Scientist Stories: Dominic Kwiatkowski, The Evolutionary Journey towards Malaria Elimination

I'm Dominic Fiakowski, and I work in the Parasites and Microbes Program at the Sanger Institute. And I'm going to be talking about how we're using genomics to understand the evolutionary processes that are going on in malaria and to translate this into a surveillance toolkit that will help in, in the process of malaria elimination. And I'm going to structure my talk in 3 parts. Firstly, I'll say something briefly about the challenge of eliminating malaria, then talk about some things we've learned from genomic epidemiology over the past few years, and then talk about how genomic epidemiology could help to eliminate malaria and what are the structures that need to be put in place to make that happen. So malaria is a disease caused by a single-celled parasite called Plasmodium. That infects human red blood cells, and it's transmitted from person to person by the bite of a blood-sucking mosquito called Anopheles. There are 5 different species of Plasmodium that can cause malaria in humans. Most malaria deaths are due to one particular species called Plasmodium falciparum, and there's about 50 different species of Anopheles mosquitoes around the world that can transmit malaria. Every year malaria causes illness in, in over 200 million people and, and kills over 400,000 African children, which is a huge death toll year on year. Beginning of the 20th century, it was in most countries around the world, but over the 20th century, it was progressively eliminated from Europe, North America, and most rich countries, and now it's primarily seated in tropical regions, particularly in Africa. In the old days, malaria was called ague. It was characterized— it is characterized by recurrent fevers. The term 'ague' was a rather nonspecific term, so it might have also been used for flu-like illnesses or even coronaviruses. And 400 years ago, Shakespeare wrote, 'An untimely ague stayed me a prisoner in my chamber,' which is a pretty good description of coronavirus lockdown. We're very focused on how we can develop ways of vaccines against coronavirus, treatments for coronavirus, how we're going to get out of this epidemic. And you might wonder whether we've got such tools for malaria. Actually, there's a fantastic toolkit to fight malaria. It— the first tool came in the 17th century when an old indigenous medicine, a traditional medicine used in South America, cinchona bark, was translated into quinine, which is an effective antimalarial. And even before people knew what malaria was caused by, they were taking quinine for the malaria symptoms. The, the most effective tool we have to fight malaria came from Ronald Ross's discovery of Anopheles transmission in 1899. And although I don't know whether Ronald Ross actually used these words exactly, the method was called draining the swamp. And actually draining the swamps throughout much of the world, and including Europe and North America, has been the way that malaria has been driven down over the last century or more. For example, the draining of the Pontine Marshes around Rome was the key to getting rid of malaria in that part of Italy. So, um, drainage mechanisms and stopping mosquitoes breeding is a fantastic way of doing it. The problem in Africa and many tropical countries is twofold. Firstly, there are many different species of mosquito that are transmitting malaria that are extraordinarily abundant, and they are prolific breeders. And logistically, it's very, very difficult to achieve that level of infrastructure to drain all, for example, all the puddles that may occur after, after a rainstorm. So it's a much— it's much more challenging infrastructurally to get rid of malaria in tropical regions. But we still have some great tools, and in the middle of the 20th century, with the discovery of chloroquine as a really cheap, safe, effective antimalarial drug, and the discovery that DDT was a potent insecticide, that led to a global campaign to try to eradicate malaria. And, and that improved things in many places, but it failed to make a deep inroad into the malaria problem. In Africa and other regions where the tropics, where malaria, which are the heartlands of malaria. And one of the outcomes of that was that after that attempt to eradicate malaria by scaling up the use of chloroquine, scaling up the use of DDT, the parasites started to become resistant to the drugs and the mosquitoes start to become resistant to the insecticides. So now we had a parasite and a mosquito population that was changed. It was becoming resistant to these chemicals. Recently, there have been some great advances. The discovery that artemisinin, an ancient Chinese medicine, is a potent antimalarial, a discovery that won two of you the Nobel Prize recently. That, that's a fantastic compound as an antimalarial, particularly when combined with other treatments in artemisinin combination therapy. And pyrethroids, which stem from natural pyrethrin pyrethrins are now very widely used as insecticides, and they're particularly efficacious when used as part of long-lasting insecticidal bed nets that people sleep under. That's a— that's probably the most effective way of reducing malaria in Africa we have right now. So there are tools, but the problem is that the parasites are now becoming resistant to the artemisinins, and the mosquitoes becoming resistant to pyrethroids as well. So we have a problem of a population that's adapting to these wonderful tools we've got. And one of the great challenges of malaria elimination is which is going to win. Are our interventions going to lead to extinction of the parasite population, or are they actually going to lead to adaptation of the parasite and vector populations and the whole thing runs out of control? So genome evolution then becomes central to the problem of controlling malaria. The Plasmodium populations continued evolving to at least resist antimalarial drugs, and I don't have time to talk about it in detail right now, but they're also evolving to evade the human immune system, which is a massive challenge for vaccine development. And also the Anopheles populations continue evolving to resist insecticides. Meantime, actually, in the tens of thousands of years that we've been exposed to malaria, human populations have developed genetic mechanisms of resistance to malaria, and understanding those could possibly inform vaccine development. So if we come back to the malaria life cycle between parasites, parasites going from mosquitoes to humans, and we think about the, the three genomes involved there, understanding the variation in those genomes could give us many clues about what has gone on in the past in terms of developing human defense mechanisms and what's going to go on in the future in terms of how parasites and mosquitoes are going to adapt to the interventions we're trying to put in place to extinguish them. So, um, 3 very interesting genomes. The malaria parasite genome, Plasmodium, is about 20-odd megabases and has 5,000 genes. The Anopheles mosquito has a genome of about 10 times bigger than that and 13,000 genes. And then the the human genome about 10 times bigger again, and 20,000 genes. And they've each got millions of different variants in those genomes. So the research work that we've been doing over the last 10 to 15 years has been trying to understand how variations in all those populations— the human, the parasite, and the vector— impacts on the malaria disease and on our ability to control it. Our starting point was to build a framework for collaboration with groups around the world in malaria-endemic countries, initially research groups who were doing different research projects on clinical studies, but also studies on parasites and on mosquitoes, and to work with them to try to embed genomics into their studies so that we had a framework that would allow us to address a number of interesting questions. And part of that process was developing policies for equitable data sharing and also building systems for local training capacity building so that everyone could collaborate fairly across this global network. We also had a lot of technical challenges— how you actually do this sequencing of parasites in mosquitoes— and a lot of analytical challenges. But a key part was building the framework for collaboration, and that's this network this very wonderful network that we're involved in, of partners around the world called MalariaGen. And a crucial part of the network was, wherever possible, producing open data in a way that was equitable to all the partners. So I now want to just talk a bit about some of the things we've learned from genomic epidemiology over the past few years. So I'm going to talk very briefly about work we started off 15 years ago, looking at the human genetic determinants of malaria severity and the question was, in populations where everyone's repeatedly infected with P. falciparum, the most severe form of malaria parasite, why is it that some people developed severe diseases and others don't? And of course, the same sorts of questions come up with COVID Why do some people get severe disease and others not? Are there genetic factors involved? And with partners in 11 different malaria-endemic countries, we performed a large genome-wide association study, and we've been publishing results of that over the last few years. I want to just tell you about one particular result, one particular finding that was informative. By the way, I should say as a background that you all know, I'm sure, that people with sickle cell trait, hemoglobin S, are protected against malaria. They have about a 10-fold protection against getting severe disease. And that was the biggest single signal we got out of our genome-wide association study. Another factor that came out of our genome-wide association study was that blood group O is protective against severe malaria, and there are some suggestions in early preliminary evidence that actually blood group O is protective against COVID. So you wonder what's going on there. But I'm going to talk to you about a different locus that we discovered, which is the glycophorin locus. And what this cartoon shows you is, is a region of the genome extending in this case over about half a megabase. In which there's a cluster of 3 different genes: glycophorin A, B, and E. And we know that glycophorin A and B are expressed on the surface of the red cell, and we know that they're used as receptors for invasion by malaria parasites. And we found a particular variant in the genome that is protective against severe malaria. One copy gives about 40% protection against severe malaria, and the homozygotes give even more protection. And it's a rather strange variant. It's one in which the glycophorin A is intact. The glycophorin B gene is removed, but it's replaced by 2 copies of a hybrid gene. And that hybrid gene looks like glycophorin B on the outside of the cell, but it looks like glycophorin A on the inside of the cell. And this is a huge structural variation. You might think something as large as this would be very easy to detect. It was actually extremely difficult to pick up the signal from genome-wide association studies because the genome was so muddled in this area because the structural variant I'm telling you about is just one of many structural variants that occur in African populations. African populations have a very high frequency of structural variation in this region of the genome, which makes it rather difficult to see signals of genome-wide association in there. Anyway, we discovered that, and it turns out that this particular variant does indeed reduce the efficacy with which parasites can invade red blood cells. And a recent paper by Sylvie Cariouki and Julian Rayner's group, just out a few months ago in Nature, has shown conclusively that this so-called Dantu blood group that's characterized by this structural variation does indeed reduce the invasion of red cells by malaria parasites, and the mechanism appears to be because it alters red blood cell tension. We don't know what all these other glycophorin variants are doing in the human population. We assume that they're probably something to do with malaria. We haven't found the answer ourselves, and we're wondering whether it's something to do with the parasite genetic type. And the next range of questions that we are working on with our collaborators is what's the interaction between human genotype and parasite genotype in determining red cell invasion and disease severity. So having just given you that taste of what's going on in the human genome, I want to move now to, importantly, to address the question of what's going on in the parasite population. And we've been working with partners around the world to assemble a large data set of parasite genomes that have been sequenced from different countries. Our most recent data release, uh, we call it PF6 release because it's the sixth version that we've produced, is of 7,000 worldwide samples from 73 worldwide locations, uh, with 49 different research groups. And in that data set, which is available openly now, there are curated genotype calls on 3 million SNPs and short indels in the parasite genome, plus copy number variations associated with drug resistance and structural variants that cause failure of rapid diagnostic tests. So it's an important reference data set for looking at all of those important types of genotypes. Almost every sample that we sequenced in this set had evidence of resistance to at least one antimalarial drug, and in Southeast Asia, some of the parasites carried markers of resistance to, to 6 commonly used antimalarial drugs. So multi-drug resistance extremely common in that area. Also interestingly, genes expressed during the mosquito stage of the parasite life cycle are very prominent amongst the loci in the parasite genome that show evidence of strong geographic differentiation, suggesting that one of the major evolutionary forces on the parasite genome is what is the vector species locally. Now, a crucial feature of this collaboration, this merging collaboration to build these large open datasets, is that we work with partners around the world and networks around the world. And we're very proud to be associated with the PDNA network, a network of African scientists led by Abdoulaye Jimdi and other researchers around Africa with whom we've been working now for well over a decade. And As part of the dataset I've just told you about, Jim Dee and colleagues led by Alfred Mambutungua produced a wonderful analysis of subpopulations of Plasmodium falciparum across sub-Saharan Africa that came out in Science a year ago. And amongst other things, it showed very clear differentiation between certain parasites in certain parts of Africa, including in the Horn of Africa And these differentiations between parasite populations seem to be associated with human demography. So it's a fascinating set of data about how parasite population structure may have historically been driven by human population structure. And also in this paper, Jim Dee and Alfred and colleagues have brought out a lot of signatures of interest for studying drug resistance, signatures of recent selection. I want to just focus on one aspect of the data set that we, we think is very important, and that's the question of how do drug-resistant parasites emerge and recombine to produce multi-drug-resistant parasites with very high levels of biological fitness. And one of the crucial discoveries that's come out over the last decade— it was suspected before, but it wasn't really nailed— was that mutation, drug resistance mutations don't just— don't— well, put simply, they're happening all the time. A single individual who has malaria, their genome is mutating, and if you look at the mutation rate, it's likely that any infected individual is going to have mutations in almost every position of the genome every few days. So we know that drug resistance mutations are happening, and when a lot of selective pressure is put on the population by high levels of drug usage, then typically what is seen are so-called soft selective sweeps, where there are multiple origins of resistance. It's not a single resistant parasite that comes out. Even if there's a single resistant locus, there'll often be multiple independent origins of resistance at that locus. So in the case of artemisinin resistance, which started to emerge in Southeast Asia in the order of 2007 or so. That has proceeded in 3 stages. Firstly, there was a soft selective sweep where we could see at least 30 independent mutations of the Kelch 1 gene that confers artemisinin resistance. And then starting quite shortly after that, a particular lineage slowly but surely started to take off. One of those mutations seems to have outcompeted all the others and seems to have been particularly durable. And very early in that process, it became associated with a lineage in a different part of the genome that confers resistance to the partner drug that's used for artemisinin in the, in the form of artemisinin combination therapy that that was being largely used in Cambodia, those two lineages came together in a single genome. So the artemisinin resistance type associated with a type of plasmepsin copy number variation that caused this combination of resistance to the two drugs, and that took off. So what we've got here is a winning lineage, and that has progressively spread across Southeast Asia, and now that lineage really dominates parasites in a large part of Southeast Asia, including Cambodia, Vietnam, Laos, and parts of Thailand. And as that lineage has spread, it's further differentiated. Now, these aren't lineages as in viruses and bacteria that are very discrete lineages, because the malaria parasite is recombining all the time. So there's a very muddled population. When we say a lineage, it's actually a rather muddled population of parasites, but what they all have is common ancestry at particular drug resistance loci. And on top of this, they're acquiring mutations in other parts of the genome. So for example, this lineage is acquiring multiple mutations now in the chloroquine resistance gene, and that seems to amplify the resistance to this form of artemisinin combination therapy. It's a very complex and progressive evolutionary process. We really need to understand how this operates, why is it that some lineages are winning. And obviously that has two parts. We'd like to understand the molecular basis of this, what phenotypes are being acquired as progressively mutations grow. But also we really want to understand what are the sort of epidemiological public health forces that are leading to that? Are there things that we could do differently that would make it less likely that that particular winning lineage would have taken off? So for that, we really do need much more time series data that allows us to see exactly how the population is evolving and align that with what we know about what interventions are going on and what other epidemiological events are going on at the time. And as I was saying, in a few minutes, at the center of what we want to do in the future is to understand, to do longitudinal surveillance, so we can really look at that process of evolution and how it maps on to what's going on in public health and disease control. As part of this process, we've been taking the genomic findings we've got and trying to translate that into simple tools that could be used by, by national control programs to, to help them to monitor what's going on in the population and to see what's happening when they're rolling out a particular therapy. Is it going in the right direction or not? And the data here I'm showing here are from a, a collaboration called GenRay that's led by Oliver Myoto in Bangkok, and he works with national malaria control programs throughout that region of Southeast Asia, and working with them to embed the collection of small dried blood spot samples as part of the control process, part of the malaria control process, and then to get those into a lab where they can be rapidly genotyped and results fed back to the programs. And what these figures are showing here in red are parasites that are resistant to this particular form of artemisinin combination therapy, dihydroartemisinin piperiquine, and in green those that are still sensitive. And you, you could— what you can see here is how in 2017 this resistance to this form of ACT was spreading up from western Cambodia into Laos and across to Vietnam. And the next stage of this process is to work with national control programs to build capacity for amplicon sequencing in their own countries. And we're starting that process already in Vietnam and in various parts of Africa, getting local capacity for sequencing so that can happen locally. Now I'm going to move to, um, the mosquito. Um, and here we have another a large data-sharing collaboration known as the Anopheles gambiae 1000 Genomes Project, and that's provided an extraordinary window into what's going on in the mosquito population. This particular project focuses just on one particular mosquito species complex, the so-called Anopheles gambiae species complex, that includes also Anopheles coluzzii, Anopheles arabiensis. There are other important malaria vectors in Africa, but Anopheles gambiae is by far the most, the most effective malaria vector in Africa. Mara is working on Anopheles funestus, and there are other important species we'd like to understand as well. Our current data release has over 1,000 mosquitoes from 13 countries, and an extraordinary feature of the mosquito genome is, is how how diverse it is. We find more than one SNP in every two SNPs. The nucleotide diversity is about 2%. In other words, if you took any two random mosquito genomes, like, uh, you would find they on average had about one difference in every 50 bases, which is well over an order of magnitude greater than that we see in humans. So in the case of Anopheles, one of the challenges we've got is not just sequencing their genomes, but in parsing all the complex variation we see there. Haplotype phasing is particularly challenging when you've got such high SNP density. Currently we're calling data on, on almost 60 million SNPs and also copy number variations as well. Now I'm going to focus on— well, I'll begin by saying that there's an awful lot one can get out of such a rich data set. Of such a genetic diverse population, so much one can learn about population structure, gene flow, migration, evolution. But I'm going to focus, in the interest of time, just on the issue of insecticide resistance and what we can learn from genomics about that. And the crucial point to get across is that the mainstay of malaria control in Africa is the use of bed nets that are treated with insecticides, long-lasting insecticides, particularly pyrethroids at the moment, which you sleep under. And that's because in Africa, most malaria transmission happens at night and indoors. It's slightly different in other parts of the world, parts of Southeast Asia, where much more malaria transmission happens during the daytime and outside. But in Africa, particularly with Anopheles gambiae, the transmission tends to be at night and indoors. So it turns out that sleeping under a bed net is, is, is a massively effective way of reducing malaria. And you can think of it as being very equivalent to what we're seeing in COVID, which is to do with social distancing. It's, uh, it's trying to reduce transmission. It's not— in this case, it's not, it's not impacting on the mosquito population, it's not reducing the numbers, but it is stopping the process of transmission. Also, indoor residual spraying with insecticides also has a part to play in malaria control in Africa, but the single biggest thing is insecticide-treated bed nets. And now the mosquitoes are becoming resistant to the pyrethroids that are used in the most common form of bed net. And what we're seeing here from the Anopheles gambiae data, 1,000 genome data is analysis of a particular gene, the voltage-gated sodium channel, which is the target of the pyrethroid insecticides. It's also the target of DDT. In fact, most insecticides are neurotoxins for insects. And what we can see in the slightly complex diagram is that there have been multiple independent origins of pyrethroid target site resistance across Africa. We can see their independent origins because they're— although some of these are exactly the same mutation in exactly the same position of the genome, the same position of the gene, they lie on very different haplotypic backgrounds, indicating they're of different ancestral origins. And we can then mark these different origins, these different lineages of resistance. We've given them named like F1, F2, S1, S2, and we can see how they spread across Africa. And in the case of the F1 lineage, we can see that that particular origin of insecticide resistance, of pyrethroid resistance, has spread right across the west of Africa from far down, quite far down south, and right across the tropical rainforest, and has gone into other parts of West Africa and has introgressed between two sibling species, Anopheles gambiae and Anopheles coluzzii, sibling species, and yet this— there is gene flow between them. And here we can see that the insecticide resistance is passing between these sibling species. And just as I was discussing with resistance to antimalarial drugs, we can see that as these insecticide resistance lineages evolve, they're acquiring different mutations which are possibly strengthening the phenotype. Now, insecticide resistance has multiple mechanisms. I've just been talking about target site resistance, but also the mosquitoes can develop metabolic mechanisms to destroy insecticides, and one particularly important form of metabolic resistance involves cytochrome P450 genes And here's one particular cytochrome P450 gene, CYP6. And what we're seeing here on the right-hand side is a 60-kilobase region of the Nautilus gambii genome, which has large copy number variations, duplications of the cytochrome P450 that's leading to amplification of its expression and therefore an increase of its ability to resist pyrethroids. And what we're seeing in this map here is how we can then parse the genomic data to look at both target site resistance in the voltage-gated sodium channel and metabolic resistance in these cytochrome P450 genes and put that information together to ask where there's a combination of target site and metabolic resistance. And that's of considerable practical importance because as pyrethroid resistance spreads, there are now efforts to bring on board new insecticides and new insecticide combinations. And one particularly important combination is combining pyrethroids with a combination called PBO, which actually can reverse the metabolic resistance. But to deploy those bed nets effectively— and they're much more expensive than standard ones— you need to know where that's most likely to work, and that's most likely to be effective in places where there's a combination of target site and metabolic resistance. You might say, aren't there better methods to do this than genomics? And you might think there would be, but it's turning out that biological assays of these things are really quite challenging in the field. And genomics is now looking like probably the most effective way to screen for these different combinations of resistance. And therefore, probably in the long term will be part of the surveillance toolkit, or should be part of the surveillance toolkit for national control programs. Another aspect of genomics is you can look at, you can look for new evidence of emerging resistance to insecticides that you didn't know about, and we can look for signatures of recent selection. What we're seeing on the right-hand side here are various plots looking across the genome in different populations. Each line is a different population, population of mosquitoes in a different location or for a different sibling species. And we can see them lining up at particular regions in the genome, and they correspond to signatures of recent selected resistance against target sites, including the voltage-gated sodium channel I've talked about—that's the really big one—but also the GABA receptor, metabolic resistance including cytochrome P450s but also glutathione S-transferases, and also actually locus here that's nothing to do with insecticide resistance, but it is more to do with innate immunity. So this is a way we can pick up emerging signatures, and that will be really important as new insecticides are rolled out, quickly being able to work out whether there's a new region of the genome involved in resistance to that. So I'm now going to round up by just talking about how could genomic epidemiology help to eliminate malaria, and what are the structures that need to be put in place. And it's quite clear that in order to manage drug and insecticide resistance, we do need to be able to map resistance to know what control tools are effective and actively manage control tools to prevent resistance. So if there are multiple combinations of drugs or insecticides that could be used, which will be effective where and when, and understanding how they're affecting the population as they're being rolled out. And that's going to require a complex structure. It's not as if whole genome sequencing will be the answer to the whole thing. It absolutely won't. We need very simple, low-cost tools, very specific genetic markers that can be used to test for particular sorts of resistance, and they should be rolled out very widely and that should be available in local clinics and should be massively decentralized. But then there isn't a value in having a reference resource of high-quality genomes that allow us to inspect details of what's going on. And certainly as long-read technologies come along, that will massively help us to understand what's going on in very complex regions of the genome. And that sort of work is naturally going to be somewhat more centralized in specialist labs, and there's all the more reason for that to be open data that is shared with everyone and to have networks that allow that data sharing process to happen. And in between the two, there's a gray area of stuff that methods like amplicon sequencing, where in some cases it should be massively decentralized, but in other cases there'll be some value in bringing that data together into common open data datasets that everyone can share. In addition to managing resistance, a massively important role of genomics is targeting interventions for maximum impact, mapping transmission hotspots, and defining feasible zones of elimination by understanding population structure. And I haven't had time to talk much about that today, but certainly genomics is a really important tool for understanding processes like population structure, migration, that can be really valuable in understanding where do I now target the limited resources I've got so I'm going to have the maximum effect on transmission. Another thing that's coming down the pike is new technologies for vector gene drive, which are designed to disable the mosquito population. And genomics is central to that, firstly in designing gene drive targets, and in managing and monitoring the spread of gene drive, which is rolled out in years to come. A crucial part of bringing genomics into the practice of disease control is actually understanding what are the use cases. As genomic specialists, we focus on the scientific interesting things, but we have to understand how we make the leap from the scientifically fascinating stuff to basic decisions like, what drugs do I purchase next year? Or how do I control this particular resurgence of malaria? How do I intervene for maximal impact? And this type of thinking actually has been really interesting as we at Sanger have also been working on COVID. Obviously, many of these types of questions are rather similar to the problems we're seeing right now in trying to tackle the resurgences of COVID under— understanding how does one get maximal impact out of limited resources. And so we need ways of mapping onto our operational use cases the analyses that will be most informative to them. Things like, in this case, the frequency of drug resistance, or mapping transmission and intensity, or locating transmission hotspots. So in the middle section here, we've got analyses that we might do And we need to understand the mapping process that works between those analyses to the operational use cases. And then over on the far left is the question of what samples we need to collect and what genotyping methods we'd need in order to serve those analytical use cases. So we have a workflow whereby our sampling strategy and our genotyping methods are leading into analytical use cases that then are producing use cases that are of value for disease control. So we've been giving a lot of thought going forward into how we can gear the science we can do for translation, and it's clear that there has to be some sort of link between scientific research, product development, and providing a surveillance service, working with different partners in each of those areas. But at the center of it all, producing a large open data set of analysis-ready data that's ready for analysis for both scientific purposes, but also public health purposes. Systematically collected time series data is really the thing that will transform our ability to infer epidemiological and evolutionary processes by genetic analysis, and it'll enhance the utility of genomic surveillance as a tool for disease control. We're systematically collecting data, and, and to— in order to systematically connect data, we absolutely have to work with national disease control programs and research groups in lower and middle-income countries, to work with them to develop genomic surveillance toolkits that are fit for their uses. And we have to help them build capacity for in-country genome sequencing and data analysis backed up by standardized software and cloud data infrastructure. So at Sanger, we're a bit of a journey, and I've talked about malaria, but it's true for the whole of the parasites and microbes. Program where we've also got work going on, on schistosomiasis, on worms, on cholera, on pneumococcus, and on the microbiome. And in all of those areas, uh, we're making a transition from what we did 20 years ago, which has produced reference genome sequences, and what we've been doing for the last, you know, 10 or 15 years, which is understanding natural variation in populations, what the phenotypic correlates of that, and how that maps onto epidemiology.. And the journey we want to now make in the next decade is to analyze time series data at a global level to understand the evolutionary process. And, and the central vision of the Parasites and Microbes program going forward over the next decade is to work with our partners around the world to generate longitudinal genomic and metagenomic data on microbes, parasites, and vectors, including malaria parasites and vectors, to analyze how those populations are evolving over time and space and in response to disease control interventions. And to do two things out of that. Firstly, to advance the scientific field of evolutionary dynamics, and also importantly, to translate that knowledge into decision support tools for public health programs. Well, finally, a feeble slide to thank all the wonderful team who did the work. I apologize to you all that I've not managed to name or picture you all in this short version. But thank you so much, all those at Sang Oxford and Bangkok, including all those that aren't on this picture—Tamara, Oliver, and others—and also to all our wonderful partners around the world in the MalariaGen network and the many networks that work with MalariaGen. Thank you very much.