The Genetics Podcast episode 52: Population-scale viral sequencing with Dr Jeff Barrett, Lead Statistical Geneticist for COG-UK. --------------------------------------------------------------------------------------- --------- Patrick Short: Welcome to the genetics podcast. I'm really excited to be here today with Dr. Jeff Barrett, who's on the podcast now for the second time. His first episode, which was episode 27, was focused on how genomic data (human genomic data) is used in drug discovery. A lot has changed since then obviously, most notably the COVID-19 pandemic has happened. And since then, Jeff has changed his day job from focusing on human genetics to now he's focused on virus genetics and specifically COVID-19 virus genetics, where he's the lead statistical geneticist for the COVID-19 genomics surveillance UK project, or COG-UK as it's known. So, first of all, welcome back Jeff, and really excited to speak to you. --------- Dr. Jeff Barrett: Hi, Patrick. Great to be back again. --------- Patrick Short: So just to start off, maybe you could tell us, what is COG-UK and what's your focus on there and maybe a little bit about the personal switch from human genetics to virus genetics? --------- Dr. Jeff Barrett: Sure. So, COG-UK is a group that was put together by Sharon Peacock, who's a professor at University of Cambridge and she very early in the pandemic brought together a group, including epidemiologists, virologists, genome experts, public health experts to try to use genomics as effectively as possible in responding to the epidemic in the UK. And that actually happened before I joined, and they've really achieved some pretty amazing things. There's a combination of sequencing efforts from a number of different sites that include both community based testing as well as hospital-based testing. There's a sort of centralized bioinformatics infrastructure that supports the whole project that's principally run out of Birmingham and Cardiff, and there's a lot of good collaboration with the public health agencies that are leading the response to the epidemic. So I came, um this summer I joined COG-UK. I came back to the Sanger Institute where I had previously worked, to lead their efforts (specifically the Sanger's effort) as part of the broader COG-UK initiative to try to see how by sequencing viral genomes at a really unprecedented scale (that's the thing Sanger is good at in particular doing things just on uh, in terms of huge numbers) - and how that could be useful in the epidemic. For me personally you know, I was, I was interested in a new opportunity and really wanted to be able to make a contribution to the pandemic response. And, I realized that actually, it's also kind of interesting to sometimes try to approach a new scientific challenge. You know, I've been working in human genetics for a very long time, and it still, in some sense is where my heart is, and it's fascinating, but this opportunity to work in a related, but somewhat different area that is analyzing the genomes of the virus is really exciting because it kind of used some of the skills that I've up, but also presented a cool and new opportunity. --------- Patrick Short: When you say sequencing viruses at an unprecedented scale, what kind of scale are you all operating at? --------- Dr. Jeff Barrett: We've been sort of scaling this up over the course of the pandemic and the, the effort that COG-UK put together can really be seen. If you look at basically all of the publicly shared SARS-CoV-2 genomes from around the world (so just for a little bit of background, the SARS-CoV-2 virus has a genome about 30,000 bases long, it's encoded in RNA - an RNA virus, and that genome was first sequenced very, very early in the pandemic.) When isolates were identified in China and a number of groups, including COG-UK really initially got together to set up good protocols, lab protocols for how to basically rapidly sequence the virus over and over again. And the reason we want to keep sequencing it is because viruses like any other organism, their genomes accumulate mutations over time. And so, you know, if I'm infected and I transmit to you, maybe the virus has added one mutation, so you can see its genome is slightly different than it was when I first got infected. And as we'll probably talk about in a minute, we can use we can use those mutations that accumulate as a sort of barcode to track and watch how the virus spreads. And so that means we want to do this as much as we possibly can. You know, in an ideal world, you might sequence every single person who gets infected because it gives you a really rich data set to track the spread of the virus. Coming back to your question about what scale have we got to, if you look at there's a website called GISAID, which is the kind of global public sharing website - it has, I think right now about 220,000 or something viruses from around the world, virus genomes, SARS-CoV-2 genomes. And about half of those are from the UK, so the UK is really contributing at a massive level and about half of the UK is contributions from the Sanger Institute where I work. And that kind of gives you some sense that both as part of this national network, we've really done this at a scale that nobody else has done. And specifically Sanger has really built now a, quite a robust kind of ongoing pipeline where we sequence somewhere around right now, 5,000 genomes taken from tests that were collected from around the UK and basically process those and upload them to the kind of shared bio-informatics platform for analysis. --------- Patrick Short: I have 5,000 tests - is that per week? --------- Dr. Jeff Barrett: Yes sorry, that's per week. --------- Patrick Short: Per week, yeah. You touched on this, the idea of using the mutations as the virus jumps around as a sort of barcode - I wonder if you could just talk a little bit more about that. Do we know how per transmission, how many of the 30,000 bases change on average? And obviously the question that's probably on a lot of people's minds is how often did those changes cause the virus to do something different, you know, be more severe, more contagious or, and how often are they kind of just like a passenger or, or a mark that we can use to map It? --------- Dr. Jeff Barrett: Yeah - so the virus is relatively slow mutating. So it accumulates, I think the widely quoted estimate it is about one new mutation per two weeks. And, you know, if you assume the estimate of the sort of interval between transmissions, I give it to you and you can somebody else and so forth is about five days. And that means most immediate transmissions have identical genomes. And then occasionally one per say three or something along those lines, you add one mutation. So that means that at the very early part of the pandemic, when it's spread very quickly around the globe, there was not a lot of diversity and almost all the sequences look very similar. We've now had some months of mutations accumulating. And so you do see a lot more diversity, but again, still things look pretty much identical when you get a bunch of immediately close transmissions in time. That's very different than say influenza or HIV where both the mutation rate is higher, but also because they've been circling humans for a much longer time, the total diversity is much bigger. So if you kind of look at the diversity of sequences in SARS-CoV-2 compared to HIV in currently humans say, it's very, very different, there's way more diversity in HIV. So what's kind of interesting is there isn't a huge amount of diversity, but there's enough now that we can use it in this barcode kind of way, where you can start to trace things and see, you know, in any given point in time, basically what's the balance between local community transmission, possibly some imports from another location where someone has traveled and brought something in that hasn't been seen in that location for a while, and then also sometimes you see suddenly the same sequence over and over again, which might be indicative of something like a super spreading event. And maybe we can come back to that in a minute. The other thing of course is as you point out these, the mutations in the viruses, RNA are not, they are a barcode - they can be used just as a label, but they're also the biological instructions for the virus. And so they might have meaning in terms of how it behaves. So far there have been relatively few examples of clear cut, biological changes in the virus due to mutations. There are a couple of exceptions and they, I think they kind of broadly fall into two categories. So category one is changes in how rapidly the virus spreads from person to person. And you can see these ones in principle because they can rapidly increase in frequency in the human population, because basically if a mutation makes a virus more transmissible, it will outcompete the previous version and will spread faster and faster. You have to be very careful in doing statistical analysis of this because you can also have mutations arise and get more common just by sheer chance, you know, genetic drift or it happens to get lucky and go on a trip to a massive super spreading event. And you'll suddenly see lots of that version, even though it's not biologically interesting. But you can develop Cisco models and a guy called Eric Falls, who's also in COG-UK has a really great example of this and together with other colleagues, they published a paper looking at perhaps the most famous of the SARS-CoV-2 mutations, so called D-6-14-G, which basically has gone from it wasn't present in the original sequence of the virus in Wuhan in China where it arose, but now it is almost completely ubiquitous - the new G version is almost at a hundred percent around the world, and the D version is almost extinct. And that one has a pretty clear signal that the increase in frequency over the past year is because it has a selective advantage. And there's actually been a whole bunch of papers - people study this in vitro of course, you can look at two lines of the virus, differing at just that one position, and then ask, you know, in this particular example, they've shown in say human lung cells on a dish that basically it infects faster. They've also done experiments, it turns out hamsters are a good model that can be infected by this virus. And you basically infect two hamsters with the different strains. And then in like the adjacent cage, you put an unfortunate uninfected hamster. And the first thing you see is that they pretty much all get infected if they're exposed in this way long enough, but it happens on average a day or two earlier in this G variant. So it's pretty compelling evidence from multiple different angles that it basically is more infective. Now, the good news is for as much as people have looked, there doesn't seem to be any difference in the severity of the G variant of the virus. So it basically seems to spread around the world, but it's not obviously giving people more severe illness or anything like that. --------- Patrick Short: Right. And I think you mentioned on Twitter as well, we will, I think we'll talk about some of the exciting new data coming out of the vaccine trials, but many of those vaccine trials will probably be a mix of the D-6-14 and G-6-14. Right? So it's not no one's really, or, or, or would you, I think you were maybe arguing on Twitter that it's likely more of one. --------- Dr. Jeff Barrett: So there's an interesting thing, which is that because humanity was so fast in responding to this virus and deploying all kinds of different responses, but including development of vaccines. So this mutation we were talking about, D-6-14-G what that means is it's in a particular protein encoded by the virus's genome called the spike protein and the spike proteins are what give the corona virus (corona means crown in Latin) - and the name is because it's sort of dotted with these spikes. And so if you look at a microscopic image of the virus, you can see all these points sticking out, and those points are this protein, gold spike. And that's what it uses to essentially bind to human cells or other animal cells. So, because that's the kind of a key part of the human to virus interaction, it has a lot of focus, both from vaccine development and from just general research, many of the vaccines that are currently in trials, including all three which have reported results so far indeed work by trying to get the body to basically recognize a fake version of this spike protein and all of them use the sequence - so that's D-6-14-G - is that's a mutation in spike and the 6-14 is the 614th amino acid, and it's - I'm not a bio-chemist I actually forget the - maybe you know your amino acids better than me - but it's a D in the wild type and a G in the mutant version. I'm giving away my, my statistical rather than biochemical background. But anyway, all of the existing counts are to this D version, which was present in the original when the virus arose in humans in China. And that could be a bit worrying because as I just told you, everyone who's getting infected today is getting the G to a first approximation because D is basically extinct. There were some papers before any of the vaccines read out that suggested basically that the immune response to the D version would also work against G. So it's not automatically the case that a new mutation means that immunity doesn't work. In fact, it is very rarely the case. And so that's really good news, of course. And as we'll talk about in a minute, the fact that there has been good news about vaccines on this particular D-6-14-G is especially important because - especially reassuring because - pretty much everybody who got infected in those clinical trials got infected with the G because there just isn't any D left. And that means that the vaccines do work against G, which is a very big relief. --------- Patrick Short: Yeah, absolutely. And is it fair to say that there haven't been any kind of credible reports of mutations that have increased severity of response? I know there was one paper a little while ago, but it was fairly widely debunked because I think some of the cases came from one particular hospital, which was only sequencing or primarily sequencing dead people. And so those people carried a - because they were from one area -they had a bias. --------- Dr. Jeff Barrett: Yeah, exactly. The world is, uh, the world is full of statistical traps. And, you know, when everyone is kind of trying to work fast, it's easy to fall into. And exactly, there's this one example that has got a little bit of chat which is, you know, anyone who studies population genetics knows that the different variants will be at different frequencies in different parts of the world. So that's true in humans - if you look at a genome from a person of European ancestry and a person of south Asian ancestry, there will be some differences which have no biological meaning, but just happened to - over the course of human history - have been more, more common in one place or the other. And the same thing happens here - the virus moves around at someplace gets a bit more common. I mean, exactly as you said, one project, you know, sequenced basically dead people because that was the project they were doing and they uploaded their data. This was in Brazil I think - they uploaded their data to the big GISAID repository - I told you, which has hundreds of thousands of sequences from around the world. And if you just do a sort of naive analysis where you say, I'm going to check for associations between variance and the status of being dead, you know, you're going to get all of this confounding that will look interesting, but really it's not at all because of the things we were just talking about. So I think more generally it is true that I haven't seen any credible versions where specific mutations in the virus are linked to increased clinical severity, basically a worse course of disease, so that's reassuring so far. The thing which I think we have to be a little bit potentially concerned about, not yet, but maybe in the future, is there have been some mutations that have arisen in the wild in humans where unlike the D-6-14-G we were just talking about, there is some evidence, in the lab at least, that they can - basically the virus isn't as easily inactivated by the immune response that you might get from the wild type. So what I mean by that is, let's say you got infected, your body will produce both a sort of antibody and T-cell immune signature, and that will have some specificity to the genome of the particular virus infected you. We just talked about an example where the D-6-14-G situation where actually, it doesn't matter - your immunity is general, but there have been, as I mentioned, some mutations where that doesn't seem to be the case. In other words, the experiments that are often done are: I take some antibodies from a person who has already been infected and then recovered. And I ask can those antibodies inactivate other versions of the virus in a laboratory. And sometimes they can't. And those specific mutations are a bit worrying. So far most of them have only ever been seen very rarely. So it's not like they're sweeping around and something to be really worried about, but they could be situations where it'd be possible, for instance, for you to get infected a second time by one of these, or if you're vaccinated, for you to be more susceptible to a version of the virus with one of these mutations. And that's actually one of the things that we think is super important about the work that we're doing in COG-UK is to have the infrastructure to monitor for just the tons of new mutations. Hundreds of mutations arrive and disappear every week, because there are so many people who are infected. And what we want to know is: what do those mutations look like? Are any of them likely to have a sort of immunological response? And again, it's certainly the minority, but are any of those changing in frequency because one might start to get worried. And in particular, once we begin vaccinating people and you know, it's super exciting that it looks like we're going to have several really highly effective vaccines very soon, you know, that evolutionarily changes the virus's environment - it puts a selective pressure where basically there'll be lots of people who are now immune. And so, you know, versions of the virus that might spread in a naive population might end up dying out. And there might be a selective advantage to versions of the virus, which aren't so fit in the kind of naive population. But if they're marginally better at infecting a vaccinated person, they suddenly have an advantage. And that's the thing that humanity has to be on the lookout for, because, you know, we want to be able to see that as it happens and then maybe change the combination of different vaccines that different countries are using at a point in time. --------- Patrick Short: It's fascinating. It feels like we're entering a new phase. I don't know if it's the second phase or third phase or what phase quite that we're in, but the questions are changing, as you said - I think we're a couple of days after most of the Pfizer BioNTech and Moderna and the Oxford AstraZeneca have all put out, you know, varying degrees of really positive results. So what I'm interested in is how does this compare to some of the examples you mentioned earlier? The ones that come to mind for me are seasonal flu (influenza) and HIV. So I'm very much not an expert on this, but obviously with flu there's constant surveillance and they try to create a vaccine that is going to match the seasonal variant that that is predicted for the next season, so it's a constant kind of year to year battle. With HIV, I guess, a little bit different where the virus is evolving multi-drug resistance, but a combination of several drugs in combination can kind of back the virus into a corner. I'm wondering what between those two, how is SARS-CoV-2 likely to be similar or different, to the best that we know today? Obviously we're not asking you to predict the future evolving pandemic. --------- Dr. Jeff Barrett: The first thing I'll say is I'm not an expert at all in this either. So I'll give you my best understanding, but I don't say any of this with a hundred percent confidence. So one thing is I'm pretty sure that the reason the flu vaccine changes year to year is there are reservoirs and humans have multiple quite diverged strains of flu. And basically the sort of global community is constantly playing this game where I think, you know, because basically depending on the Northern and Southern hemisphere, the flu seasons kind of chase each other around the calendar. So what they're doing is trying to figure out in high flu right now, what's the mix of the most prevalent strains, so they can try to vaccinate in the other places you know, when the flu season gets to them. And so it's (again, as far as I understand) it's not per se that flu is kind of really strongly evolving, but more that there's a pool of standing variation. And basically as you vaccinate against different strains that selects back and forth on that pool of standing variation. So one thing is that I don't think there is nearly as much yet standing variation of SARS-CoV-2. So the question is, will we get to that point where, you know, it evolves enough variation that it becomes... I think one possibility is it becomes endemic and every season there's a SARS-CoV-2 wave alongside of flu. You know, I would definitely not want to predict, but I think that's a possible outcome. And so I think there's a kind of long-term, which is we, we ideally want an effective vaccine, even if it has to be administered every year, you know, that's certainly a much better situation than having a kind of low level simmering amount of this deadly disease all the time. I'm still focused really on an even nearer term thing, which is just, you know, we're still really in the grip of this current pandemic. And the vaccines I think, are going to make the difference in 2021. We just don't know yet I think whether it's possible that a mutation can rise so quickly that even within one season, it could escape a well vaccinated population. I don't think that's likely, but I don't think we know enough. And that's sort of why our focus is just keeping this surveillance going, such that week by week we are just seeing what is the distribution of variation in the infected population? And if things start to change, we can see that and hopefully react. And then, yeah, like I said, that will sort of end up playing itself into what is the five or 10 year future for humans and this virus. --------- Patrick Short: So what does the next six months, say, for you all look like? I think it's fair to say that there won't be large scale vaccination in that time. So is it, is it getting a better understanding of how the virus is changing and evolving and what's going on at a population level to then inform that longer term strategy? --------- Dr. Jeff Barrett: So I think there are broadly speaking three ways in which genomics is useful in the epidemic, one of which is perhaps the way that genomics has most widely been used in infectious disease epidemiology before this pandemic, which is to investigate a specific outbreak. You know, so for example, there was a really nice paper that came out from some data from the spring wave in Cambridge, where there was a really high rate of infection in a renal dialysis kidney dialysis unit. And those patients are at a super high risk group. So, you know, there's unfortunately a very high fatality rate if you get infected and you're on dialysis, and the genomic data, they used alongside other kinds of epidemiological data, you know; which healthcare workers worked on which shifts and this kind of information, help them pinpoint that it wasn't actually the dialysis unit itself that was the most important thing to change, but the transport of patients for the dialysis. So they had been using, you know, pre pandemic, like shared taxis to bring a bunch of patients together for their thing. And that turns out to be a risky transmission route. And so they moved to a mechanism of spacing people out and sending them in isolated cars and so forth. That made a difference. So that's one area. For that you kind of need, generally speaking, to sequence a bunch of specific individuals from this hospital or whatever, you know, and have a bunch of information, detailed information about them. What we're doing at COG-UK is doing a lot of that kind of work. What we're doing at Sanger is, is two other use cases, which kind of sit alongside that, which I group into surveillance. So one we've talked about a lot already, which is vaccine resistance surveillance. And for that, you just need, again, at large-scale, random geographic sampling every week of infected individuals. You sequence them and you watch the frequency of mutations change. So that's one piece, and we want the infrastructure that we have now to run for two years, I would say, you know, I don't know how long it's going to be, but we want it to be able to run for two years so that we can provide this service. And then the third one is also a surveillance use case, and it is more immediate than the vaccine one. And it's to try to provide an additional data feed, to detect super spreading events as they happen. So what will happen if there's a super spreading event in a factory, say, there will be a lot of transmissions and then those transmissions will, you know, there'll be secondary and tertiary transmissions as the people who originally affected the first event transmit to other people. And you'll see, maybe weeks later, that in a particular location, there's suddenly a big surge in cases. And there's been lots of these reports in the news. There's one in the US - this famous Maine wedding where a huge number of secondary people were infected. And there are a number of deaths and none of the deaths were guests at the wedding. It was all the kind of ripples that went out from that. So it turns out these things also, as you might expect, leave a very strong genomic footprint. So basically, if you have a bunch of transmissions, as we talked about earlier, all in a short period of time from one source or a very closely linked set of sources, the genomes basically look identical. And if you're just sampling in the community, usually you see this kind of random mix of different versions of the genome, depending on what's circulating at that point in time, in that point in space. But instead you sometimes see all the exact same genome over and over again, and that's a signal that something has happened differently. And basically we think... so we can currently get that - we can currently run an analysis across the UK on any given day and say in the previous seven days, what did this look like? And we can put a pin in places in the map where basically there's evidence for this kind of pattern where there's likely a super spreading event. We're currently lagging by about two weeks from when the people were infected to when we can get the genomes into analysis. And that's, if I do say so, a very impressive turnaround time, given the logistical complications of, you know, the samples get sent to these central testing facilities, the Sanger runs vans up and down the country every day to bring the residue from the tests to our facility, we have a huge robotic line that sifts out all of the negative tests and disposes of them. It puts the positive tests into the sequencing pipeline. We do the genome sequencing, we do the bioinformatic analysis, we stick it into a statistical model. So that's about two weeks. We need to get it done to ideally about one week and we're working on that. And so the plan is, as the UK comes out of the national lockdown, it's going to go back into a tier like system, and genomics isn't going to magically solve any of the problems, but it can be one piece of information that sits alongside the local number of positive tests, the number of people who are going for tests and other factors to basically say, do we need to increase or decrease the restrictions on a local basis? And the genomic data we think can maybe move some of those decisions forward by a day or two days or three days. And that's a little bit, but the key thing in an exponentially moving epidemic is any intervention you move forward by a couple of days, saves you a bunch. And if you multiply that by a lot of interventions, that can be the difference between a third national lockdown before vaccines become widely distributed. So that's really our, our near-term focus is, can we just provide a useful feed of data that sits alongside other things that will make public health decisions a little bit better? And then the hope is that that basically just gets us with fewer infections, less disruption to society, to widespread vaccination. And then we're in this mode of just watching to make sure that mutations don't develop. --------- Patrick Short: So how does that genomic data driven decision-making kind of actually work from an operational standpoint, do you all have a dashboard that senior members of government and NHS can say, you know, here's a hotspot developing in this particular part of the country. Is that where it's headed? What what's it like today and what do you envision it to be? --------- Dr. Jeff Barrett: Yeah it's a good question. And not an easy one, because the response is really a very multi-agency, multi-level kind of activity. And one thing I've learned is that there isn't this kind of panopticon somewhere that is flipping the switches. So what we're trying to do, we do have a dashboard like that - at least a beta version of it, what we're trying to do is find as many audiences as possible who are interested in this kind of feed and just get it in their hands - that might involve the central public health agencies. And those are actually devolved, so there's public health, England, Scotland, Wales, Northern Ireland. And some of the devolved nations are, well in Scotland for instance, have extremely strong teams who are using genomic data in those places already. There's also the regional directors of public health who have really kind of, they're kind of amazing jobs in that these people would also, as I understand it, have responded in the pre COVID era to the Novichok poisoning, from a couple of years ago, or I think like an anti-smoking campaign, because, you know, they're covering all aspects of public health, and now they're swamped with also obviously this one enormous public health piece of work. And so they might also be interested in this kind of information, and as they look week by week in their localities. And the reality is we don't yet have quite the right thing to put in their hands. And that's what we're sort of feverishly working on right now, so that hopefully early in the new year, we'll have a tool that is, that is useful to those kinds of people and that will be in their hands on that kind of timeframe. --------- Patrick Short: Yeah, because it seems to me like you've got really two streams, you've got the longer term scientific discovery, but you also have what is ultimately a product that you can put in the hands of decision makers and help them make better genomic data-driven decisions, right? --------- Dr. Jeff Barrett: Yeah, I mean, it's one thing that's been really interesting for me personally. So I know I was in academia for a long time. And then I moved into into a role in a company - in a small company. And, you know, doing science in a non-academic context was really interesting to me, in the sense of you have a product and you have to really, that's the focus, not necessarily on publishing your results, which is great, but it's on building a product that serves needs. And this project is, although it's in an academic institution, is very much like that. You know, we have to have a product that has to actually be something that people want, that these customers who are the public health officials want to use it and that it fulfills their needs. And I've really enjoyed that kind of mindset in, in trying to build something that hopefully has some use. --------- Patrick Short: You mentioned that the UK was the biggest, if not one of the biggest contributors to the kind of global viral sequencing - who else is doing this on a large scale? And, and why isn't the US on that list? For example, it's got certainly a lot of, a lot of viruses to be sequenced. --------- Dr. Jeff Barrett: For sure. It's an interesting question. I think the US is probably second if you add up all of them, but as far as I know, it is much more that there are many disparate initiatives which are producing data, it's just because there's a lot of data that they have contributed. But actually it's, there's data from all over the world - this is clearly a global thing. The UK is clearly the biggest contributor, like I said, it contributes about half of the total. So it is as much as approximately everybody else put together. I think there are three things which make that possible here. One is the fact that a lot of the community-based testing is centralized in the UK. You know, this has got some criticism, but one advantage of what happens is basically there are all kinds of different places, home testing kits, carpark, testing sites, et cetera, where the swabs are sent to a small number of facilities, the tests are done, and the results are given back to the individuals separately - of course, there is testing in hospital and healthcare settings, but that centralized community testing means that for this kind of thing, you can access the samples and sequence limit at a scale that is really hard in places which are completely distributed. So that's one, one event from the UK. The second is it has a facility like the Sanger Institute and more broadly, a wide variety of genomics expertise. You know, there are like 16 other sequencing centers that are part of COG-UK and, you know, the UK really is a leader in that. And so I think it was always kind of fertile ground for being able to do genomics at scale. And then the third piece is really the leadership of Sharon Peacock, who, as I mentioned, put COG together, in coordinating across those different levels of government, of academics, of universities, of hospitals to make this thing really work, to put all the data in one single repository, such that everyone can analyze the full data-set instead of fighting over different bits. And I don't think anywhere else has quite done it on that, on that scale, or in fact, I'm sure, nowhere else has done it quite on that scale - it's I think something that everyone who's involved should be proud of. --------- Patrick Short: Yeah, absolutely, it's really great. I'm just conscious of time here - I have one final question, but before the final question, there was one thing that was in the news a week or two ago about the minks and Denmark that I wanted to ask you about. So this was a, and I'm not actually sure where the story has gone, but there was discussion that millions of minks would need to be culled because the SARS-CoV-2 had jumped from humans to minks, mutated and had potential to jump back. But I know there's a ton of questions around, you know, this happens all the time - it's jumping from humans to animals. I'm just wondering what was up with that story, and is that, you know, is the virus jumping into animals, domesticated animals, and back, in a wider context, makes this something that you all are thinking about or that we should be thinking about more going forward? --------- Dr. Jeff Barrett: So the first thing is, this virus can infect other mammals, and there's various documented versions, including domesticated cats and dogs for instance. I don't think it's super common, but I don't know how closely people have looked. You can clearly infect mink. And the reason mink became a headline grabber is (I didn't know this until a couple of weeks ago), but being apparently kept on farms that have really quite grim housing conditions where there's basically like thousands of minks kind of piled on top of each other. And that, as you can imagine, is a super transmission friendly environment if they get infected with anything. So it turns out that SARS-CoV-2 has jumped to minks, at least a couple of times from people. In the press I've seen at least in the Netherlands and in Denmark, but it may have happened to other places. So that I think is not a huge surprise. Again, I'm not an expert in sort of zoonotic transmission, but it seems like that is something that happens with respiratory viruses. What seems to have happened are a couple of things. One is, it looks likely that there is one mutation that makes the virus more adapted to mink. Again, it's in the famous spike protein we've been talking about, I think it's called Y-4-53-F - that one seems to keep popping up in mink. And so there's the inference of (I'm not sure it's been proved), is that it basically makes it easier to transmit within mink - that variant can go back to humans. They also in Denmark observed some other variants that involved a few mutations that had been piled on top of this Y-4-53-F in one particular group of minks. And they were worried about that one because it seemed like it could, it also could have this property that we discussed earlier, where it might make it less, that version of ours is less susceptible to being activated by the antibodies from essentially previously infected recovered individuals - that was what made them very nervous. I think the authorities in Denmark were potentially worried about some mutations that could in theory, confer resistance to the immunity that one might get from the widely circulating versions. You know, I can see why the authorities got concerned. There's a sort of nightmare scenario that in another animal reservoir, a version which can escape our vaccines or our existing immunity evolves, and then it comes back and we have a whole new pandemic to deal with. From people I've talked to, I don't think anyone needs to kind of hit the panic button. Because as we've already talked about, the virus is mutating all the time and there are millions of infected people. If you think about the effective population size of - each of those people have, you know, millions or billions, you know, there's lots of mutations that are happening to people anyway. So that's, I think the bigger concern right now than a particular reservoir, I do think they did a huge cull in Denmark because they were worried about it. So I think it's good. I think that my take home message from that whole story is it's good that people are aware of it. It's probably not something that we need to right now be terrified about, but it is probably a slight element of reality we should keep in mind that as you know, we're excited about vaccines, making things better, there are risks that are out there about vaccinating a population, but then having a second proper wave at some point in the future. And obviously that's something we want to avoid. --------- Patrick Short: Just a final question to wrap up here. I think we've explored the case of what happens, if, let's assume that this round of vaccines works really well. If there is mutations that mean the selective pressure pushes this current version of the virus down and others evade, we may be in a kind of yearly, or, you know, some frequency battle with the virus, but the alternative case is actually, this doesn't happen, the vaccines work really well. And the virus is more or less pushed down to eradicated, pushed to a minimal level. And the population in that second case, where, what do you see as the, as the longterm goal of, of what it is that you all are building? Will you be able to build a kind of pan-viral surveillance system, because this won't be the last time it happens to us as a species. I'm just interested in what that, that long-term looks like. --------- Dr. Jeff Barrett: Yeah. Um, I think, you know, a number of colleagues involved in COG_UK are in particular, hoping to establish, genomic surveillance of infectious disease as a capability for the UK into the future. And that might be for another, you know, SARS-CoV-3 that comes from another zoonosis sometime in the future, if we're so unfortunate, but it also might be to monitor other infectious diseases that circulate at some much lower level, but still circulate; you know, methicillin-resistant bacteria or other things. And I do think we've deployed this particular tool of genomic surveillance on a scale that I don't think has been done anywhere for any disease in the past. And so it does seem like there's an opportunity to try to take this, this capability and perpetuate it, and I think that hopefully is a lesson that society will apply in lots of different things where we learn what are valuable things in response to this kind of event. And I think the sad history is that often after these, everyone says, oh yeah, we must preserve this and be prepared for the next thing. But then it often falls by the wayside because there are other priorities come, and it's hard to spend money supporting something that is for the future. Maybe this pandemic has been so damaging that that way of thinking will finally be kicked out. And so I very much hope that there'll be some mechanism found to basically keep this capability going for, for other infectious diseases in the future. --------- Patrick Short: Yeah. It's a really good point. One of our earliest COVID-19 focused guests was Dr. Angie Rasmussen. And she told me about a large scale - this came out pretty early in the pandemic - but global project and research program that was focused on zoonotic surveillance to try to identify things like this before they jumped, but it was defunded or, you know, massive cuts a couple of years before this happened. So hopefully we can learn, as you say, from past mistakes and not defund this kind of forward-looking research that you know, I think of it at, at worst, it will help us learn something new about, from a basic science perspective and at best it can help to, to nip a future pandemic in the bud, which seems like a pretty good ROI for me from a government science budget perspective. --------- Patrick Short: Well, thanks, Jeff. I really appreciate you taking the time. I certainly learned a lot. I had a bunch of questions that you very helpfully answered, and obviously you've caveated where you're not an expert on some of these, but it seems like you've you've made the transition from human genetics to viral genetics with with no problem at all. And I'm just really glad and thankful that you chose to take this career jump at this time, because you couldn't have picked a better time, I think, to jump to a new organism, new part of the tree of life. --------- Dr. Jeff Barrett: Thanks a lot. It was great to chat with you again.