, although I know this is not your major area of focus, just to say we've done this here, uh, uh, with a group called the Pediatric Cardiiaac Genomics Consortia. I'm, was, and actually still am the PI at Columbia, and then we have a group here, uh, with, uh, uh, in pediatric cardiology here and we've been doing this study for almost 15 years now. Um, as we've been doing the study for congenital heart disease, it looks very similar, um, and I won't go through all the details, but in a Similar way, we've been able to identify the genes associated with renal heart disease. Uh, it is also those genes also associated with worse outcome when we find one of those de novo variants. As we are now following those children over time, we're seeing associated other birth defects or congenital or, uh, rather neurodevelopmental problems and enrichment for those types of genes. In an interesting way, we're just starting to look now at greater sort of granularity of cardiac function. So, even after, um, We have great cardiac surgeons that fix the plumbing is the way I describe it. When we look at the long-term function, so the pump, the squeeze, long-term. We realized that those same genes are associated with systolic, with heart failure and problems with systolic function over time and interesting, also associated with arrhythmias. So, the primary substrate, even though you can fix the connections, fix the holes, be able to hook things up correctly, the underlying substrate of the cardiac myocytes is still on a cellular-based system damaged so that you don't get the same contractility and you don't get the same conduction going through the heart. So this has helped us now as we're getting into our adult congenital heart disease program in terms of being able to predict long-term who's going to have problems, being able to think about how to monitor individuals over time and as I said, uh, be able to think about um future roads in terms of who may need additional support, uh, even after fixing the anatomy. Just to uh let you know, it's not all about those de novo genetic events. Uh, we've identified even within certain communities. I was in New York, uh, for many years and as an example, GDF1 for congenital heart disease is a gene that we see with one specific mutation in the Ashkenazi Jewish population. So when I was in New York, I got to the point if I saw a newborn or a fetus with transposition of the great artery or double outlet, my right ventricle, and I took a family history and they were Jewish on both sides of the family, I could use my X-ray genetic vision and make the diagnosis with that one specific genetic variant almost instantaneously. Um, we've now gone on in work, uh, actually with, uh, Doria Shuram, uh, and Rabbi Eckstein and added this now to the Ashkenazi Jewish carrier screen. Um, so this is done within the Orthodox community as well as done for patients that are getting, uh, clinical testing. So that this is something now that for individuals who want to know about this can know about this even prior to conception or prenatally and be able to use this information as they're planning their families going forward. For congenital heart disease, a scorecard of how we're doing, uh, and we use this all the time when we think about prenatal testing. Uh, for complex congenital heart disease, in about 40% of cases, there's an underlying genetic identity that we can identify, uh, even prenatally now with prenatal exome genome sequencing. And even the cases that appear to be isolated, about 21%, uh, are due to underlying genetic factors. I say this because we recently, uh, completed a study prenatally where we interviewed the couples who had gone through or were offered, uh, prenatal genome sequencing, uh, when their fetus had an anomaly, and we interviewed specifically the parents that decided not to do the genetic testing. And interestingly enough, most of them said they didn't do the testing because their doctor said That the gene, genetic testing wouldn't yield a diagnosis for isolated anomalies, uh, specifically congenital heart disease. And so I think there's some education to be done in terms of understanding probability of a diagnosis and what the value of that diagnosis is. And I would argue for many of those couples, we learned that information was powerful to them, uh, for many of them, at least in terms of planning, planning their mode of delivery, planning where the, uh, surgeons that they would work with after birth and having a team in place. Um, the last example that I'll give in terms of structural anomalies is esophageal atresia, tracheoesophageal fistulas. Uh, I will say that we're farthest behind in terms of numbers for this, so we've got the most to learn is the exciting part. And I'll also say that this looks like it's the most genetic of all of the anomalies that we've looked at. Single genes that drive, I would argue, most of these cases. Um, in this particular case, we started out and we pushed ourselves by looking at less than 100 cases to see if we could understand what was going on. And even with less than 100 cases, we've now got pretty good intuition about which cases and which genes are involved and you may not recognize this, but the SOx genes over here on the right are something that are coming up again. So this group of transcription factors, again, very important. Interestingly enough, there are all sorts of connections which I'll get to in a second. This gene, ITS1, which we saw first, uh, with esophageal atresias, we now know, believe it or not, is related to Parkinson's disease risk. Uh, and so there are genes that are doing all sorts of things over the life course if you look closely and if you look long enough. So, as we're doing this and as we started looking at these, we challenged ourselves, uh, within the small number of cases to see if we could sniff out the genes that were involved. And as we did this now, Turned to frogs, uh, because it was a higher throughput system for us. So with Aaron Zorn at Cincinnati Children's, uh, have a PO1 grant that looks at this, and we picked the genes that we thought were the best genes, and we then started knocking them out in the frog to be able to see if we could recapitulate the phenotype. And shown here on the right is what this looks like. This is the normal separation of the esophagus up here and the trachea down here. And you can see that in each one of these cases that we Yes. Uh, in fact, we were right. We see abnormal separation of the esophagus and the trachea, and as we were doing this, uh, our batting average wasn't bad. 13 of the 18, we immediately saw were correct, even though it was based on just one single case that we saw, being able to recognize that. And now that we're starting to go back in those 5 that we left on the table, as we look more carefully, we realized many of those genes also actually were involved just with more subtle defects. So, as a scorecard for putting this together, and, and I'll just briefly say that um I still remember the day I had a phone call. I have a lot of phone calls, but I had a conference call at 8 o'clock for congenital heart disease, I had a conference call at 120 for autism, and I just remember saying to myself, this is deja vu all over again. We have had the same exact conversation for congenital heart disease as we had for autism. There's gotta be a connection here. And that was over a decade ago, but in fact, that's been true, uh, that many of these genes, in fact, I could be on many different Calls from many different consortia, but they're the same genes that tend to be involved. Uh, interestingly enough, it's now sniffing out which genes tend to be associated with which conditions, and we can now make computationally with more data, with artificial intelligence and machine learning, make predictions even for genes that we've not seen before based on where they're expressed, what cell type, where, when they are over the life course, what we will see in terms of the future even without having data behind it, but making predictions about this. We can now see that there's certain things that stay in their range, so certain genes that tend to stay, uh, and not affect the brain in particular in terms of ultimate neurodevelopmental outcome. That's incredibly important to parents to know is this something essentially that you all can fix and then their child will be fine after that, their head will be fine, or is this something that is going to be associated with long-term neurocognitive issues and potentially significant, uh, disabilities and, and limitations on independence. And so we've been able to map these, as I said, based on the data. The numbers that we need are big. And so just to show you what we have been able to do and, and I do think big in terms of some of these things, uh, I put together an autism cohort called SPARC, Simon's Foundation Powering Autism for Knowledge. Uh, the same thing but now thinking about autism and the heterogeneity behind autism, now a very common condition, 2% of the population. This is very much male enriched 4 to 1, males to females, but knowing again that prognosis or things associated with it over the life course in particular might be guided by the genetics. I won't go a lot into detail, but I will highlight this ITS1 gene again. So, that same gene that we saw first with esophageal atresias, we now saw One of the most common genes associated with autism, in particular, in our inherited cases. And so we now know and have been following, and that's how we made the connection with Parkinson's disease. Now looking at those families and looking at the older generation, seeing other neurological conditions associated with that and realizing, as I said, there's more complexity. In doing this, we wanted to be able to build communities and empower patients and their providers to use this information. So with this in the autism and neurodevelopmental space, we've now created a group around the world called Simon Searchlight. This is now based on gene, so individuals get a genetic diagnosis either through research studies or through their clinical testing, and we do this around the world internationally. Within this, we have 200 chapters under Simon Searchlight, so each one of these is a different rare genetic condition. Many of these, we now have upwards of 500 individuals per condition and we now have meetings. I've got one of these on Saturday, for instance, online, and we do and we're bringing many of these to Boston. to right here, uh, we've had them right here in this room. Meetings where we bring the families together, we learn from them, we intensively put them under a microscope through all sorts of batteries of tests to understand what they're like over the life course, and importantly, to learn from each other and to develop things that they can take home to their providers wherever they are. As we're doing this, we are trying to understand this as early as possible so that there's an opportunity for intervention, we get in as early as possible as we can do this. And with folks like Mike Tarkowski and Ron Wapner, we are exploring in terms of prenatal sequencing. So we have a consortia called Prenatal Seek. Uh, we've done this with amniocytes or, uh, chorionic bill sampling, and we're pushing ourselves now to be able to do it from a maternal sample as early as 1011 weeks of the pregnancy where we can read out the fetal DNA sequencing from that maternal sample and be able to very, very early in the pregnancy make these diagnoses even before we may see something on ultrasound. As we're doing this, uh, speed matters. Uh, so we're doing this even in our NICU here, uh, where if we need to, we can get this genetic testing, these results within a week. Uh, we're pushing ourselves in the same way on the prenatal side to be able to get the information, uh, in a time to be able to make decisions about management, uh, as families are trying to cope with a lot of anxiety and uncertainty. Um, and as we're doing this, just to give you an example, and this was from one of our cases at Columbia, as we're doing this prenatal testing, we do identify some cases that can be associated, in this case, FOXP1, uh, with very, very severe pulmonary problems, pulmonary hypertension, pulmonary dysplasia, and, uh, can, it's, it's tough, but we have tried to even think about, uh, newborn lung transplant. Uh, it's not a very large, uh, group of donors, uh, unfortunately, to do this, so we're very oftentimes Not successful in terms of identifying those individuals, but by knowing this very early on, we know what our options are and we know what we need to do. And so, uh, informing transplant decisions ends up being, uh, was very important for us, whether it was lung, uh, liver, kidney, but trying to understand, uh, who would be the best candidate and making sure, even in cases, for instance, of kidney transplant, making sure a family member might not be at risk for the same condition, uh, and therefore using the wrong donor as we do it. Um, the most recent thing, and I'll end on this, is I've been worried about equity. Um, so in thinking about who comes to see me and importantly, who doesn't see me for whatever reason, uh, trying to understand how we don't leave anyone behind in this process. And as I had been thinking about this for a long time, I realized that from a genetic point of view, the one time over the life course in the healthcare system where everyone passes through the same point is in a process of newborn screening. So I actually started here in in New York in terms of the entire process a little over 50 years ago. We take a single, a few drops of blood, uh, from newborns from a heel stick, put them on filter cards, Guthrie cards, and have whole public health processes for newborn screening. And so it came to me that that was the way to be able to get access for everyone. If that was just a system of care, people didn't have to think about it, it just happened. That would be the point at which we could intervene and be able to get everyone the same and equitable access. So, I started out with newborn screening pilot studies when I was in New York in 2016. We did one for spinal muscular atrophy. Uh, I timed it and Very intentionally timed at exactly the time we had ASOs that were just starting in clinical trials, knowing that we need to identify babies pre-symptomatically for that degenerative condition so that we could get them access to treatment in time to do that, and, and that was actually right. Uh, and the newborn screening actually powered those clinical trials that, uh, allowed those. ASOs in part to demonstrate efficacy and get FDA approval. That FDA approval, we very quickly turned around as the evidence review for the recommended universal screening panel. And so, in record time, that was added nationally to newborn screening, and now I'm proud to say it's been implemented nationwide. We now have 3 FDA approved medications for SMA every baby who's screened, and we within 3 years, a condition that used to be the most common genetic cause of death for children less than 2 years of age to now a completely different, uh, condition where with Basil Darius, uh, we follow these children and they are strong and they walk and they live their lives and I'm not saying they're absolutely perfect 50 years out cause we don't know that, uh, but it's a very, very different, uh, situation. So we've uh taken that same platform and now, uh, we do a study called Guardian Genomic Uniform Screening Against Rare Diseases in all newborns. Uh, I started this in New York City, although I'm glad to say the Department of Health in Massachusetts is going to allow us to move forward in the Commonwealth, so coming to, uh, babies near you, hopefully soon. Uh, we take that same newborn's green blood spot and do whole genome sequencing, so generate data on their entire genome. We don't read out all the information. We limit ourselves in terms of looking at the information for actionable conditions that we can do something about. Right now, that includes about 450 conditions. We read out that information. We're successful almost 100% of the time in generating the data and returning it to families within 3 weeks. And I won't go into a lot of detail except we do focus on conditions for young children. So we don't talk about risk of Alzheimer's disease or Parkinson's disease or things long term. We focus on the newborn period with things that we're quite certain about as we do that. And as we've done that, we, uh, find that most parents want to do this within New York City, uh, in terms of that cohort in incredibly diverse community, uh, which is actually majority minority, about 31% white, but many, many other, uh, parts of the world represented and about 75% of parents electing to do this consented research study. As we're doing this, it ends up that about 3% of the babies are positive for something and that we're almost always right in terms of that something. We go from screen to diagnosis to onboarding to care and walk families through that entire process. Uh, and we have learned along the way, uh, because we do have an incredibly diverse group of individuals that sometimes we did interpret the information wrong, but we have learned from that and as we've done that, improved in terms of our false positives over time. We're also thinking, and I know that, uh, people may have, uh, things they have to say about EPIC, uh, but I will say EPIC is going to help us at least in the rare disease community. And so we've been working with EPIC, uh, in trying to understand this superhighway of information and how we can make all of us smarter faster and not have to remember all of the millions of things. that go with rare conditions. Rather have people like me as the brains behind you, so that as you have a patient with a diagnosis, surface the information just in time when you need to know that information without being distracted by where do I go for it, what's the most up to-date information? Can I trust that information, be able to deliver and serve that up to you as well as to our families. We've worked on that and I've worked with the ICD uh committee to be able to now get a series of rare disease genetic diagnoses, so at least we can code these correctly. As we code these, we can use that as well as natural language processing. I'll tell you coming to an epic instance near you, uh, i.e. here, we're gonna be having governance in terms of being able to make decisions about where that information will live, consistency, so everyone knows where to give it. Uh, we won't have liability issues that you missed it because we'll surface it and make sure that you can't miss it. Um, we are publicly and freely, uh, publishing the clinical care guidelines that go with this, so it's not behind a paywall. Everyone can access this and be able to put that, like I said, in bite-size bits that'll be in Epic so that as you're seeing this, this information can be fed out in an age-appropriate way for whoever the provider is. Again, equity in terms of giving the information to more people at just the right time and minimizing the burden that goes with that. As patients and their families have access with MyChart to be able to see this, I fully believe that they can be part and should be part in terms of, uh, this journey. And so surfacing that information to them in multiple languages, and information and language that they can use this, is also part of the mission that we have. So, I hope I've, uh, uh, taken you through a bit of a whirlwind tour, uh, but trying to think of how we can use genetics and genomics as we're going forward in this. Um, I'll say that as we do this, and I'll, I'll say this again for anyone who wasn't on at the beginning, it takes massive numbers of people to do this, both patients and participants and also smart people like you all in terms of being able to be partners going forward. Not, no one of us have the same information or perspective and so we're better together as we go forward. Um, I just wanna personally thank all of the many people who've been part of the teams over the years, part of my team here. This is, for instance, uh, the Dreams team in terms of going forward. You can see just the huge number of people it takes to accomplish these things, uh, the oesophageal atresia team, and, uh, within, uh, many of these, uh, just want to acknowledge our funders and like I said, the patients and families. So, I'll stop here and take questions. Wow, Um, So I'm pretty sure that we could invite Doctor Chung to give 100 rounds in a row. And she would have a different talk every single time, because I've had the great fortune of being exposed to her and her work in many different fora. Um, I guess I'll admit publicly. So I was on the search committee that recommended the choice of Doctor Chung to the administration. To take what I believe is the most complex, important job in this institution. And even having a deep dive into our background. I am still stunned at how much I didn't appreciate um because there's um There, you know, so many of us dive deep into something, right? We, and we like to spend our careers focused on something. And she just gave in just this one talk, examples how she's working at the level down here, right? It, at, at the actual gene level, and they're working at the collaboration level in diseases that we thought were surgical, right? That we spend our lives treating and trying to improve the care of. She only barely scratched the surface of things that she's done at the policy level. Um, at the collab international collaboration level. Um, I, I think I'll get this right. Doctor Chung was the principal plaintiff in a Supreme Court case that prevented the ability to patent a human gene, which could prevent research and therapy. I have that right, um. On and on and on. Um, and then she took a job here at a pretty complex place with an enormous bandwidth responsibilities. Um, just learning the acronyms, uh, uh, in this place takes most people 6 months. Um, and nobody told her before she got here that somebody put a banana peel on the floor before she stepped, um, on the ground in Boston, and she knows what I'm talking about. Um, and somehow, um, I, I, I like we think we work hard. Um, I just don't know. I just don't think you sleep, um, and I can test that phone calls because she and I have calls all the time. It's, how about Sunday at 1:0 and she's like, you know, in an airport on a layover like between, and it just gets all these things done. Um, so, um, we are thrilled that you're here. We're thrilled to have, um, this kind of science and this kind of population impact on the things, uh, the children that we try to help, uh, we believe we do help, um, and, uh, and we're, we're, um. Uh, I don't know how you keep your practice going, your lab going with all the things that where I spend my time with you, um, but it's stunning. I, I would open up to, to questions for her. Thank you so much for that, like, incredibly inspiring and amazing talk. Um, the question I have for you is, you know, now you have, you, you have the, you've worked very hard to prove that when you have an idea, it's generally gonna be a high yield result on the other side. When you were starting out, how did you convince funders, collaborators, etc. to take on these projects for rare diseases? Especially in scenarios where you need to accumulate a big N and the payout, the payoff, the result is not gonna show up, maybe for many years, um, what arguments do you use to make those projects seem appealing to those people? Yeah, um, it's a great question. Uh, number one is young people have a lot of advantages, so use it to your advantage, um. Number one, you do have early investigator status at NIH. Uh, you've got, um, uh, just alacrity with and facility with ideas that older people don't. So one of the things I often say is that the younger generation should be skating to where things are going. Don't do just the things that other people have done. And so I use that to my advantage, the fact that I was young and that I had some ideas and, and over the time, and I've only known this afterwards, there were some people that were champions for me at study section who, who sort of knew me as a person and championed me and were willing to be able to give me some rope to go off on. And I, like I said, I didn't know it at the time. I've since known that as people have whispered to me and turn and, and sort of been able to put the pieces together. One of the things that I've been doing over my life course, but here at Children's in particular. So number one, all of the data that I talked about today, which includes on autism, we've now got genomic information on over 100,000 individuals as an example. I've, I've made all of this freely available and accessible so that the next generation doesn't have to go through what I went through in terms of building those cohorts and learning together. Specifically, since I've been here at Children's, number one, there's already been this ecosystem with that idea, but I don't even know that we've used it to its full potential. One of the things I've realized as being a newbie coming in is that there are a lot of things, but there's not a good way for me, at least to navigate and find out what's where and make those connections. So I do think we're gonna, um, at least in the Department of Pediatrics and, uh, Steve knows I've Extended that to everyone in surgery. I'll just say that now in case you didn't know it. We are happy to help in terms of navigation, uh, specifically for people that may need help in terms of knowing here's the, who's here at Longwood, uh, at Children's, at Longwood, at the Broad, um, and not only for this, but for, you know, other both methods and clinical trials and other things, uh, we know that it's hard to be able to navigate that as a junior person, especially. Um, so those are the major things I would say, and then, uh, the other portion of this is that, um, I had to have surgeons working with me. I mean, I'm not a surgeon, right? And so in terms of doing this, these are really your conditions and so, Uh, I've always worked with surgeons who have really been the champions and my co-pilots. And so we've always had, as we did these, what we call MPI or multiple principal investigators, and that's always the way I've done things. I have a subject matter expert. I'm the genetic expert, and we put this together. And so as you were hearing at the introduction, there's no way I could be an expert in everything from breast cancer to CDH, right? I, I have to have clinical partners in this, and, and that formula has worked very well. So find friends, I guess is the other portion. Thank you very much. Uh, I say, uh, I was initially scared because of PTSD from genetics, uh, in, in med school, but you took us along for a ride and I was there with you the whole time. So thank you. Um, and it just got me thinking a lot about what a lot of us do, which is prenatal counseling, um, and, you know, that's always been a privilege, but also one of the big, you know, challenges with it is, you know, we're asked, as all of us do, to counsel from a place of uncertainty. Um, and sure, we have the anatomic numbers. OK, it looks like there's an esophageal tresia, maybe long gap, or is your LHR, and these are the general, you know, anatomic variables we can use to help prognosticate. When it comes to the other things, you know, when it comes to, OK, is it, is it complex, in some ways, you know, we say, well, this, these are things we can't predict, so we're going to leave that up to the, who knows, but we can tell you the anatomic factors. It's a little intimidating now as we see, OK, now we're going to see that place of uncertainty is going to become more certain. Any advice for how we use this genetic information and, you know, prenatally, like I don't, I don't know when I'm getting a prenatal consult. Am I going to start seeing more of these genes and, and how confidently we're going to be able to use that to now help a family might be making decisions about whether to keep the pregnancy or not? Yeah, so I'll throw out a provocative idea, knowing that I don't make these decisions, but I would say on the prenatal side, when you see a structural anomaly, and we'll have a paper hopefully coming out soon that you'll see the data behind this from prenatal seek. Um, I would argue that everyone should be offered the opportunity in terms of genomic prenatal diagnosis. Um, we've done, we've literally studied this across 4 sites in prenatal seeks, so both I can, you know, tell you in terms of the results, but we've also gotten in the head. We've done these semi-structured interviews and surveys with the families to know as they go through the journey, what they're thinking and how they're reacting to that and Uh, what I'll say is that different people want different information and they get overwhelmed. Some people are information seekers, some people are overwhelmed and they're at the max of what they can handle at the time. And so, I don't force anything down anyone's throat as they're doing this. Um, but one of the things we've realized is that, number one, we're not as smart as we think we are sometimes, and so, one of the things, if we do do prenatal sequencing as an institution, I hope we will also think about If we don't find a diagnosis, also looking postnatally at that same information again when we've got the baby in our hands and have a better sense of who that baby is and then also be able to look at that again and do that rapidly enough so that in the NICU we can use that information for management. Um, I do think one of the things we've talked to families about is a lot of women are afraid of the needle, so they're afraid of the amniocentesis and they're concerned about the fragility of the pregnancy. As we get to non-invasive methods of doing that from a blood sample, my prediction is that if cost is not a barrier, the vast majority of information of of people are gonna want the information. Uh, I think we can also think about What we give and what we don't give. Again, I don't know that fatally we need to be predicting Alzheimer's disease or breast cancer diagnosis or things like that. I think if we constrict and focus on the things that families are worried about, about that sort of early life course, um, I do think most of them want it if they can afford it. I think that's the next question is from an equity point of view, how do we make sure that everyone has access? So, on the policy side, working on things for that as well. Well, I think it's clear that we could go on with, I have another 10 questions myself that just from, from today's talk, um, we're obviously out of time. Uh. Well, we would like to invite you for the next 9, lectures. So that we, we continue to learn and and and um. Uh, and I, I just wanna say on behalf of the entire community here, thank you. I thank you for um. Taking a chance on us. I know coming to Boston Children's is a big deal and everybody wants to do it. And you have described, I heard you described to some of my colleagues here that it was even more amazing than you expected, having admired it from afar. Um, but, um, um, People in your department, some people in this room whom you've already met with, um, and certainly people in, in the boardroom and amongst the chiefs know that you are the agent of change, uh, for the good of this institution. So, um, thank you for everything you're doing. Well, it's my privilege and I'm, uh, very proud to be amongst the number of uh esteemed faculty here. So thank you for having me.