 So, willkommen zusammen. Heute Abend gibt es den Talk von Andrea über den Coronavirus Structural Taskforce. Und ich bin Messai, der Herralt für die Session. Wir haben ein Signal-Agentual hier. Das heißt, der Signal-Agentual wird die Fragen sammeln, die ihr in den Chat gestellt. Und dann gehen wir am Ende vom Vortrag über diese Fragen. So viel zum Ablauf. Der Vortrag wird aufgezeichnet und ist dann dementsprechend nachträglich verfügbar auf media.cc.de, irgendwann in den nächsten Tagen oder Wochen. Und damit würde ich mich freuen, Andrea, du als Nachwuchskruppenleiterin an der Uni Hamburg. Du hast dich in den letzten zwei Jahren mit dem Coronavirus beschäftigt und daraus wunderbare Visualisierung gemacht. Wie lief das denn ab? Und vor allem, wie sieht Corona eigentlich aus? Ja, vielen Dank. Erst mal Danke für die Einladung. Und genau darum geht es eigentlich in dem Talk jetzt. Um die, was wir, die Coronavirus-Structural Task Force nennen. Und ich werde die Präsentation in Englisch geben, sodass International Listeners können auch auch inlisten. Aber du kannst dir Fragen in, well, any language, anyone here speaks. Ich verstehe German, Englisch und Japanese. Und ich will start with a quote by Marie Curie. I know the room Mary is not named after Marie Curie. But she said something that is very true in this pandemic, which is nothing in life is to be feared. Only to be understood. Now is the time to understand more, so that we may fear less. And indeed this holds true for the Coronavirus more than anything. Because as you all know, you cannot see the virus. You can only see it indirectly visualized by science. Or you can see the measures against it. Or you can see ill people. But the virus itself is invisible. And I'm going to start this talk with questions. There will be many questions. And the first one is what does the Coronavirus look like? Now you may think you know. But the reality of it is that even German news has no idea. And I know that ZDF is now using a different picture, which looks more similar to what I'm going to show, but it's very wrong as well. This picture is what most people think the virus looks like. And I also brought you like two top modern models of the virus. One can even make sounds. Any spiky ball these days really passes as a Coronavirus. Because no one seems to know what the thing really looks like. Only it's like ground and it has spikes. That's the only thing that all the models have in common. But some things look like you can just, you know, like they are little shrek ears type things or have tentacles. No one really knows. So how do we know as scientists? And can viruses be seen? If we imagine, so this is an electron microscopic picture of a human hair. It's 0.1 millimeters is the length of this line. So the hair is a little bit less. The little red dot, which you may or may not be able to see inside that circle is the size of the Coronavirus. Now if we zoom into the picture of the hair, we can see a hope, a little red dot here. And that's the Coronavirus to measure. So it is 150 nanometers or 0.0001.5 millimeters large. That is tiny, even by scientists' standards. However, even smaller than the virus with 150 nanometers is a single atom, which is represented here again by a dot, which is barely visible, and is 0.1 nanometers in diameter, or one angstrom. Atoms are tiny even compared to the virus. So the virus is composed, being matter of very, very many atoms. How can we visualize something this small? Can we see it with a light microscope? And what color would the virus be? This is to scale, yellow light. This is the 90 nanometer wavelength, meaning from this point to this it is 600 nanometers. So the wavelength of visible light, which ranges from 400 to 780 nanometers, is actually longer than the virus is wide. So there is no chance whatsoever to ever observe a single virus with light. nicht möglich. Wir brauchen etwas, das ein kleiner Waiflinge hat. Und da sind zwei Dinge, die wir benutzen, X-Rays, die 0,1 Nanometer Waiflinge haben. Also sind sie sehr, sehr klein. Sie sind sehr lightig. Sie sind auch Photon. Wir nennen sie X-ray-Light. Rücken, Strahlung, Rückenlicht. Aber sie haben so viel Energie, dass ihre Waiflinge klein sind. Alle Dinge sind Elektronik, die sogar kleiner Waiflinge haben, wenn man sich nicht als Partikel, sondern als Waiflinge schafft. Wir können Elektronik und X-Rays nutzen, um die Virus zu observerieren. Und für diese haben wir mehrere Möglichkeiten. Eines ist große Partikel-Acelleratoren, wie die, die ich in Hamburg arbeite, die einen sehr intensiven X-Rays produzieren. Einer ist ein Elektronik-Mikroskop. Hier ist ein Modell eines Elektronik-Mikroskops. Wenn du sie nimmst, brauchst du ein Scientist. Und dann schaffst du einen Elektron-Mikroskop, ein Elektron-Mikroskop. Das ist der offizielle Scientistik-Mikroskop. Es ist ein Elektron-Mikroskop. Mit einem Elektron-Mikroskop. Elektron-Mikroskop, Elektron-Mikroskop. Du schaffst deine Elektron-Mikroskop durch die Magnetik-Mikroskop. Die Magnetik-Mikroskop sind negativ geplant. Also kann ein Magnetik-Mikroskop als ein Lenssystem. Auf ein Sample, zum Beispiel das Virus. Dann hast du einen Detektor. Was sehen wir auf diesem Detektor? Wir schießen Virus mit Elektro-Mikroskop und rekordieren, wie viele Elektro-Mikroskop in den Sample ist. Was wir sehen, sieht so aus. Natürlich ist es black und white, weil Elektro-Mikroskop keine Kohle ist, und was du siehst, ist ein schwarzer Schaddach und dann ein bisschen breiter Spots. Das ist quasi ein Corona, während der Sonne klingt. Das ist warum der Coronavirus ist, den Coronavirus beinhaltet, weil es auf Elektron-Mikroskop spielte. Diese Bilder kommen keine Farben, aber Wissenschaftler wollen sie colorieren. Das ist ein gaggres Virus und das ist die Rückw эп. Wenn wir sie colorieren, könnte es so werden. picture released very early in the pandemic by the national institutes as one of the first pictures of the new coronavirus. We can also do scanning electron microscopy which is a similar measure where you coat the entire surface and then you get a pretty three dimensional picture. What you can see here are lung cells, the long carpet like the like hairy structure here that's lung cells. There are single type two alveolar epithelial cells so they're like a little like like their job is to get rid of stuff the lungs don't want. They're like a carpet they can like move and they get rid of stuff for you. However these cells have a problem they're infected with coronavirus. You can see some slime or mucous here and you can see the viruses here. Because of the coating they look a little bit like cauliflower so that's nice but it doesn't give us the full picture but so much for people who say we cannot isolate the virus we can actually even like make it visible. So we can make this invisible enemy visible. It is just a question of having the right equipment and a good sample of the virus and many hours of work. The virus therefore exists and can be made visible using electron microscopes or for example also particle SLRH but I'm not gonna go into detail here we have not enough time tonight. I'm only going to talk about electron microscopes here. So what is the virus made of? This is the virus. We're going to talk about this picture later in the talk when we talk about to model a little bit more but here is one spike I think you've all heard in news from spike proteins which cover the surface of the virus or better the viral. If we draw this schematically as Thomas Schbletschduzer did for us he's an illustrator burst in Berlin it looks like this and then we take only the head of the spike which is the region of the spike we know most about and then comes an animation that I did so it's not quite as pretty. If we zoom in we see the surface and beloaded surface we see things represented as a ribbon however we can show this differently we can show this as individual atoms connected to each other. The problem with this display is it's really hard to find anything. It is super difficult to get an overview with this picture so we prefer to show these complicated and big molecules made up of atoms as surfaces and ribbons. And this ribbon diagram by the way has also been found by a great female scientist Jane Richardson several decades ago which was quite revolutionary for my field. So the virus is made up of atoms and molecules and their structures can be found out by NMR x-ray crystallography and electron microscopy. For now this is all I'm going to tell you we're now going to dive very deeply into the biology of the virus and what the structures tell us and then in the end I'm going to tell you a little bit more exactly how we actually get from the measurement to the model which holds some pitfalls and problems for us and is an area in which I do usually my research. But first I'm going to talk about the model you all know this picture from the CDC right and by the way this thing for scientists this thing is not a virus it's a variant it's only the transport form of the virus. The virus does a few more things that are not like contained in this little shell that's just a transport form for its RNA to get into a host cell we call this a variant. But most people even scientists can also refer to it as the virus. So this is the CDC model. Right? That's the picture that went all through the past around the world. That's like the picture of this pandemic. And it was made by the CDC very early on by two scientific illustrators there. However, already then it had some problems. They made it in quite a rush and it has some errors. So we decided to make a new picture which looks like this. And if you compare the two pictures, here are the differences. The head of the spike in this illustration sits directly into the surface while in reality it is sitting sitting on a long, very bendy, grope-like structure that havers it. So the virus, the head of the spike which binds to the host cell is quite flexible. The surface is not quite as coarse as shown here. It's smaller. The virus is actually relatively large for a virus. And it's got other proteins swimming in its surface. If you look exactly, you will see the virus is also not exactly round. Now we thought this is not enough. It's nice to have a picture, but wouldn't it be nice if we could actually touch it? So we made a 3D printable model. For those interested, you can find also all the information on our homepage. I'm going to show you the 3D model. Let's see. So this is the virus model. As you can see, the virus is not exactly round because it's outside is very soft. It's like a soap bubble. It can change its shape quite drastically. It's wobbly. And the spikes are actually stochastically distributed and they are not like regularly arranged and they are swimming in the skin. And there are other proteins in the surface as well, which you can perhaps see. Whether they really form this little flower shapes, we don't know. And the virus is huge. So this, on the same scale, one to one million as a rhino virus for the common cold. The coronavirus is huge. 20,000 base pairs RNA makes it one of the largest virus genomes we know. And the virus is therefore soft on the outside. While rhino virus has a very hard and rigid shell that is always composed the same. And we made this model in the hopes that like other scientists and perhaps schools would like to like print it at home and actually like get something tangible. And it turned out they did. So we get quite a few requests from people from childcare facilities and from schools and from other scientists. And even our administration like to have them printed. One even proposed we may have them as Christmas ornaments, but I found that a little bit like tasteless. We didn't do it. And like just before I left Hamburg, we got a new model. This model is now already a year old. And in 2021 science made quite a lot of progress. So we now know there are fewer spikes. And the virus all in all is a little bit smaller. So it's not quite like this, but perhaps less. So it's not 15 centimeters in diameter. The model is 12. And it is still like potato shaped. It's not round because now was important for me to show. And I would have like to show you this model like in front of the camera tonight. But it went into the museum. It was the first one we had. And we brought it to the opening of this exhibition in Hamburg, pandemie rückblickende Gegenwart. So you can now see it in the museum and will be back in my office when they have assembled theirs. And this also holds true. I'm going to talk about the task force in the second half of this talk. But one thing that is really true and was true for this project as well is the task force is typically more interested in new communication project than in all the pile of stuff we need to finish. You may know this from home. Right. So this is our model. Now let's dive deeper into the thing. Because so far, we have only talked about the virion and only about its outside. So the virion has the spike proteins on the outside to other proteins M&E protein. It has a double membrane Hau, which is very thin. A nuclear capsid, which is wrapping the RNA. The RNA is actually containing the genes for everything that the virus needs in order to like take possession of the host cell. So the RNA is the important bit, the virion is transporting and the nuclear capsid and everything else kind of packs it and make sure it gets into the host cell. And I'm going to show you quickly a video because I think this is so nice to understand. And it's the answer to the question, why does soap help against Corona virus? Because very many like other viruses are relatively hard to wash off, but not Corona virus because it's so large. It has only a double membrane shell. And this is a very nice video from the protein Data Bank from our colleagues there. So this is the virus double membrane. It has lipid molecules. You can see they are hydrophobic on the inside and hydrophilic on the outside. So they love water on the outside, but not the inside. Green molecules are soap. So also has a hydrophobic tail and a hydrophilic head. So unlike the water which stays outside, the soap just gets into the membrane and kind of like goes in between the lipids. And then they make holes so water molecules can get inside the virus. They can even assemble in around bits of the membrane and get it out of the virus hull or around a spike and remove the spike, which is hydrophobic at the stalk out of the virus. This leads to a totally decomposition of the membrane and therefore it can be completely dissolved by soap. As soon as the nuclear capsid and the RNA are exposed, the virus is no longer infectious. It needs its spike in order to infect. So you don't need to disinfect your hands. You can just wash them with soap, which I find is so lucky in this pandemic because it would be really ugly if we would need to disinfect everything all the time, but we can actually just use soapy water. Although, let's be honest, I like to use disinfectant every here and then. It gives me just a feeling of more safety. It's kind of like a ritual to protect me. I suspect several of you are the same. So much for the virion. That's like the outer shell. That's like the transport form. But there is more. Much more. This is the coronavirus life cycle. Or I should be more correct. It should be called infection cycle because technically speaking, viruses are not alive. They need a host cell to reproduce. So, I'm going to come back to this picture. We're going to take this apart bit by bit. First of all, there is entry. Entry into. Oh yeah. Let's quickly go back. The thing at the bottom here, the big thing here, that's the host cell. And the little one is the virus, right? We're clear on that. I hope we're clear, right? And this is the outside. So, this is like your lung outside. That's your lung cell inside. Or hopefully not your lung cell. Right. Here's the virus. And this is the spike. The spike is a vaccine target. As you know, it's what's encoded RNA vaccines and it's also contained in all the like vector vaccines we see. And I brought you another model for that, which is also pictured here. So, this is one to ten million scale model of the spike. And this is an antibody. Now, if you are vaccinated either by RNA or by vector vaccine, your body either produces or is injected with spikes, which are usually on something to carry it, a host cell or a cell of your body if it's an RNA vaccine or a vector. Your body needs a few days to recognize this thing. So, once it has, it will build antibodies that like perfectly fit onto that. They can recognize the spike very, very exactly. They have a specific binding site, which is much more rigid than anything else the antibody can bind to. Then, this gets decomposed, because the vaccine is not viable. What remains in your body is the information for these antibodies. So, now, if you get infected with Corona, the immune system antibody recognizes the spike it has previously seen in the vaccine. And that is how the vaccine actually works. So, having this protects you. One of the biggest problems with COVID-19 infections is that the immune system responds too late and then too much. So, having these makes much more certain that you will not get severe COVID, which is, I think, very nice. And what we also did together with the animation lab in Utah is not only make this life cycle or infection cycle, we also made an animation, the scientifically most accurate animation available on how the virus actually binds onto the whole cell. There's a lot we don't know, but everything we do know, we have shown here. So, here is the virus. You notice the spike protein already. There's nucleocapsid inside. There's RNA, which encodes for the rest. We're going to go into the rest after this. And then at the bottom here is the host cell. So, lung cells actually have ACE2 receptors shown in purple. And the spike protein recognizes those specifically. Meaning, the atoms, like a puzzle piece, fit exactly onto the purple receptors on the lung cells. The spikes bind there. And then something else happens. Another enzyme also being in the membrane of the host cell cuts the spike. So, it's not like, the name spike suggests it would shoot something into, but that's not the case. It gets cut. And then comes the bit where we are a little bit unclear how this happens. So, what we know is, after it's cut, it somehow ends being tethered into the host cell. And we don't know the mechanism of this. So, we decided to illustrate it here with, like, a refolding process. And then it's energetically unstable. So, in order to become more stable, the whole thing clamps and folds together. Thereby dragging the virus and the host cell membrane together. The two membranes, two biolipidic membranes fuse. The virus material is inserted into the cell. The RNA is now inside the host cell. And this is how infection happens. So, we felt it was particularly important to show this to people. We have made this animation Creative Commons. But unfortunately only American television caught up on it. So, we're really hoping that one of the German, like, documentaries will show this. Because we think it's really nice to see this process, like, as accurately as we can depict it. So, here's the spike. That's the role of the spike. Now, the nuclear capsid and RNA have entered into the host cell. What happens now is that the nuclear capsid dissolves. How that exactly happens, we don't know, but it dissolves. And the RNA gets immediately translated into protein. So, the genome gets read inside the cell. Because the cell believes its RNA, that comes from the host cell. And it starts building the proteins encoded in it. Proteins are again molecules. And actually, it makes one very, very long protein chain. Really long. Which is called the polyprotein. Which then gets cleaved. So, the scissors in this case, the thing that cleaves all the long protein chain into individual molecules, that then actually can work, is called the main protease. Because it's cutting protein, it's called the protease. And these are the little triangle-shaped things here. Only when that happens, do these bits become functional. So, if it doesn't happen, if we can like hinder this like cutting of the long polypeptide chain, we have an efficient drug against corona. Which is why main protease is a major drug target. And what you can see here in red is a drug actually bound to main protease. So, that's what it looks like. We are looking specifically for inhibitors, which will stop this molecule from function. So, you can imagine this like a screwdriver, that we put at the right point in a machine where it fits and it stops the entire machinery. That's what we want. And that's like called structure-based drug design. When you actually know the structure of the molecule and then you find a molecule, a small molecule, a drug molecule, that like specifically stops whatever that protein molecule is doing. It's also how many antibiotics work. Or for example, if you've been up long yesterday, aspirin. Then, from here, where we have the polyprotein thing, something happens inside the cell. The cell is making some kind of foam, which is connected to the endoplasmatic reticulum. That's this one. And inside it, these like enzymes that have now been cut and are functional, start like a copy machine to make more and more copies of the RNA genome of the virus. The whole cell kind of like foams up and makes more RNA. And it does so by a copy machine, which is called RNA polymerase, because RNA is a polymer, and a polymerase is an enzyme that's making polymers. And an RNA polymerase is an enzyme that makes more RNA. So one way to block that process of making more RNA, which was similarly, you know, stop the infection cycle, because more, if it can't produce more RNA, it's got nothing to put into new viruses, is to use Remdesivir. Also we thought for a long time. So this is Remdesivir. Remdesivir looks to the whole cell and to RNA polymerase in particular, like a nucleotide, like a building block for RNA. So it basically thinks, oh, that's new paper I can copy on. When in reality, it's kind of explosive and like blocks the entire polymerase. So you can imagine this like a trion horse, Remdesivir is a trion horse, and a dependent RNA polymerase says, sure looks like RNA to me, and just builds it into the strand. So we'll have a look at this molecule. This is, by the way, I should possibly go back and explain this for a moment. This is what we get out of our measurements in cryoelectron microscopy. So this is the so-called reconstruction density, and that's what we build the molecule in. So that's like what the measurement gave us. Everything else we have to do by hand. So here is the density, and this is what the researcher built into it. The template strand, the old one, which is to be copied, is green. The new one is orange. Remdesivir is purple and connected to the end, although the program doesn't display it as such. What you can also see is three magnesium ions, that's their own thing, and a diphosphate. So there are more molecules that the researchers modeled in there. We are, around it is the protein, so I've just quickly depicted it without, so you can see what's happening, because it's all very crowded and difficult to see. Around it is the protein, and the job of the protein is to copy the RNA and to attach one after the other another building block to the bottom of the orange chain. It's gotten a Remdesivir, and it tried to put it there, and it successfully did. And this structure was taking us to prove that Remdesivir will stop corona, but if you look at the density, you can see that the three magnesium ions in the diphosphate have no density. They are not covered by this gray cloud, right? So we think they were never really there, and the Remdesivir itself is also not having so much density. So it turns out that the structure is not quite saying what the researchers did, because the density doesn't match up with the structure they built. And as we later found out in clinical trials, is that Remdesivir in fact doesn't really do what people hoped, which among other things has to do with the fact that RNA polymerase from coronavirus is able to prove, read, go like, oh, there's a Remdesivir, then go back free nucleotides, take, rip out the Remdesivir, throw it away, and get the proper nucleotide to build in. So we're hoping that Molnovapir will be better, which by the way uses a similar mechanism. It's called a nucleotide. Yeah, it's similar to a nucleotide. Here's another error we found in the bottom of the structure. I'm only going to show you, because I think the software by Tristan Kroll ist so cool. It does a real-time molecular dynamic simulation while you're dragging around your structure. We found that there is an error in the way the whole molecule was arranged. If there's any specialist, there was a nine amino acid out of registration. Okay, I'm going to stop like kicking out. There was just an error. Let me show you how he corrects it. So first he marks up the wrong region, and then he releases it, and it goes where it's out to go. I wish my life would always be like this. You sit with three days glasses in front of your computer. You need hours to do this by hand, but his software just does it. Sorry, it's just, if you've spent months doing this, you're totally excited by this wobbly molecule. I just wanted to show you, because I think it's so cool. All right, back to more general content. We have now the RNA. Let's imagine we didn't get any good drugs so far. So the infection cycle is still ongoing. The RNA now is exported from the endoplasmatic reticulum. By the time we made this, we didn't know how, but Hamburg researchers and Dutch researchers have now found out how this actually works. There's a pore here, and the pore gets the RNA out. The RNA then gets packed into new nucleocapsids. The nucleocapsids were also coded by the genome of the virus, then gets wrapped into new double membrane, which is host membrane. Just it has now spikes, which also were produced from the genome. And then, over to Golgi, the Golgi apparatus is somehow involved, it gets exported. Of course, for everyone that infected a cell, there will be thousands that are produced. And that's the entirety of the SARS-CoV-2 viral life cycle. I know this was a little bit much, but I think this is cool and exciting, and just about what I hope the public can understand about this. It hopefully also tells you why molecular structures are important. So they let us understand how the virus works, how host cells are infected. They can help us to find drug targets and to do structure-based drug design, where we find drugs that specifically block these big molecular machines from doing their work. They also help us to understand the structures of vaccines and antibodies. And they also let us understand changes due to mutations. I haven't got an example because of the time, but when we get a mutation with this structure, I can kind of tell you that bit is going to change or that bit is going to change only by knowing the new genome. I can already make a prediction about what the mutation is going to change in the functionality, and it's really important. So in my group, we have some theories what Omicron actually does. We haven't published them. We haven't even tweeted about them yet because we're still waiting for research results. But it's important to understand these molecular structures. However, it's not very easy to get them. So what are the problems? When we do a measurement, we get density. In this case, the density is blue. It's from the spike head that I've shown in the beginning. So the top bit, you may recognize this, looks a little bit like the blue density here. This is the result from our research. In this case, I have built almost all of the molecule already. So I'm going to show it. This is like the software we're actually using. At tenfold the speed I'm usually using and I would usually be sitting there with 3D glasses. So here's the density. I sit there with my 3D glasses and this bit hasn't been built. So you can now see me by hand at one bit of molecule after the other, trying to get them where they ought to be. And you can quite well see that the computer is not able to do it all automatically. So I have to help it a little bit. And as I said, that's about ten times the speed. It's even got the warning from the bin program not reacting. All of this software is also not commercial. This has been developed by other scientists. So the usability is like so-so. Coot is an amazing program. But if you want some new functionality, you better program it yourself because perhaps no one else is going to do it. And we have actually contributed with a plugin or two to Coot. Yeah, you can see it's not always easy. So I try and go like, yeah, now it's sitting nicely in the density and here it's fairly easy. My students love doing this. It's like for them, it's like computer gaming. They do it for three months straight. If you don't get them off their chair, go right up your thesis, they'll do it forever. And secretly, I'm jealous because I also like doing this. I don't know if you can follow, but this is like playing a game. By the way, if you're interested in playing this game, we are also having practical places and this stuff. Right, so building, building, building. Going like, oh, there's another alanine. I need a proline. So I mutated it and I go like, yeah. Okay, now it's all nicely sitting. So that was easy. But what do I do here? So I've built something here, but is it correct? The density does not really tell me what's going on here. I'm going to show it just to you in slower again so you understand the problem. In this then part of the density, I can really not tell only from the density what's going on. I know approximately what the molecule must look like due to the sequence. So I've got some information. I know which atom is connected to which, but how it all three dimensionally fits in here, even if I know which lines have to go in there, is super hard. So it would be very easy for me to make an error here because the measurement data don't tell me enough about what the model actually should look like. And several interpretations, several models would all be possible. So this is kind of difficult. I'm just seeing a question there that I may want to answer right away. Within Qt, is it possible to verify if you have chosen created the right building blocks? You know the building blocks because of the genome. So the gene tells you the order of amino acids you want to put after each other. So if you started at the right point, the rest will also be like the correct atoms and the correct connection. But how it like three dimensionally folds, you don't know, you have to make that on the density. So it is possible to do it. You have to write buildings blocks at hand, usually. If there, however, if there is like, you know, if a magnesium ion, for example, is sitting there, the magnesium ion was not in your genomic information, you just need to know what you're doing. To recognize this is a magnesium ion or this is a diphosphate or something like this. Right. Going to answer the rest of the questions later. Molecular models need to be built by hand. This can lead to errors. There are a few automatic algorithms that work under favorable circumstances, but most of the stuff still has to happen by hand. As of today, and I get my postdoc from like holidays for 15 minutes today and he gave me a new number. So we've got 1,909 molecular structures today. New structures come out every Wednesday. There are 1,334 from X-ray crystallography. That's the thing with the particular accelerator in Hamburg. They have been measured at synchrotrons all over the world. Not only Hamburg, where Biontech structures have been measured, but also at the ESRF. There have been large screens at Diamond in Japan, in China, at Sesame, at Soleil in America. Synchrotrons all over the world contributed to this. 35 from nuclear magnetic resonance, which is a little bit of a niche method for this type of study, and 566 from electron microscopy. So 1,909 molecular structures of different states of 17 macromolecules, 17 proteins from a total of 28. So Coronavirus in total has 28 different genes for proteins. There are 28 proteins, and we only structurally know 17 of them, and then we have like different versions of them. Different pH, different temperatures, Spike-Head, Bit of Spike-Head, Spike-Head with antibody, things like that. So 1,900 Datasets in total. Errors in structures, as we've just seen, can happen because fit between the data and the model is bad, because complete automation is not yet possible. Models are built manually. Expertise in many different areas is needed. You need to be good with software. You need to have done all the lab work. The measurement needs to be right. Processing needs to be nice. You need to know your statistical validation. You need to know your chemistry. You need to know your biology. So it's really not easy, and you need to know your 3D goggles, unless you get sick from them, in which case you can't use them. One of the major aspects is software, the methods you're using. And this is where we come in. Small structural errors can lead to big-structured problems downstream. Imagine the bit with the Remdesivir. The diphosphate there, the fact that there was something bound into that structure that was not really there, but the model had those additional free magnesium ions, down the line, as I know from insiders, led to waste of hundreds of thousands of dollars and many hours of work time in drug discovery, because they kind of fed this model in order to find a drug that then ultimately never bound because the magnesium ion wasn't actually there. So if we make small structures, that builds up hugely for the downstream applications, where they need these molecular structures. Errors are common, and now we add to this an ongoing pandemic. And the scientists are there to rescue today. Lockdown happens. You're sitting at home. There are no colleagues to help you. Your child is homeschooling. The dog wants to be fed. Your grandma calls because she's worried. And you've got to solve the spike structure on which lives will depend as fast as possible. Normally we take a year to five to solve a structure in the pandemic. You only got three months. Course errors are gonna happen. So that's, well, just a matter of fact. We've got to arrange ourselves with it. It's not the fault of individuals. It's how the whole thing works. It's such a complex process. Errors are going to happen. Now in normal life, my team and I are methods, developers and structural biology. We are the people who give others the experimental techniques and the software to solve their structures as best as we and they can. We're not usually in the stage light. We're usually, you know, for every Nobel Prize, there have been like dozens of Nobel Prize in structural biology. Has been methods, developers in the background who developed the methods that made it possible to see, you know, the structure of the DNA double helix or structure of the ribosome or the structure of the influenza virus. It's just that normally we're just enablers. However, here was a pandemic and very many structures that had errors. So we did what we needed to do. We came together as a relatively large team under my leadership. We are today 23 people. We peaked at 27 last year from nine countries to check and improve the molecular structures of SARS-CoV-1 and SARS-CoV-2. So we are methods developers. Most of us are methods developers. We are specialists in solving structures. We evaluate all the published structures in the protein data bank or PDB from SARS and SARS-CoV-2. We reprocess all of them and we remodel them. Although not all are looked at manually because that would be just too much. We also do scientific dissemination, putting these structures into context for people who want to start doing molecular research on coronavirus and we do public outreach. I'll give you a quick insight into our pipeline because I thought the software bit might be interesting for you. So every Wednesday come new entries of molecules in the protein data bank which is by the way the worldwide protein data bank is an open resource. Everyone in the world can download the new structures and all the journals require people who make new structures to put them there which is nice. I'm really privileged to be in a field where the data are public. We compare the new structures with the NCBI proteomes. So the genes from coronavirus to find the structures that belong to coronavirus. Put everything into an SQL meta database. Then we calculate how different the structures are from the ones we already know. We look whether all measurement data are available so we have a big problem with not all measurement data being published. I hope we are going to be like astronomy one day where everything gets published. I am sitting in some German committees to that end among others DECOM which is a very new hub. And then we run a number of specialized programs which all do validation and put the results on GitHub immediately. So on Thursday at the latest researchers can find our validation, remodeling and the quality indication for the structure. Everything that can be done automatically online. We then for some structures manually rebuild them. So we actively look whether there are problems. That was for example the case with the Remdesivir structure. We tried to do this for those structures that we think drug designers will use the most or that are really important. And when we find errors we contact the original authors first and tell them we found an error. Here is the corrected structure. Use our data you don't need to cite us. Just correct your structure in the database please. This means we won't get credit but it meant that at the beginning of the pandemic people were adjusting their structures already when the preprint was out. So there would not be problems downstream. And I have often been asked I would like do like this again. That was really like why people accepted our corrections because they didn't need to give anyone credit. They could just like change them. And the databases also online. So everyone from the Philippines to the US can just use them. Whether it's like a commercial person or a taxpayer or a private person, research institution, a foundation everyone can use our data. They're just online there for everyone and we only ask them to give us credit in the form of a reference citation. We disseminate the data via GitHub, via Twitter, 3D Bionodes which is a three-dimensional database is linking directly into our database. We contact the offers. We also have entries in Proteopedia, MOL SSI which is a virtual bioinformatics institute links directly into our database. So they're always like up to date showing what we are doing. We have a home page and we do reviews. There's a lot of downstream users. The biggest ones are the EU Jedi Grand Challenge Folding at Home which peaked out in July last year I think at 2.4 exaflops computing Powerful Molecular Simulations and they used in the majority our models to start from and also very big as IBM Open Pandemics. But there are a number of others plus many individual labs so we found a great new many friends. Here is our home page that's where you can find it there's also an English version you can find blog posts for the public and for scientists and in the end I would like to talk for like five minutes about daily life in virtual mobility. So my team is over 20 people from nine we recover nine time zones right we're nine hours time shift apart and we had several lockdowns so actually the majority of people in the task force about half of them don't work for me. They are volunteering their researchers elsewhere that volunteer to be a part of this effort and there are many people in the task force who've never liked personally met so how do you make group coherence if you are working for 20 months or 22 months by now in an environment like this right we founded ourselves in March 2020 as a chat group called the coronavirus structural task force which was a joke back then it didn't remain a joke but that's how life plays we have every day Zoom meetings at 10 o'clock one time per week in the afternoon so the people from Oregon can join us we do a lot of like media outreach in international and german media we've been like on nano and terra x and planet e we've been in breakfast television that was a particularly interesting experience my email got leaked to querdenka and i got a few very interesting email exchanges i also must say i never got insulted or threatened by anyone so i just discussed with them and it worked out and i'm happy because it like i understood like how what the theories are and that was very interesting the media also like to write about my hair my eyes blah blah blah all these things i just want to note that streak is about my age and drosten is only nine years older than me so really okay whatever most importantly they are talking about our research we also did a lot of social media i got a twitter account everyone else did as well you can watch us work on twitch if you're interested we found out people find it soothing to see us modeling these structures i was requested to make stickers for the team as soon as we got a grant and we got funded by the federal ministry for research and education last in 2020 we have a youtube channel where you can for example see the entry animation and the students brought a cactus which is called corona cactus i know it looked like we had fun and we did i can tell you being car you know being at home having to care for your children having people dying having an ongoing pandemic knowing that IBM open pandemics is going to spend another million based on your research and also that ctf wants to talk to you in the bellina abendblatt this is so stressful think about all the responsibility we had and it was really terrible for us so we needed to cheer up every here and then the whole group kind of like grew together and we all became friends this was for us it was very uncommon as researchers typically our behavior with each of us much more formal than the behavior in this group was but these were like exceptional times and we wanted to fight the pandemic and inform the public that's what we were there for and was not so much about personal gain and that was nice we have a group chat you know i mean i'm talking to right audience right we have a group chat you know what that looks like let me tell you typically professors don't communicate this we have a virtual space we have a virtual space we hope to change to work adventures soon um we also play games in the evening occasionally with the group so we do some team building efforts when people can meet in person they usually do and sometimes they go and travel and like beat each other but this has been very limited and we did grow together as a network that will be there after the pandemic so we are 25 people all over the planet that did this together and even if we wouldn't have made any difference against the virus i would still be happy to have done it for the friendships i made however we did make a dent we don't quite know how big it is because our science was open science and the results could be gotten by everyone without reference but we know a few things wouldn't have happened without us and i'm deeply grateful for having had a purpose during this pandemic in the end i have a little bit of a more serious topic my work contract is gonna run out in may i think it's gonna be prolonged when it will i'll be signing my 14th work contract since 2008 14th like i had 13 work contracts already out of all my task force members there is only one holding a permanent position and two which are retired everyone else including six people whose contract is gonna run out next year are on temporary contracts and it's not students and students are extra that the students are on time limited contracts is okay but germany's got a problem 84% of academics in german universities are on time limited contracts so are we only one task force member has a permanent position in science and that's not me and this is not so much on my behalf because i'm going to find my way through life look at all the exposure i had but there are people out there who are single moms and dads who come from less privileged backgrounds or had a harder life and who can just not afford to be on one year contracts all the time while holding a phd we're losing all these bright people and i'm seeing them right they leave my institute and they go to industry and then the universities complain that they're not competitive we need to change the system we need to have more permanent positions in science i promise we won't perform worse if you give us permanent contracts we love what we are doing so back to the corona topic in order to understand the virus and its life cycle we need to understand its molecular biology this will help with the design of therapeutics we evaluated these molecular structures with a bespoke pipeline and expert knowledge provided context and reached out to the taxpayer and the general public to inform them we also wrote a paper together with long awfulists making the invisible enemy visible which is our motto and as we started this all with questions i'm going to end with questions structural biology remains difficult what can we learn from our findings should ask we as a scientists community change our attitudes towards errors should this serve as a model for other projects can it serve as a model for other projects i hadn't thought about this but a nature editor asked me when i was writing a comment whether we believe that science should always be like this my god that would be awesome i would totally be up for it if science would always be like this come together with a bunch of friends but without funding start doing something to you know fight a global pandemic then get some funding still having like no senior people on a project can open science compete i don't know we get pretty little quieted it would be definitely open science compete i don't know it doesn't quite look like it the still getting mattered only by the citations my paper gets and well at least the paper is not behind a paywall but if we would have published all our stuff in like papers we'd possibly get more credit i really don't know but we need to work on this if we want open science we need to change how people are rewarded how senior do you really need to be i'm 39 it seems i'm called a junior group leader all of us are young the youngest is 24 he's writing a first author article about a coronavirus protein i feel that the german academic system and all over the world actually you need to be older and older to become a professor and be permanent and be like a grand holder i don't think it's necessary i think that professors could be 30 and the world wouldn't you know like go down will what will change in science after the pandemic we had like large exposure it certainly will also have to do with like questions like did the virus now come from a lab or didn't it that you know would change how people view science i'm sure how will scientists be viewed by the public right um right now of course you know mom and dad are very proud but what's gonna change are we going to still be the bad guys because we are off mar but i'm like you know exclusively taxpayer funded i never took any money from the pharmaceutical industry i have like you know no stakes in this game i'm i'm just like earning tax money so i feel that there is a whole complex of difficult things there how people regard science but definitely the pandemic is going to change how science is going to be viewed what's that gonna be in the end i'd like to thank all the task force members and all our collaboration partners and our scientific fairy godmother arwin pierson who had little role in this research but much role in our mentorship and bringing us forward my home the university of hamburg the corona virus structural task force our fund we are funded by the durcherforschungsgemeinschaft and the federal ministry for research and education i would like to point out that we are looking for student assistance not only for scientific work but also for social media video and programming work and 3d printing so if you know anyone who's interested please point them in my direction my email is there we're also offering bachelor master and phd visas in areas that cover both lab work and computer work which is a rare thing these days you can find us on youtube and the internet and on twitter and i'm looking forward to the discussion and thanks for listening let's rewrite it so i just brought it up with the bmbf because i'm now sitting on these commentees more and more i'm reaching an age where i'm sitting on commentees and i told them legacy software is a problem the software is written in fordron every second line is go to so it's written like assembler no one knows what's in there anymore and if the person dies we're not going to be able to do anything about the algorithms they're just going to be lost exactly and so you have to first influence the grant writing institutes that they are grants out so that this software can really be published and this software is not easy so that's not your typical webpage so you need to work together in large groups and so it's a very interesting piece of problem i have to say so sadly as a PhD student we weren't we weren't looking into this but it's like this is too big of cake to eat it was very interesting so i can also there's a newer tool suite and that's also only maintained by i think three people or so so even that one it's not written anymore in fordron it's still not good so um just my so there's my PhD thesis i worked on shellix and that faces similar problems yeah it's it's used in the entire world to serve every small molecule structure there is more or less it's like yeah so questions further questions directly so we have got a first question i think around 20 30 minutes ago and that was if the yeah that's that's why lalu is so great as a signal angel she's picking up all of these ones um um so the first question was how do the virus variants affect the shape or the form of the virus i think they so no matter like okay this is the old model as we know then there should be fewer spikes and should be a bit smaller but nothing would change on this view like the scale is way too large the mutations each change about 10 atoms so every like amino acid at this different is about 10 atoms and those changes would be so small you could not see them on the virus model you'd actually need to look at the head of the spike in order to see the mutation and what it actually does so the changes are too small to show them in the model that doesn't mean they're not meaningful so as you've seen like the head of the spike binds to the host to the host cell to the ac2 receptor and that binding is highly influenced by this by the mutations now we thought that we may end up lucky because the same part where the antibody is binding is also the piece that binds to the host cell so everything that would make the spike change in a way that the antibody couldn't bind anymore to its head would have also changed how it bound to the host cell however it seems that Omicron is still able to bind human cells very efficiently while antibodies cannot recognize it so easily that may have to do with like this thing is actually like packed and you could imagine like putting cotton wool around it and it's called like oscillation it's got long sugar changes that like are wool and they're there to obscure the immune system like the antibody goes like oh there's wool I can't really find where I'm ought to bind is it here I don't know and that's changed in Omicron but it's not fully understand yet I saw there was a new structure this week but I haven't looked at it yet however the changes are too small to show it in the virus model they're like really tiny changes and another change that happens in Omicron as well is the proofreading mechanism when the RNA is copied is like damaged and we we think it's damaged so they're so called endo RNAs which is a proofreading protein its job is like to go like is the RNA correct yes correct correct that seems to be a little bit broken so it could be that Omicron is accumulating so many mutations because its RNA copy machine is like not working as it should is basically not proofreading that would mean that more viruses are produced that are not viable and cannot survive but it would also mean that it mutates much faster and we think that may have an influence but that's just theory so far this hypothesis we haven't proven it yet so this is why I haven't tweeted about it yet because it's just a theory but it would make sense right connected to that I would have a question the ACE receptor of small children is a little bit different than one of the adults do you know about that because Omicron is going towards more the smaller children at the differences I can't, I have read that but I haven't looked into it properly so I'm afraid I'm not going to answer this question because I feel I'd rather not say anything about it and say something that's wrong especially well we're looking forward to the structures in the public PEB so that everybody can look at them we live in an age of preprint and very often the PEBs are there when a preprint comes out which is how we called so many errors they published a preprint they put out the structure we went like there's errors in these structures and then when they published the actual paper everything was corrected luckily the changes I think are tracked on the PEB this is the second question yes that's what IT people like because then you know version history is very important there's a second question and it goes towards the tools that are used to simulate those molecules wait wait wait I have a follow up to this something that I would really like to see but it hasn't happened yet the PDB is a very static repository where only original offers are allowed to change their structures now imagine if the protein data bank would be like GitHub with pull requests where we could go like change the molecule around go like now it's a better fit to your data please pull that would be a very subversive proposition I would say yeah wouldn't that be nice I'm like why aren't we doing this it's like the system is already there it happens that we have repositories for a while in software development we could do the same thing with models fitted to our experimental data but I think I need to go into more committees yeah it sounds like this I would agree for that proposal there was a person that was asking about the tools that you would use to simulate those molecules and structures and so on and if you then create these on more usable pictures for the public how do you balance artist impressions, simplified models while keeping the scientific truth as much as correct as possible yeah the question doesn't you don't need the public even to to have this problem you have it already when you make pictures scientifically because sometimes you want to show a certain aspects of a molecule very clearly which means you have to cut away for example a part of the molecule so one answer to this is the program that does the modeling is not a program that you use to make the pictures that's the first thing you do so you make your model with one program and then you take all the coordinates of the atoms and you throw them into a professional rendering program that like will do it all pretty but you still have to make an executive decision on what to show in your graphics in your paper and I think that in particular as structural biologists who deal with three-dimensional and two-dimensional pictures we would do very well to think a lot more about scientific illustration so all my team likes thinking about these things so just I think how we got them where we are now right they actually like stuff like this they go like we can print it 3D and then we can put a magnet on it and it's gonna stick but it turns out that scientists also need these tools to understand what's going on it's like actually having a 3D model helps you so much with thinking about things Crick and Watson they build a model of their DNA in metal for a reason because we're looking at three-dimensional things making them like understandable with our hands so yes as a good researcher you're not only able to explain your research to the cleaning woman you should also be able to visualize it properly and it's part of the art if you are a structural biologist and you're not able to make good pictures of your molecules you're not a good structural biologist end of story you're in the wrong discipline you should have possibly chosen something where you need less graphics and I think one of your illustrators is actually a structural biologist by training, right? I think some of them Shanuka Thomas-Sello is a proper biologist Thomas Splittstrausser is a proper biologist proper like a PhD in bioinformatics and Janet Iwaza and Liu are both scientists who are having a group that only deals with illustration so it's actually in science we have several groups in the world in structural biology who only do illustration as science so David Guetzel with his watercolors is very well known but Janet Iwaza is another one so actually making these animations and pictures is so complicated that television can't do it and the ZETI-F made a series of animations so they made very nice ones for Planet A, Viso and Nano with us but then they asked my expertise to make another animation and they only had like a call with us and they never came back and then published a completely wrong animation of the entire viral mechanism under my name and I'm still sad every time I see it three times a day heute journal shows a wrong structure of the virus for which they claim that I helped them make it and I wrote to them and told them your depiction of the virus is wrong I can help you make a new one but it seems that the zweite Deutsche Fernsehen der heute journal at least didn't care and I guess they think it's not important enough but I think with a thread like this where we really cannot see the virus it is important to bring to the public the best possible depiction we can deliver sometimes however as I said you omit certain aspects for example to show the effects of a mutation you'd only show the sight of the mutation you don't show all the atoms but that's really an important part of what we're doing like pointing out the important bits and that's why scientificization is so important I think we have one final question which I would say comes out comes towards the direction of the immune system and the question would be can you can you define a vaccine on purpose so that the immune system can forget how to produce the antibodies after a defined period of time and I think the concern here is about increasing your your financial gain in a crisis for example so programmed obsolescence is a word that was mentioned so we're a little bit late so you could keep it the short that would be great okay um the quick answer is no that's not possible um you can make you can enhance how long vaccines and how much immune response you get from a vaccine with certain additives but you cannot like make them a definite time because everybody is different so even when you get your booster shot they say six months but you may need it after three months or you may need it after twelve without a titter it's very hard to tell and the pharma industry does not have tools that would allow them that to my knowledge so i think that's not a risk i mean it's it's a human system it's so complex it's easier to shoot a rocket to the moon so but i think it's a it's a valid concern it's just technologically not possible luckily i guess the final and really last question i think is where can somebody find the 3d models oh um you can go to insidecorona.de and find a blog post about a 3d thing or you go to thingiverse and you look for insidecorona um i can yeah yeah i just go to i'll just put it in the that's our home page and then on thingiverse it's also called insidecorona and um you can also write me a message on twitter if you don't find it at at tornlab and remember we're going to put out a new model soon in january but i'm still waiting for the final files and holiday so it will be a few more weeks so better print in january not in december than you have today yeah exactly or print two yes thank you very much for having me yeah it was great thank you and i think if there are any no more questions everybody all right oh you know how it looks like i think that's really important so i think one and a half years ago i came across the first picture so i was like this is how it looks like now i can tell it to the people who who don't read the scientific original papers because they are so difficult to read so yeah all good and thanks for being here and looking forward to hearing and seeing more and hopefully once this will be over i think i hope so too going back out of the spotlight to being just a methods developer that would be nice that would be nice yes