Name:
Decoding Genomic Diversity with deCODE Genetics CEO Kari Stefansson on “Close to the Edge”
Description:
Decoding Genomic Diversity with deCODE Genetics CEO Kari Stefansson on “Close to the Edge”
Thumbnail URL:
https://cadmoremediastorage.blob.core.windows.net/154a03ac-dd9e-4fcf-9324-50a4e7da83ac/videoscrubberimages/Scrubber_10.jpg?sv=2019-02-02&sr=c&sig=95vVZBTtx8QN8CGTvFk%2F8LLPT5yDTc31t%2FTKygrexyk%3D&st=2024-12-08T21%3A39%3A57Z&se=2024-12-09T01%3A44%3A57Z&sp=r
Duration:
T00H37M31S
Embed URL:
https://stream.cadmore.media/player/154a03ac-dd9e-4fcf-9324-50a4e7da83ac
Content URL:
https://cadmoreoriginalmedia.blob.core.windows.net/154a03ac-dd9e-4fcf-9324-50a4e7da83ac/CTTE - Ep 15.mp4?sv=2019-02-02&sr=c&sig=5kMHSvQz52sMkE5q2dxGsxWqgm9mZy%2FhxPmSKz0INX4%3D&st=2024-12-08T21%3A39%3A59Z&se=2024-12-08T23%3A44%3A59Z&sp=r
Upload Date:
2023-10-19T00:00:00.0000000
Transcript:
Language: EN.
Segment:0 .
[MUSIC PLAYING]
MALORYE BRANCA: Welcome, GEN Edge subscribers. This is Close to the Edge. And today we have the pleasure of speaking to Dr. Kari Stefansson, who is the founder and CEO of deCODE. Now it's one of the most fascinating companies in this whole field. And we have a lot of questions for Dr. Stefansson, including what went into their latest paper which has some of the biggest data analysis of whole genome sequencing and some very interesting findings.
MALORYE BRANCA: So let's get started. So what are the major findings from the recent Nature paper, and how did you achieve them?
KARI STEFANSSON: You see, this paper that we published was based on 150,000 genomes from the UK Biobank. It's just one of these many steps we have taken in our study of human diversity. And human diversity, like all of the diversity in the biosphere, begins with diversity in the sequences of A, G, C, and Ts. And it isn't all that many years ago that the first human genome-- where the sequencing of the first human genome was completed.
KARI STEFANSSON: And now, all of a sudden, we at deCODE, we are a part of a group that is sequencing the whole genomes, so 500,000 people.
MALORYE BRANCA: Right. And then the name--
KARI STEFANSSON: And we have--
MALORYE BRANCA: --of that group again.
KARI STEFANSSON: That group consists of deCODE genetics and the Sanger center. We took on the task of sequencing all of the genomes in the UK Biobank. And basically, this paper that we published was based on the first 150,000 genomes. And there are-- the thing that stands out in my mind are, for example, the fact that most of the regions in the genome that are fairly-- that are intolerant of sequence diversity are outside of the exomes.
KARI STEFANSSON: So basically, if you look at sequence conservation as a sign of functional importance, most of the functionally important regions are outside of the exome-- the coding exomes. And another thing that makes it important is that when you contrast the whole exome sequence with the whole genome sequence that it is not just that by sequencing the whole genome we get the regions in between the genes, we also get-- about 10% of their coding exomes are not captured by the whole exome sequencing.
KARI STEFANSSON: And so there are a lot of things that we get by doing the whole genome sequence that is not obtained with the whole exome sequencing. The whole exome sequencing of the UK Biobank provided extraordinarily important opportunity to mine the coding sequence of the genome. But it is-- the whole exome sequencing is the poor man's way of capturing the diversity in the sequence.
KARI STEFANSSON: In the paper, we provide several examples of discoveries that we could make of correlation between variance in the sequence and variance in the phenotype that could not be captured with whole exome sequencing. But this is always a balance of cost, on one hand, and the completeness of the data that you get. And up until now, the whole exome sequencing has been much cheaper than the whole genome sequencing.
KARI STEFANSSON: But that is probably changing. We are told by several companies that sell sequencing technology that, come next year, the cost of sequencing a whole genome will be down to $100 a genome.
MALORYE BRANCA: That's very, very important. Besides having the actual samples and the technology, was there anything, for example, on the analytics side that is a deCODE-unique feature?
KARI STEFANSSON: We use our own analytical platform that we have developed over the past 25 years at deCODE. And it comes very easy to us to handle large amounts of data. It comes easy to us to do variant calling in a very large data set-- for example, do joint variant calling of all of the samples in the data set. So we could bring that to this task, our experience in working with large amounts of data and all of that.
KARI STEFANSSON: And we have now, I think, at deCODE sequenced about 350,000 whole genomes, which is a very large amount of-- or generates a very large amount of data. And we believe that we are in a very good position to manage data of that sort.
MALORYE BRANCA: Excellent. What is the next step for this particular project?
KARI STEFANSSON: The next step for the UK Biobank-- and keep in mind that the UK Biobank is a very, very unusual enterprise. It is the biggest gift ever to biological sciences. It is developed by a partnership between-- let's put it this way, the whole thing basically comes out of a partnership between the public and the private in Great Britain. And they have made the data available to the entire world to work with, which is beautiful, which is one of the things that has turned out to be a little bit more difficult for you Americans to do with data coming out of things like this.
KARI STEFANSSON: In spite of all of the promises to do so, it has proven to be somewhat difficult. So the Brits are leading the way when it comes to that. But the next step for all projects of this sort-- and keep in mind that what we are doing now is that we are studying human diversity. We are trying to figure out what are the biochemical foundations of the differences between people, and then we are putting diseases in the context of what we learn about the diversity.
KARI STEFANSSON: And the diversity begins with diversity in the sequences of A, C, G, and Ts, but there is more to it than that. A very large part of our diversity is rooted in the interaction between the genome and the environment. The diversity in the sequence of the genome generates phenotypes that are under to the influence of the environment.
MALORYE BRANCA: You're talking about epigenetics?
KARI STEFANSSON: And so-- no. So how are you going to capture the environmental influences? The genetic component is easy to study. Or easy-- it is becoming easier today. We can now sequence the whole genomes. But the study of the environmental influences is more complicated. We have no systematic way of scanning all of the environment. But one of the things we know is that nothing really happens in a body without involving the proteins.
KARI STEFANSSON: The proteins are the business molecules in our body. So when we have the sequence of the genome, and we have measured, let's say, level of proteins in blood, we can determine what part of the diversity in the level of proteins is rooted in the genetic control of the expression. And what is left over, it has to be because of environmental influences. So by using proteins in blood, using sequence of genomes, we can infer, to a certain extent, what is due to environmental influences.
KARI STEFANSSON: And it's also-- one of the interesting things is we are stuck with the same sequence in our germline genome from conception to death. But the level of proteins, they rise and fall as a function of time to and from event, all right? Event that can be an onset of disease, or can be the end, or the pathogenesis, the atherosclerosis that lead to a heart attack, et cetera. So being able to put the genetics, the sequences of the genomes, in the context of proteomics gives you substantial power to capture not just the pure genetic influences but also environmental influences.
KARI STEFANSSON: So one of the things that has started in the UK Biobank is an enterprise or a project of measuring level of proteins in blood. And the first results of that are now coming out. There is a paper coming out on the measurement of the 1,500 proteins in the blood of 50,000 of the participants in the UK Biobank. And soon thereafter, there will be a report of measurement of 3,000 proteins in the same 50,000 individuals.
KARI STEFANSSON: At deCODE genetics, we have measured about 5,000 proteins in the blood now of, all together, 50,000 to 60,000 individuals, some of them from Iceland, some of them from clinical trials elsewhere. So I think that the next avalanche of discoveries when it comes to the biochemistry of human disease will come from the combination of genetics versus proteomics.
MALORYE BRANCA: I see. Interesting. You made a comment that this was happening more slowly in the US. And do you think that that's simply because of privacy or issues that have gotten in the way?
KARI STEFANSSON: No. No, I think that it has been a little bit of clumsiness when it comes to making data from public resources available to people all over the place. For example, you have the All of You process that is-- All of Us process that is going on now. The data coming out of that has not been made as generally available as the data from the UK Biobank.
KARI STEFANSSON: And I think it has a little bit to do with just cultural differences. But for me, it's somewhat interesting to see that, over the past decade or so, the Brits have taken the lead when it comes to big genetic projects like this one. And they-- but I'm convinced that Americans will catch up with them one day.
MALORYE BRANCA: Yeah. Yeah, they also took the lead in the coronavirus sequencing.
KARI STEFANSSON: Yes, they did that. I actually-- I insist that we in Iceland were the ones who took the lead because we sequenced the genome of everyone-- the viral genome from everyone who was infected in Iceland.
MALORYE BRANCA: Interesting.
KARI STEFANSSON: And we did that at deCODE genetics. We started to sequence everyone who was diagnosed from the middle of March 2020 and basically until February of '22.
MALORYE BRANCA: That's quite an achievement. How has your relationship with Amgen evolved?
KARI STEFANSSON: They have been-- they could not have been better. We continue to function like we've always done, which is basically focused on studying human diversity using genetics, and now of late, proteomics and even some metabolomics. And we have been-- what has been coming out of our work has been applied, number one, in the attempt to figure out what is a good target to work with, and secondly, to use it also both in the design of clinical trials and the monitoring of the clinical trials.
KARI STEFANSSON: So I think that has gone extraordinarily well.
MALORYE BRANCA: How has it changed, though?
KARI STEFANSSON: It has changed in the sense that the field has evolved. The data available have become greater. So what we have of instruments to help with drug discovery and development has increased. For example, the probability that we can help with, for example, the stratification into clinical trials has increased dramatically because even though polygenic risk scores can be useful, as I told you before, the variance in the sequence that are used to build up the polygenic risk scores are variants that we're stuck with from conception to death.
KARI STEFANSSON: So they are not particularly good to use to predict when something is going to happen. But when you have proteomic biomarkers, they basically have a level that rises and falls as a function of time to and from events. So with the advent of the technology to measure very large number of proteins in blood, we had all of a sudden in a position to help much more when it comes to stratification into clinical trials, when it comes to interpretation of what comes out to clinical trials.
KARI STEFANSSON: So that is basically probably the biggest change is that our usefulness when it comes to the developmental side of the R&D has become substantially greater.
MALORYE BRANCA: That's always been a goal of using genomics in guiding clinical trials and finding subgroups. Is it actually beginning to be a reality?
KARI STEFANSSON: Yes, it is a reality. It definitely is. And the goal is not necessarily to use human genetics, but the goal is to use understanding of human diversity-- the human diversity when it applies to risk of disease, response to treatment, et cetera. And it is becoming reality. There is-- if you think about human disease, and you leave-- let's leave cancer outside of it for a minute.
KARI STEFANSSON: We call it a human disease when the physiological function of an organ is not functioning as it should. It is either functioning too much or too little. And what we are trying to do and what the field is trying to do is to figure out what the physiological function is that isn't working in these various diseases and what is the biochemical pathway that impacts this particular physiological function.
KARI STEFANSSON: And then the goal is to develop a drug that either downregulates or upregulates the biochemical pathway. And the component of the biochemical pathway then become biomarkers that you can use. And this is becoming reality. We have several examples where we have had major impact on approach to clinical trials, et cetera.
MALORYE BRANCA: What are some of the examples?
KARI STEFANSSON: I'm not in a position to discuss. I need to get approval from Amgen to begin to discuss the clinical trials that are ongoing.
MALORYE BRANCA: One of the examples that I saw in the literature-- or in the press, though, were for atherosclerosis.
KARI STEFANSSON: [LAUGHS] Yeah. You see, we have done a fairly detailed study of the way in which you can use, on one hand, polygenic risk score, and the other hand, proteomic risk score to stratify into clinical trials. And when we analyzed, for example, FOURIER, a clinical endpoint trial for Repatha, we demonstrated that if, in the beginning of that trial, the kind of proteomics that we have today would have been available, they would have been able to get away with using only 75% of the number of participants that they eventually used.
KARI STEFANSSON: And because you would be able to stratify people into much higher risk groups by using proteomics than traditional risk factor, and even substantially better than if you use polygenic risk score.
MALORYE BRANCA: You said the proteomics we use today, how is it different from the proteomics we had 10 or 15 years ago?
KARI STEFANSSON: We were not in a position to measure this large number of proteins simultaneously in a very large number of individuals. I mean, let's-- we took, for example, the first-- so the big study we did of proteomics was to take plasma from about 40,000 Icelanders and measure 5,000 proteins in the plasma all of these people. That would have been completely impossible 10 years ago.
MALORYE BRANCA: And that's because the instruments have changed?
KARI STEFANSSON: That is because the technology has evolved, the technology to do this. And there are two leading groups-- two leading companies that are doing this. One of them is Olink, which is a Swedish company that is using antibody methods to measure proteins, and the other is an American company out of Boulder, Colorado that is using aptamers, small synthetic oligonucleotides that bind to proteins.
KARI STEFANSSON: And these are the two companies that are empowering the field to measure several thousand proteins simultaneously in the blood of a very large number of people.
MALORYE BRANCA: Now, you said that there's a paper coming out. Is it-- will your group be on this paper?
KARI STEFANSSON: There are two papers that have been published out of-- that have been submitted for publication coming out to the work on the UK Biobank. One of them is a consortium paper that was written by a group of pharmaceutical companies that participated in this. And the second paper comes from deCODE, which was a separate-- a little bit different approach to the analysis. You see, when you live on a small rock in the North Atlantic, you have a tendency to become eccentric and do things your own way.
MALORYE BRANCA: I can imagine. As having been in this field for as long as you have-- and I've been there too-- what do you see as the inflection points that happened in genomics? The inflection points-- I mean, at one point it was hyped, and then it was not so hyped, and then a lot of companies went out of business, and then a lot of companies reinvented themselves.
MALORYE BRANCA: Certainly, you did.
KARI STEFANSSON: Yeah, we did not reinvent ourselves. From the get-go, from 1996, we have been focused on studying human diversity. In the beginning, we were focused on just genetics, then very gradually we added on other methods to study human diversity. The methods to do human genetics, when we started, were very immature, very weak. It was not very powerful.
KARI STEFANSSON: But we have continued to work on human genetics. The inflection points have come with the technology. The technology has, in many ways, generated the field. First of all, when the ability to do genome-wide associations came with the SNP chips, that was basically-- that was a revolution. It allowed us to go from hypothesis-dependent research where people had candidate genes that almost all of them turned out-- when people found, through candidate gene approach, that a variant in a gene associated with a disease or whatever, and almost invariably, it turned out to be false positive.
KARI STEFANSSON: But once we were able to scan the entire genome through genome-wide associations, we lost the necessity of having a hypothesis. So the freedom that came with a hypothesis-independent approach generated avalanche of discoveries. That basically transformed the field. Then, gradually, these SNP chips started to become bigger and bigger with a larger and larger number of variants that were being created.
KARI STEFANSSON: That was a very, very big one. And people made a fuss, or for a while, were somewhat critical of the common variants that you are picking up through genomic-wide associations, ignoring the fact that almost all of normal human diversity is conferred on us by this common variance. But then came the sequencing ability to sequence a large number of genomes. Now, that's also transformed the field.
KARI STEFANSSON: Now, we can search for the rare variants that have very large effect on risk of disease. And actually, it's interesting, when you look at-- we just went over it the other day. If you ask the question, when you sequence whole genomes, how frequently do you find variants that are considered to be actionable by the American College of Clinical Geneticists, these variants in the genome that if you find them by-- if they are serendipitous findings, that you should report them to the carriers?
KARI STEFANSSON: Basically about 4% of people contain such variants. So the ability to do whole genome sequencing has not just transformed their ability to make discoveries, these technologies are destined to have a remarkable impact on the delivery of health care. I mean, 4% of the population with variants in the genome that confer on you very large risk of a disease that something can be done about, all right?
KARI STEFANSSON: I certainly hope that, within the next few years, we will begin to sequence the whole genome of everyone born in this world.
MALORYE BRANCA: The Amgen/deCODE deal was a huge event in terms of the genomics field. Do you think that there were specific-- I mean, was it your overall program, or were there specific findings that really drew Amgen to you?
KARI STEFANSSON: I think that what drew Amgen to us is that we are just simply the best in the field. [LAUGHS]
MALORYE BRANCA: But you had several noteworthy discoveries right around that time, didn't you?
KARI STEFANSSON: We had a lot of-- I mean, we had made a lot of contributions from the get-go. And I don't want to go into any specific one of them. At the time, if I remember, around 2012, we had been making a lot of discoveries in Alzheimer's disease, for example. We had been making discoveries in cardiovascular disease. We had been making discoveries in a large number of cancers. We had been making discoveries in inflammatory diseases.
KARI STEFANSSON: I think it was the-- I think it's basically the overall competence that was reflected in what we had published. We had extraordinarily good informatics. We had extraordinarily good statisticians and mathematicians who had demonstrated that you could manage large amounts of data. I think that was probably the reason. And then when Amgen bought deCODE, Bob Bradway had just become the CEO of Amgen.
KARI STEFANSSON: And we had met him many times when he was a banker for Morgan Stanley in London.
MALORYE BRANCA: Interesting. So what do you think are some of the things on the horizon for genomics?
KARI STEFANSSON: I think it is very likely that this kind of mining of data on human diversity-- and you'll notice that I'm constantly trying not to use genetics in this because I think this is more than just genetics. I think it is using proteomics and transcriptomics. I think it's going to dominate drug discovery and development. Because when you begin your drug discovery and development on the basis of data on human diversity, you have the human validation, which was most often an extraordinarily difficult step to take when drug discovery started with animal models, et cetera.
KARI STEFANSSON: You don't need any sort of conversion coefficient to get it into humans. It turns out that Homo sapiens is the best animal model for man. So it is-- and when you think about it, when we have data now on hundreds of thousands, and in some instances, of over a million individuals with a particular disease, your ability to basically use these experiments from nature to figure out the biochemical pathway has become astonishing.
KARI STEFANSSON: All right, so with this-- basically, the-- currently, I think that there is hardly anything about the human condition that is not possible to figure out by studying a large number of people and look at what makes them different in all kinds of aspects. And one of the things we desperately need to do is to begin to use these methods to learn more about the function of the human brain, which is sort of the last frontier.
KARI STEFANSSON: Because, remember, not just if you're thinking about diseases of the brain, when you think about all of the common diseases around because they have both genetic component and environmental component. And what directs us into an environment and what makes us avoid an environment is the function of our brain.
MALORYE BRANCA: Right. So have you determined what a minimal data set is for a really effective study? I mean, is more always better, or are there situations where you only need a certain number of samples and you can still make a discovery?
KARI STEFANSSON: You see, this is a question I was hoping you would not ask me because I am extraordinarily greedy when it comes to data. But I can tell you, about 2006 or 2007, we started to work on atrial fibrillation. And we got permission from our ethics committee to start the atrial fibrillation at 11:00 in the morning.
KARI STEFANSSON: And at 2:00 in the afternoon, we had found variants that confer-- a common variant that confers a very large risk of atrial fibrillation based on only 450 individuals. 450 is too small a number to-- it's unreasonable today to expect that you're going to find anything. And one of the things that has started to happen is that with increasingly larger data sets available, we are entering into in an era when we have to be somewhat careful on what we get out of it, because we are beginning to pull out sequence variants that have so subtle effect, so little effect that it becomes difficult to separate it from just what faces us because of statistical fluctuations and stuff like that.
KARI STEFANSSON: So the very, very large amount of data present certain problems that we have to learn to deal with. How can we sort of-- how can we navigate these big data sets without running into what will, in the end, turn out to be false correlations of some sort? But we are learning that as we go along.
KARI STEFANSSON: And the bigger data set, the better, the more of human diversity. That is one thing I have to emphasize, that we continue-- not in our rhetoric but in our actions, we continue to be somewhat [? colonial. ?] Almost everything we have sequenced is of people of European descent. We are dearly missing data on people of African descent. And health care disparity in the world, it begins with lack of understanding, of detailed understanding, of the differences between the common diseases in people of non-European descent compared to what we know about people of European descent.
KARI STEFANSSON: And we have to plug that hole. We have to begin to sequence African genomes in large numbers. We have to begin to sequence Asian genomes in large numbers. And there is an attempt, both within the pharmaceutical industry and within academia, to try to build large collaborations with Africa and with people of African descent in America and elsewhere.
MALORYE BRANCA: So what do you think, overall, deCODE's contributions to the field have been?
KARI STEFANSSON: Still another question I don't like to answer. It's come to-- we have been at this for 26 years, all right? We have contributed to-- we contributed an awful lot to the genome wide association era. We contributed more than any other institution there. We have been contributing an awful lot when it comes to discovering rare variants that correlate with large risk of disease.
KARI STEFANSSON: We have contributed a lot to the study of generation of new diversity in the sequence by studying recombination, gene conversions, and mutations. We have contributed a lot to methods to call variants in a sequence generated. We have contributed a lot to the way in which the narrative has developed.
KARI STEFANSSON: I think we have-- given the fact that we are just a few individuals in a backwater country up in the North Atlantic, I think we can be proud of what we have contributed. When we came here, there was nothing here. There was no tradition. There was no know-how. And we built this up from the ground. And we are happy with our contribution.
KARI STEFANSSON: We probably could have done better, but I think this is a reasonable day's work.
MALORYE BRANCA: And what's your goal next?
KARI STEFANSSON: What is our goal next? It is to-- I think the big question that is sort of burning on our back today is to understand better epistasis within the genome. And we feel that we have managed to find a way to do that, to study it. I think that we have interesting stories to tell there, because that is one of the things that has been missing.
KARI STEFANSSON: Good examples of epistasis within the human genome are very few. And I think we have sort of figured out how to approach that. That is one thing that I look forward to do. I look very much forward to see how our methods and our ability to analyze data, how it is being now systematically applied to everything at Amgen, from drug discovery to development, to the possibility of using it in the way in which drugs are being given to patients in precision medicine.
KARI STEFANSSON: So we have a lot of exciting both basic science processes ongoing, we have a lot of processes that have to do with applied science. And then it is always this big question of, how can we contribute to man's understanding of himself?
MALORYE BRANCA: That's fascinating stuff. Thank you so much for your time.
KARI STEFANSSON: Thank you. Bye-bye.
MALORYE BRANCA: Thank you. [MUSIC PLAYING]