Dr. Stephen Kingsmore – A Vision for Transforming Medicine with Rapid Genome Sequencing (Part 1)

Dr. Stephen Kingsmore – A Vision for Transforming Medicine with Rapid Genome Sequencing (Part 1)

Dr. Stephen Kingsmore, MD, DSc, is the President and CEO of the Rady Pediatric Genomics and Systems Medicine Institute. Dr. Kingsmore spoke at the inaugural Precision Medicine Leaders Summit last summer. We spoke with him about his program for rapidly sequencing the genomes of sick children and what he’s looking for in the field of next generation sequencing and genomics.

AllSeq: Please tell us a little bit about what prompted your move to the Rady Pediatric Genomics and Systems Medicine Institute.

Dr. Kingsmore: Obviously multi-factorial, but from a career perspective this is very logical. So you know I’m fifty-five, going on fifty-six and for most of my career I have been pushing the idea that biotechnology is going to revolutionize medicine. Over the last decade or so that has focused on genome sequence information. Two jobs ago, I was at a research institute and we were one of the first to start to sequence human genomes. And at some point during that process, we realized, “wow there’s information in here that’s going to transform medicine.” And so then very purposefully I moved to a children’s hospital where they had lots of kids with genetic diseases and we could start to get experience. Over the five years or so that I was at the Children’s Mercy in Kansas City, the results completely stunned me. We published a lot of papers but two that really stand out. One was thirty-five babies in a neonatal intensive care unit. We decoded their genomes and fifty-seven percent of those babies got a diagnosis and two-thirds of those diagnoses changed how the baby was managed. So that made me realize that this is ready for primetime. You know this is stunningly successful. And so at that point you know it was sort of “this has to go big.” You know doing thirty-five babies is nice. It’s pioneering, but we know I need to think about ways to assess how generalizable is that. And then if it is generalizable, how do we implement that for all babies. I couldn’t do that at Children’s Mercy Hospital, that wasn’t my job description. And so I started the search to figure out where might that be possible and I believe that where we are at is the right place. San Diego is a big urban center. It’s west coast so people are fairly progressive in their thinking, and San Diego itself has eight hundred biotech companies and genomes are a big industry here. We’ve got companies like Illumina being major, major employers. And so again I felt this is going to be a community where the idea of testing genomes is going to be a mainstream, kind of normal bar conversation, you know, not some kind of exotic weirdness. The fact that we had all of those biotechnology companies, and in particular Illumina, meant that we could really forge strong commercial hospital partnerships which were going to be pivotal to being successful at scale. And then last of all there was the actual hospital itself. The physicians really wanted to get into genomic medicine and the executive team did and the board of directors did and then a guy called Ernest Rady did who had given a $120 million to get the thing kick-started. So when you add all that together, it was clear that it might be possible to do something here in San Diego at Rady Children’s that you just couldn’t do anywhere else in the world.

AllSeq: You’ve been using NGS in the clinic in a very specific way – applying rapid whole genome sequencing to the diagnosis of rare childhood diseases. Why is speed so critical in these cases and what are you doing differently from other hospitals that are sequencing patients?

Dr. Kingsmore: You know I think you’re being kind. I think that we’ve been a little bit lucky that we’ve sequenced only a relatively small set of children. And yet we did pick winning application areas – one in developmental disability kids and the other in the NICU kids. We stumbled across the idea of speed. Illumina approached us in 2011 with a new sequencer that was very fast with kind of the idea, “might anybody want that speed?” And we did think “well, there are high acuity clinical situations where you need fast answers and if it’s going to be relevant, it had to be relevant there.” And it turned out that there was some merit to that idea. I think what we are now trying to do is something a little bit different. Which is to understand, instead of using a hammer for everything, to really start to assemble a suite of tools that will fit all clinical situations. So some kids need a rapid genome but probably not very many – who are those babies? And how fast do they need their genome? You know we don’t know the answer to that. How do we pick those babies? And then for kids who are more stable, what’s a good turn around time for them? And again which kids should be highest priority for this, and then how generalizable does it go? These are the questions that we would like to answer over the next year or two years so that we can start to map out “OK it works here, will it work there, and how well does it work, and how do we make it most cost effective? Those are the sorts of things that you know we’re now pushing toward and I don’t think we’re any different than any other children’s hospital in the country, of our size. I just think that we have a unique opportunity given the fact that a new research institute is being created to actually have budgets to be able to push these questions into really powered studies. Our experience across the entire United States over the last five years has been in pitifully few patients and pretty much nobody has been able to afford to do randomized controlled studies. That has to change. We need evidence, we need to know when it works, where it works and how it works.

AllSeq: Most of the clinicians we talk to wrestle with choosing between targeted panels and exome sequencing. For them, whole genome sequencing is simply out of the question, primarily due to lack of reimbursement and difficulty in analyzing the results. What makes your program different?

Dr. Kingsmore: I guess partly we’re just a little stubborn. And you know to be honest there is not much computational difference between a genome and an exome. One good exome will give you a quarter of a million variants and a good genome will give you five million variants. That’s not that different honestly. A quarter of a million variants is a big computational problem. Five million is only twenty times more. So you know I don’t think it holds water to say the reason not to do genomes is because it’s too complex. It’s not. In fact there’s a lot of argument to be made for “well, if you’ve got the entire genome, you can start to rule things out.” Whereas, in an exome you never really know if you’re finished because you’re only looking at two percent of the genome. So maybe genomes are a lot easier to analyze because you’ve got complete information. The main reason I believe why people do exomes is because they are a lot less expensive. It’s a budget thing and we’re going to hit that really soon as well because our goal here, our vision, is to be able to provide genomic sequencing for every child who might benefit and we certainly can’t afford to do a genome on everybody. So I think going forward we just want to start to do some randomized studies to say “OK, can we measure how much better a genome is than an exome?” Unfortunately nobody’s really done that study and it probably needs to be done three or four times in different disease areas so that we actually know the net difference. I would love to do that study next year and then we’ll know. I imagine that probably genomes today are not a lot better than exomes. Exomes are really excellent. The genome is nice if you can afford it because you can then store all of that other information knowing that at some point in the future we’re going to know an awful lot about variation that we can’t understand today that’s not in the exomes – regulatory variants and structural variants that you can’t get from an exome. So in the long run you’re creating a valuable database as our knowledge improves. But from a practical standpoint of “I’ve got a sick kid and I need a test”, an exome is a great solution. So we’ll be bringing up exomes (over time). The point is a lot more moot than it was several years ago when exomes were really quite awful. They just weren’t the whole exome – they were a partial exome. But now exomes have gotten really good. And there isn’t, I don’t think, a huge amount of difference today between the performance of the two, but we need somebody to formally study that.

Published with the permission of AllSeq Inc.