Scientist Stories: Mark Blaxter, Genomics and the Tree of Life

My name is Mark Blackster. I'm from the Tree of Life program at the Wellcome Sanger Institute. It's my pleasure today to present you with an overview of what we're doing in the Tree of Life program, some of the work we've been doing so far, and what we plan to do in the future. So I usually start these overview talks with a quote from Darwin. Now this is one of my favorite quotes from Charles Darwin. It's from his Origin of Species in 1859. After he'd considered his tangled bank of creatures interacting with each other and evolving, he puts forward this realization that evolution has happened, is going on now, and will happen in the future. And it's this realization that evolution has created endless forms in the past, is creating them now, and will create them in the future, is, if you like, behind our hopes for the Tree of Life program. We hope to understand the evolution of these endless forms, not only to understand how they came to be, but also what the processes are that are acting in real time now that are generating them. The other thing that happened a long time ago, in fact 100 years before The Origin of Species, was this other amazing book from Carl Linnaeus. So this was the start of what I call one of the most successful science megaprojects ever, bigger than the Hadron Collider, bigger than anything you can imagine. This has been going for 300 years, nearly. This has been going globally. This has been a huge collaboration between tens of thousands of people. I hate to think what it really cost. But what it's given us is this universal good, this way that anybody on the planet can talk about species using a universal language. And that's what we hope to do with Tree of Life as well. Is to think about producing a universal good for the planet that we can all use to talk about biodiversity. So since Linnaeus's first volume of the Systema Naturae, we've identified a lot of species, and the Catalogue of Life currently stands at about 1.9 million species, of which about 1.5 million are eukaryotes, that's organisms with a a nucleus. There's a bunch of extinct species in there and a few thousand bacterial species being named. So we have this huge catalog of diversity, and one of the aims of the Tree of Life project is to take this catalog, which is like a catalog to an amazing library of biodiversity, and actually print the books, actually write the books, write out the books, which is the DNA sequence of the genomes of all of life's diversity. So why do this now? There are a couple of technical reasons why to do it now, but actually the real reason for doing it now is because of imminent danger. So we now as humans have suddenly realized that our biodiversity, our planet's biodiversity, is in crisis. We're into the sixth great extinction where anthropogenic change has likely caused the increasing rate of extinction of species across the planet, and that climate change, which is largely anthropogenic in origin, is driving major changes in ecosystems which threaten ecosystem survival. This picture is taken from Kew's State of the World's Plants and Fungi 2020, and it's unusual for Kew to put a forest fire on the front of one of their books. They usually put beautiful flowers or amazing forests, but they quite rightly put this forest fire on the front of their 2020 State of the World's Plants and Fungi because that's where we are. This is a map from that publication just earlier in the summer showing where we know that plant species have gone extinct within recorded time. Obviously Many species will have gone extinct without us noticing them. Many species may have gone extinct before we've even described them, and this is not that many species when you think about how many species of plants there are, somewhere around 400,000, but each one of these species is a unique experiment in evolutionary biology. It's a unique experiment by the biosphere in how you can be a plant. Or a fungus. And it's really sad to put up the names of these taxa. These are some of the species that have gone extinct. Each one of these we will not see again on this planet. Each one of these was a unique way, and unique successful, except for the fact we drove it extinct, way of being a plant. And some of them survive and may survive for a while in plant pots in botanic gardens. But, for example, for this euphorbia, that's it. That's the only one left. And I can make these tombstones for any number of groups. So here's a group of larger animals that have gone extinct. We all know about the Tasmanian tiger or thylacine, which which was driven to extinction by hunting, and the last one died in zoos. We all know about the manatee, the stellar sea cow, which died out because we overfished the sea otters and caused an ecosystem collapse, which meant that the sea cows, which thrived on the kelp, could no longer survive because the kelp was gone because the sea urchins had eaten it all. The sea urchins were super abundant because we'd taken all the otters for their pelts, and they were the ones that preyed on the sea urchins. So we'll never see these organisms again, and we want to be part of a process that helps stop this happening in the future. Another thing I'm really aware of is that these are what I call charismatic megabiota. These are the really big things, the large organisms which lots of people can recognize, which you can spot from a long way off. But what about the smaller ones? How are they doing? Well, in fact, they're not doing very well either. So we really don't have good catalogues of how many insect species, for example, have gone extinct in the last while, but we do know that insect populations are in extreme danger. And so again and again in the news there are comments from ecologists and environmentalists and research scientists across the globe pointing out that however you count it, insect populations are under great threat. And if you think about it, of that 1.5 million eukaryotic species, about 1 million of those are insects. If insects are in trouble, the whole biota is in trouble. I've spent most of my research life working on tiny little organisms, much smaller than insects even. So the no-see-ums, if you like, the microscopic life. And microscopic life is hugely abundant and hugely important to the functioning of ecosystems. So while we might want to preserve an ecosystem because it has a stellar sea cow or a tiger as its charismatic major proponent, if you like, it's the small organisms that make these ecosystems work. So these small organisms are of real importance to humans. And there's a lovely picture from the Wellcome Picture Library of when people were just inventing microscopes, finding a microcosm under the microscope. This is monster soup in Thames water. This is a protest about Thames water being so contaminated. But these small organisms— this is a bunch of plankton from the open ocean— These small organisms actually sit right in the midst of the food chains and food webs on which the rest of ecosystems survive. And we don't know about these organisms. We don't know how at risk they are. We have no measure of the threat to them. And yet the way they make their lives is just as astounding as the way a tiger or an endangered euphorbia makes its life. So the question is whether this wave of extinction that is coming on us, whether we can do anything to mitigate it, or if you like, whether we can not quite ride but go with this wave and use the renewed excitement and interest of people in biodiversity and in its conservation and preservation to help, using genomics, mitigate the effects of the sixth great extinction. There are other reasons for doing this genomic sequencing, I'll come to them later, but this is one which really motivates us. It also motivates a global project called the Earth Biogenome Project. So this was launched 3 years ago with the goal of sequencing every eukaryote on Earth in a staged series of phases, starting with a representative for every family of eukaryotic organisms, working through the genera, and finally sequencing all species. So this is an international consortium project where individual groups bid in or join and promise to deliver what they can to this global project. The Earth Biogenome Project folks, in the paper they published in PNAS, made some very cogent arguments as to why we should do this project, why we should sequence all life now. One of the exciting things they point out is that it would cost about the same to do all life now as it did to do the human genome the first time around. So 20 years ago, the Sanger Institute and many others participated to sequence the human genome for the first time. And that cost something like $4 billion. And that's the estimate of what it'll cost to sequence all life on Earth. So in an equivalent cost, we can get sequences of everything. So that tells you something about the change in the technology that underpins what we want to do. But really, we think this is a good idea. This is worth doing because we can do amazing science with it. So for example, build the trees of life, understand how everything is related and how genes have transmitted through generations vertically, but also been transmitted horizontally. Understand, if you like, some of the basic rules of evolution, and by careful selection of taxa and looking at taxa in particular areas, understand how the biosphere has responded to and may respond in the future to climate change. The data we produce from these genomes we hope will be really useful for things like biomonitoring and conservation. We'll be able to track and trace invasive species very effectively, We'll be able to track and trace endangered species, including promoting genetics-led breeding programs for endangered species. We'll also be able to do something really exciting, which is whole ecosystem conservation genomics. Rather than just looking at an apex predator or some charismatic piece of the megabiota that we use to represent an ecosystem, we can look at the whole ecosystem. We think also these data will be really useful for society. So Improving the health of the ecosystems and making sure that the ecosystems we rely on, including agroecosystems, don't collapse will help society, is part of societal health. But the other thing we want to do is provide raw feedstock, if you like, raw data that go into bioengineering and synthetic biology. So looking at a post-oil economy, for example, what do we need? What biochemicals? What synthetics do we need that we can make using the new enzymes and systems we'll identify in all these amazing organisms we will sequence? So that's all a series of arguments that were made by the Earth Biogenome Project in their original paper, and it's ones that we sign up to. So the Tree of Life program at Sanger was started just over a year ago. And we're one of the 5 programs at Sanger. And what we aim to do is better the human condition through large-scale genome sequencing across diversity. And so we're going to do that with 4 major parts of our work. One is we're going to innovate technically in terms of how we can produce these genomes. I'll spend much of this talk talking about that since we just started. Deliver genomes across the diversity of the tree of life. So by deliver I mean make them public, make them accessible. We'll contribute to global projects delivering towards this goal of the Earth Biogenome Project, and then also do particular unique science based on these genome sequences. So we have individuals inside the faculty who are interested in, for example, speciation, and are interested in ecological interactions between species, and how that drives species distributions. Interested in the evolution of symbiosis, and especially of parasitism. And also interested in the pattern of process in the evolution of genome structure. That's why do we have the number of chromosomes we have, and why are the genes that are on those chromosomes on those chromosomes and not elsewhere. So that's our program goals. Inside that program, we've got a number of collaborative projects that we take part in. Very often we are the sequencing or the informatics analytic partner, because the Sanger Institute is not a biodiversity institute traditionally. We sequence a lot of things, mostly humans, but many other things. But all these specimens and samples have come from outside. So with people across Britain and Ireland, we're collaborating in the Darwin Tree of Life project, which aims to sequence all the eukaryotic organisms in Britain and Ireland. That's about 60,000. We have a project funded by the Gordon and Betty Moore Foundation which is working globally with partners to look at genomes involved in symbiosis in the aquatic realm. So marine and freshwater, all these different associations that happen between eukaryotes and other organisms, whether eukaryotic or prokaryotic. Often this is in the form of symbiosis that is involved in photosynthesis. If you think like corals, that's between a cnidarian, a jellyfish relative, and an alga. Or also for metabolic symbiosis, where the two organisms together can do something metabolic that they can't do apart. We're collaborators in the Vertebrate Genomes Project, which is one of the major projects of the Genomes 10K Consortium, which aims initially to sequence a reference genome for each order of of vertebrates on the planet, and then go ahead and sequence one for each family and each genus. And all these, if you like, are parts of the Earth Biogenome Project. So I'm going to talk first about the Darwin Tree of Life— well, I'm going to talk mostly about the Darwin Tree of Life project. This is the one we've had running for the longest. It started in November last year after a couple of years' worth of gestation and discussion across all the partners. And I'll talk about the data we've been generating for that and how well we are doing. So we call it the Darwin Tree of Life Project because of the way Darwin did his science. Darwin spent most of the latter part of his life at his house in Downe in Kent, and he wrote his book The Origin of Species there. He wrote all his books there, and he basically did research in his back garden and corresponded with people globally in order to understand the world. And he managed to understand evolution and the way the world works from the work he did in his back garden. And we think we can do the same by looking at our back garden, which is Britain and Ireland. So this little archipelago that sits off the west coast of Europe is not known for its biotic richness, but it has got a very well-known history. And very importantly, it has very well-known biota. So we know very well both the species list and where those species are found across these islands. And even more importantly than that, we have people who know where to find them. So the tradition of natural history, both academic and amateur, is really strong in these islands. And so we can actually find all these species. So with this behind us, we set up the Darwin Tree of Life project. Darwin Tree of Life project is a collaboration between biodiversity institutes, so things like the Natural History Museum, Royal Botanic Gardens at Kew and Edinburgh, the Marine Biological Association. And a very important site for us is the Wytham Woods Field Ecology Site, which is owned and managed by the University of Oxford, where they've been following a particular ancient forest for— ancient woodland for many decades. And then we're also working with analytic partners, so Sangha will be doing a lot of analysis, but we're going to be collaborating with University of Edinburgh, University of Cambridge, and the Earlham Institute to analyze the data we get and turn them from raw genome sequence into something which has biological information attached. And then our data presentation partners, if you like, are EMBL-EBI, and especially the Ensembl group at EMBL-EBI, who'll be presenting all the genomes we sequence. I should say that the Earlham Institute is also driving a kind of real blue skies bit of our project in collaboration with the University of Oxford, looking at protists, single-celled organisms, where although we have 2,000 or 3,000 species described from Britain and Ireland, the likely true species list is in the tens of hundreds of thousands. And so they're doing a blue skies project which involves both species discovery as well as genomics. The people involved are legion, and at every site we've got a large number of people involved, some of whom are funded— very few of whom in fact are funded directly by the the project, and many of whom are giving their time and energy and enthusiasm pro bono. So there's a lot of contribution both from the Sanger Institute and from our collaborating institutes. The core of the project is funded by the Wellcome Trust through a major discretionary award, 50% of which goes to the partners and 50% of which goes to Sanger to do the sequencing. So our primary goals in the first years are to sequence a reference family, a reference genome for every family of organisms based or found in Britain and Ireland. So this, as I said, Britain and Ireland are not known for being very biodiverse, but interestingly, or encouragingly, of all the families of organisms on the planet, nearly about 40% are found in Britain and Ireland. So if we sequence a family representative from British and Irish taxa, we'll have covered 40% of the diversity at the family level across the globe. This will be a major contribution just from these small islands. Once we finish the families, that's our first goal, about 4,000 genomes, we'll move on to sequencing generic representatives and species representatives, and hoping to complete all 60-odd thousand in about 10 years. At all times, we're also looking to sequence not just one representative per family, but maybe a few if there's something particularly interesting going on, or diving deep into a particular genus if there's an interesting story to be told about genome evolution there. And what we've set up is an amazing pipeline which is now working, which is astonishing, that goes from finding a specimen in the field, identifying it, extracting DNA, doing the sequencing, doing the primary assemblies, curating those assemblies so we're sure they're correct, and depositing those data publicly for all to use. So this pipeline, this engine that we built, this processing factory, is now up and running. It's running at about 24 species per week. It starts in the field where we have enthusiastic people. So only one of the people pictured, of the three people pictured here, are actually employees of the the Darwin Tree of Life program. One is a faculty member and the other one's a colleague looking for things in the field. So these are experts who know what they're looking for and can find the things that we need to sequence. And all these specimens are caught using a very ethical standard operating procedure, fully permitted, And excitingly, we are only taking 1 or 2 or a few specimens per species, because we can sequence whole genomes from single specimens. These specimens are identified in the field, put in special barcoded tubes, snap-frozen at -80 degrees, and shipped to Sanger. The other thing we do with these specimens right at the beginning of the process is DNA barcode them. So DNA barcoding is a process whereby you take a small amount of DNA you amplify up a marker gene. For animals, it's the cytochrome oxidase 1 gene from the mitochondrion, and you use this as a tag to tell you what taxon the organism belongs to. So there's lots of arguments we have, discussions about whether or not DNA barcoding works for every species all the time, but what it does act as is a very quick affirmation that the taxon identification is likely correct, and also the DNA sequence acts as a marker for the rest of the process running through the institute. So we standardly DNA barcode the specimens, and those DNA barcodes will be made public even if we don't actually sequence that specimen that was barcoded. We then have to extract long DNA. We want to make genomes. We want to make really good genomes. And if we There are lots of ways of getting DNA out of organisms that usually results in very, very fragmented DNA, such as is shown in the bottom of these three graphs. However, we need very long DNA, and so we've got an active process of research at the Sanger Institute and with our colleagues globally, in fact, to devise the best methods for getting DNA out of these organisms. Importantly, these methods have to be robust across taxa, so the biology of the organism really affects things. But it also has to be robust across sample size. It's fine if you have a badger— there's a lot of DNA in a badger— but if you have the flea that's feeding on that badger, there's not much DNA in one flea. And that flea is a specimen which we want to sequence just as much as the badger is. So we're developing methods to get DNA out of even very small organisms. And we've— we're looking at how to get the specimens from the field to us the most efficient way, and also trying to make sure that we don't co-extract other things. DNA is just a long chemical, and there are lots of other long chemicals which are charged in organisms, and so it's possible to co-extract those, and those interfere down the line with extraction. So we have a method which we really like now. So the top two graphs show this method really working. We can make good DNA from very small organisms, and that DNA is enough to do the sequencing. So what sequencing are we doing? So there's two kinds of sequencing we're doing for the genome, and then we're also doing RNA-seq for predicting genes. For the genome, we need very high quality base calls. We've got colleagues out in the field collecting these organisms, which may or may not be rare, and it costs time and money and enthusiasm and buy-in from those colleagues to get those specimens. We don't want to have to do it again. We don't have to send them out again to catch that species once more. And because of that, we want to do the best we can off the specimen they deliver. So we don't want to produce a low-quality genome, we want to produce a very high-quality genome. So we need really high-quality base calls, so that every base that we call in the final assembly is correct. The other thing we need is high-quality linkage between sequences. It's quite possible to generate— and I have generated in my career previously to joining the Sanger— many genomes which have high base quality, but which are extremely fragmented. So these fragmented genomes are great for exploring gene space. They're great for building genetic maps. But they don't give us the contiguity. They don't give us the contiguity between genes to know how genes are organized on a chromosome. They don't allow us to associate promoters or enhancers with genes, because the fragmentation of the genome means that that information is lost. So we need high-quality base calls, we need high-quality linkage between sequences, so that we can go from our individual specimen through to a chromosomally complete genome. So by chromosomally complete, we mean we are convinced that the pseudomolecules that we're presenting to the world come from one chromosome and go from telomere to telomere. So the two types of data, genomic data, we're finding most useful are Pacific Biosciences Sequel HiFi data— I'll describe that in a minute— and Hi-C, or chromatin conformation capture, data. So the HiFi, PacBio HiFi data, are a really exciting new kind of data. They're delivered on the PacBio Sequel machine using the standard zero-mode waveguide technology that they developed, have developed. What happens in this case is they build a circular construct by putting dumbbell adapters at the end of the fragments you want to sequence, and they read that construct multiple times, so you get multiple estimates of the genome, of the sequence of that fragment rather, and from the same zero-mode waveguide well in the plate. And because all those reads come from the same DNA fragment, and they also read the forward and the reverse strand. If you put those reads together and call a consensus, that consensus will of course be more accurate than any one read. So the raw read accuracy is relatively low, maybe 10% error rate, but the final accuracy of the consensus, because you've read it so many times, is much higher. Also because the PacBio error rate per read is, uh, primary read is actually relatively random with respect to sequence. There are a couple of caveats to that. The final sequence isn't just a blessing of lots of errors, it's actually a correction of all the errors in the individual ones. So we get data coming out of the PacBio HiFi, which are of the same high quality as the Roy Illumina reads, but they are much longer. So this is the first 360, uh, SMRT cells, single molecule real-time sequencing cells, from the PacBio Sequel 2. In Sanger. These reads are 13, 14 kilobases long on average. Some of them are smaller, and that's because the DNA wasn't so good, or there's some issues with the DNA. Some of them are longer. We're hoping to push this average higher, but that's a 15-kilobase read, which has basically got 1 error in 10,000 bases. So there's maybe 1.5 errors in that 15,000-base read. And that means we can assemble it really well. So the bases are high quality, and they've got high overlap identity, and we can assemble it. We still have some problems. This is a plot of the N50 on the y-axis, so that's the same data as I showed in the last graph. And on the bottom is the yield in gigabases from the cell. Most of the yields are well above 10 gigabases, and we're hoping to get most of them above 20 gigabases in the future. And most of the dots on this graph are insects, and most of the insects are doing perfectly well. They have good high N50s, 13-14 kilobases, and good high yields. However, some don't, and these some that don't are often just because the DNA was very poor, but also we have, um, uh, some cases where we think there's a phylogenetic or a taxonomic bias. So for example, for our mollusks, the purple dots here, about two-thirds of the mollusks were not getting good yields, even though we got relatively good N50s. And the same with the annelids. 4 out of the 5 annelids we've done so far have got less than 10 gigabases yield, even though for at least 2 of those we had very good N50s. So there's still places to go, still improvements to go, And that's what we're working on now, to make this much better. So I mentioned that we wanted to be an open project, and one of the things we're trying to innovate on— this is driven by Shane McCarthy and Richard Durbin in the project— is to make sure that you have access to the data as early as is possible. And so Shane has set up an open portal at tolqc.cog.xangar.ac.uk, where you can go and you can find out what data we have for which species, in our pipeline. So the minute species has gone into the pipeline and some data has been generated, you can find it on this site. And in fact, the graphs I showed in the previous two slides are basically pulled from this site, so it's not private at all. You could do— you could, you could draw the same graphs if you wanted. Okay, so for each species, we will show you what specimens we have. And this is for peacock butterfly. But also what data we have. And importantly for the data, you can also see the quality of the data, including things like, for example, a GenomeScope profile. This we find really useful, and it's something just to comment on. This is a GenomeScope profile where the histogram is of k-mers, so unique sequence words, if you like, in the reads, and looking at the distribution of the frequency of those k-mers. So of course there are a lot of k-mers which are only seen once. These are ones with single errors in them. And then we have two peaks, a peak at about, in this one, about 40-fold coverage, and a peak at about 80-fold coverage. And these peaks correspond to homozygote, the one at 80, and heterozygote, the one at 40. And this shows this organism is quite heterozygous, so it's about 0.7% heterozygosity, much, much greater than a human. If you remember, I said our sequencing error rate was somewhere in the region of 1 in 10 to the 4, so 1 in 10,000. This means that the two alleles are present in this butterfly— we can separate them. So we can actually generate separate assemblies for the two haplotypes of each chromosome. We're not quite there yet. We can generate a chromosomal assembly for one haplotype, and there may be haplotype switching in there, and a good assembly for the second haplotype, type. But we are getting towards a place where we can actually submit not just the haploid estimate, or consensus estimate, of the genome of an organism, but actually the diploid estimate, giving both alleles. The data are very, very good. Okay, the other sort of data we're using is Hi-C, and I've put a couple of slides in to explain what Hi-C is. So this is a method which was originally used to, as a a technology to look at the interaction between pieces of DNA, especially when they were being involved in, for example, long-range interactions at enhancers and so on. And it was realized quite quickly that if you could do this genome-wide, you get a picture of the enhancer or repressor landscape of a whole genome. For us, what it means is that because chromosomes are more likely to interact with them DNA and chromosomes is more likely to act with itself, so within a chromosome than between chromosomes, we can actually use these data to separate our scaffolds that we get out of the PacBio data into chromosomal-sized pieces. So just to illustrate this, this is the human genome. Obviously in metaphase we can see all the chromosomes and we can sort them out, and it's possible to paint those chromosomes by taking chromosome-specific sequences labeling them with fluorescent dyes and hybridizing them to the chromosomes, and in the metaphase plate you can see on the bottom left, you can see each chromosome is effectively colored differently. If you do that in an interphase nucleus, and the picture in the middle at the left is an interphase nucleus colored with the same sort of paint, you can see that the chromosomes aren't smeared out across the nucleus but they're actually localized in the nucleus, and it's possible to label each domain in that nucleus, which happens to be from a fibroblast with the chromosome it corresponds to. This means that if we cross-linked the DNA and proteins inside that nucleus, cut the DNA, diffused away any intervening fragments, then re-ligated the DNA, we'd be more likely to re-ligate within a chromosome than we would be to re-ligate between chromosomes. And that's how Hi-C works for us. We turn scaffolds into chromosomes by doing this cross-linking, restriction enzyme digestion, re-ligation, and then purification of the ligation products and sequencing. That gives us two markers in the genome which can be a kilobase apart or can be megabases apart, and we can use those to join scaffolds together. And this really does turn scaffolds into genomes. A couple of illustrations from work in progress. So these are comparing two different assemblers on the same genome, and the red, deep red color indicates the density of these Hi-C links between or within scaffolds and fragments. And I hope you can imagine that from these two plots here, you can see where there are segments of the genome which haven't yet been joined together, but which Hi-C says should be joined together. So if you look in the right-hand plot, at the top right, there is a bar of bright red color which I think, I hope you'd agree with me, probably belongs with the second, needs to be joined in with the second of the two contigs at the top left of that plot. So we can use this to curate the genome, and that's a really important process. So what we do is we do an assembly, do the best we can, including using the Hi-C data to scaffold everything together. Then we give it to Kirsten Howe's Genome Reference Informatics team, and they turn it into a chromosomally complete genome where all the joins and all the ordering between the scaffolds is actually affirmed and correct. The other thing we're doing is, because we're sampling from the wild and not from a tissue culture dish, or not from an animal that's in an in an insectary or a mass house, and not from a plant in a sterile growth system in a greenhouse, is we're actually sequencing more than one thing. So we know we expect the nuclear genome will be there in the organism we sequence, but we know there will be other things there as well. There's the organelle genomes, the mitochondrion and the plastid. There's also any symbionts. Are present. So these could be mutualist symbionts that are essential for the life of the joint organism. Think about corals and their algae, or aphids and the Buchnera bacteria they live with in order to be able to eat only plant sap. There'll also be parasites. So it might be that the individual we pick from the wild has got a parasite, and we don't want to submit a genome that says, hello, here is the genome of this moth, and in that statement we're actually including a bunch of other genomes which are not entirely relevant. We also want to separate anything that got in there by accident, um, something that was part of the processing, a human skin bacterium for example. And so we want to separate these out and submit it as a group. So here's the nuclear genome, here's the organelle genomes, here are the other genomes. So that's a process that's working now, it's not quite at production level, but for example, here's the buff tip moth. This comes with 5 free genomes: the nuclear genome, represented by the Hi-C plot on the right, the mitochondrial genome, assembled independently by the circle on the left, and it also has 3 different Wolbachia, which is an alpha proteobacterial symbiont, It's a reproductive parasite, so it's not a mutualist symbiont, but it does bias reproduction so that females infected with Wolbachia are more likely to transmit to the next generation than females without. But this one actually has 3 different Wolbachia in it, so we've got 5 genomes out of this one organism. So the last few minutes, how are we doing? I'm going to go through this quite quickly. All the data again are public. Currently we've got about just under 3,000 samples in the Sanger Institute from various places across the UK. These represent about 680 or 700 species. As I said, we're doing about 20-odd extractions and sets of sequencing per week. After the coronavirus shutdown, this is ramped up very, very quickly, and currently we have about 250 or so species in sequencing, and about 70 species that have all the data they need in order to finish off the assembly. This Venn diagram also mentions a data type called 10x. This is single molecule read clouds. This is really useful for correcting any remaining errors in the PacBio, and also in just affirming the correctness of the assembly. Where have these species come from? Well, we seem to have an inordinate fondness for beetles, butterflies, moths, flies, wasps, bees, and bugs. We've actually— because we know we can do insects, and because that's what we had in the freezer at the start of lockdown, we focused been forced to focus on extracting from insects. This has been really useful because it's allowed us to ramp up, but it's also going to be really useful for the insect community and the arthropod genomics community because we're sequencing quite densely from this as well. So about 46 families in total are represented in the sampling that's been done, and nearly 40% of what we sequenced so far has been Lepidoptera, plus a bunch of other of things. And just to focus on the Lepidoptera, this is a slide from Marcello Uliano, who's been assembling these. Currently we're doing an assembly, an assembly race, I guess, which is to sequence as many, to assemble as many Lepidoptera as possible in as short a time as possible. And Marcello has been assembling this. So in the green dots are Lepidoptera genomes from LEPBase, which essentially is the state of the art prior to our starting this project. So 2 of those green dots are up near the pink dotted line. That pink dotted line is a line of the contiguity we'd expect in the genome if the genome was chromosomal. So we had this idea of chromosomal pseudomolecules, so everything was in chromosomes. So only 2 out of the 40-odd LEPBase genomes made it to this chromosomal level of contiguity. What Marcella and Xenia and Shane have been able to do with the data we've generated thus far on these, these 48 Lepidoptera is, with the contigs— this is just the raw contigs that come out of the PacBio HiFi data— generate many assemblies that are looking to be chromosomal without any scaffolding. So where the contigs N50 is at what we'd expect of chromosomes. And then once the scaffolding happens, so the Hi-C scaffolding, these genomes are all coming in exactly where we'd expect them. They do look to be chromosomal. There is one dot way above the line on the left-hand side of this graph, the light blue dot. That's a, a species which has many fewer chromosomes than we— than the norm. So most butterflies have 31, this has less, this species has fewer, and therefore they can't again 50 size, and scaffold at 50 size, much bigger. Okay, so we're getting there. We can deliver chromosomal genomes at scale, and congruent with what we expect from their karyotype and biology. And what we can do with those genomes, for example, here's another butterfly, the ringlet, is suddenly start— is be able to start looking at patterns and process in genome evolution. I said that Butterflies and moths usually have— well, the standard number of chromosomes is thought to be 31. The ringlet has 2 fewer, so it has 29 chromosomes. And we can clearly see in the assembly that 2 of the chromosomes in the ringlet are fusions compared to the genome of the moth, Tutela xyllostella. And so we can immediately map where these fusions happened. Those of us with a sharp eye will see that on chromosome 24, the last two markers on chromosome 24 appear to come from the wrong chromosome, and this looks like a small breakage and fusion event, a translocation event, which just affects the top end of chromosome 24. Okay, so we're starting to get ideas of patterns of chromosome evolution in the butterflies. Again, we're releasing these data openly. These genomes are not in the public domain yet, but these preliminary assemblies that Xenia, Marcella, and Shane are producing, and the curated assemblies that Kirsten Howe's Grit Group are producing, are available in our open data release portal. This is on GitHub, or you can get it through darwentreeoflife.org. And again, this data gives you— this portal gives you access to the raw data, the sequencing data, but also to the preliminary assembly. So this is not released yet. Assemblies. They will change between this point and being released, but that allows them to be accessed openly. Finally, what we want to do is publish these genomes rapidly and openly. So traditionally in a genome project, one would sequence the genome and then spend a year or more analyzing it, and finally at the end of that process produce a paper which described the genome, and the genome would be publicly available. For the genomes we're producing, we want it to be faster than that. Yes, we will be publishing overarching papers,, but we want each genome to be available when we finish it. So this is one from an earlier project, the 25 Genomes Project, which we published openly as a genome note. This is for the red squirrel. This genome note does two things: it acknowledges everybody involved, so everybody who collected— the collector, the person from the Lancashire Red Squirrel Trust who collected the red squirrel that was used, right through to the person who did the data submission and analysis. That's really important because that gives them credit, and for the Lancashire Red Squirrel Trust, they are now the people who contributed to the sequencing of this red squirrel genome. Since the red squirrel genome— red squirrel is found, and the genome it contains is found from Ireland all the way across to the far east of Asia— they've contributed to a project which can be used by people across half of the world. Okay, so this is— we want this to be open and rapid. And so we're working on this genome note publication mechanism. So as I said, I've just gone through the Darwin Tree of Life project, because that's the one we had going for nearly a full year. And we're also working on aquatic symbiosis genomics. Those samples are just starting to come in. And we've also, over the last 3 or 4 years, been sequencing for the Vertebrate Genomes Project. And those genomes are also publicly available. And we hope to add to this set of projects we're working on both taxonomic-based ones, vertebrates or other taxa, also regional ones collaborating across Europe and across the rest of the world, and all doing this in an open and ethical way so the benefits that we believe should accrue from these genomes are actually available. So we're not keeping them behind paywalls, we're not keeping them with embargoes, we're releasing them openly to the world for everybody to use. Okay, so I gave a thanks, a people slide earlier in, in the talk. Those are people that do the work. I get to give the talks. I'm very grateful for this opportunity to talk and hope you found it interesting and that you find the genomes that we produce useful and transformative. Thank you.