TWiV 1357: Furin, Naturally with Paul Bieniasz

0:00This Week in Virology, the podcast about viruses, the kind that make you sick.

0:10From Microbe TV, this is TWIV, This Week in Virology, episode 1357, recorded on September 11th, 2026. I'm Vincent Racaniello, and you're listening to the podcast all about viruses. Joining me today from Ann Arbor, Michigan, Kathy Spindler. Hi, everybody. Here it's 76 degrees, and it's just beautiful. Blue sky, white puffy clouds, really, really nice. Again, I don't, the app that I used to have that gave me the temperature in Celsius is not here anymore, so just imagine it's perfect. It's close to room temperature, so it must be close to 25-ish. Here in New York, it is 27. It's sunny. It looks nice out, but I have no windows here, so I can't tell.

1:05Also joining us from Western Massachusetts, Alan Dove. Hello, and it is gorgeous here, too. 77 Fahrenheit, 25 C. And it is a CAVU day, as they say in aviation. Ceiling and visibility unlimited. Clear blue skies. Gorgeous light winds. CAVU. Is that a thing? CAVU? Yeah. Ceiling and visibility unlimited. You'll love it. those conditions. Wow. Just means no clouds and it's clear. Our guest today is right here in New York City on the other side of Manhattan from me. You can just take the cue and get there very reasonably. He's at Rockefeller University. He's been on TWIV before. Paul Beanash, welcome back.

1:47Hi, Vincent. Hi, everybody. Great to be with you. Lovely day here. Not greatly different weather here than on your side of the island. And taking the queue for those non-New Yorkers doesn't mean getting in a line in this case. It's a train. Yeah. So yes, I am in my office. This is all natural light. Yeah, he's got a big window looking out on the East River, right? Nice. It's very nice. I bet you watch the boats because they're so cool, right? I do. After a while it gets a bit same old, same old. This time of year you can have some really quite impressive expensive yachts sailing by and sometimes I Google them and it's like a million dollars to spend a weekend on.

2:31You're near Hellgate, right? You can see the intersection of the three rivers? A little south of there. You could probably look to an angle and see. Yeah, we could. If you enjoy these programs, we'd love to have your support to produce them. You can go to microbe.tv slash contribute. And yes, today is 9-11. It's 25 years after that terrible incident here in New York City. And there are a lot of remembrances. This morning on my way in, I passed a high school in New Jersey. Many of the states had people who died in the towers on that day. And so this was a high school. They had put a flag in the ground for every person who had died.

3:16So there were thousands of flags. It was very moving. So RIP everyone. Well, Paul is here because at ASV, he gave a talk, which was amazing. So first of all, just the two days before ASV, Theodora texted me. She said, make sure you go to Paul's talk. He's going to explain where the furin cleavage site came from. And so with that, I said, well, I would have gone anyway. So he gave a talk about that. And the reprint is out. And so I wanted to get him on to talk about it. The name of it, it's on BioArchive, so it's just under review now, I presume. The name is Genetic Innovation in Coronaviruses Driven by a Viral Nuclease.

4:09And before we get into the science, Paul, you've got a bunch of authors on here. Can you tell us something about them? Yeah. So first author is Christopher Bianco, and he's responsible for the bulk of the work in the paper. He's a postdoc, he's been with me for a number of years. Before that, he was a graduate student with Ian Moore, but a terrific young scientist. Others on the paper, Alex Stable, He's our resident bioinformatics expert. Manevel Loda, Miranda Aldis, Michael Tartel, they all contributed in important but less large ways to the paper. And, of course, my wife and scientific partner, Theodora Hatsiwainu, we sort of do everything together.

5:02Cool. So take us back to the very beginning and talk about how this got started. So, okay, it's a little convoluted, so you may have to bear with me for a minute. We like convoluted. We love convoluted.

5:23So basically, towards the tail end of the pandemic, it became much easier to do genetics in coronaviruses. As you know, they have large RNA genomes and the systems for actually doing molecular biology in coronaviruses has been really quite difficult to handle. but a group came up with a method it's called CEPERF it's basically a circular polymerase chain reaction method where it's become infinitely easier to assemble recombinant coronavirus genomes as retrovirologists we are very used to carrying the entire genome of our virus on a single plasmid transfect cells get virus out it's been frivily easy to do genetics but when this method came out for studying coronaviruses, I thought, ah, this is something we can really do.

6:19We've been, historically, we're HIV retrovirologists. During the pandemic, we did a lot of work, as you know, on coronavirus neutralization. But I've been sort of looking for ways to get into really studying coronavirus molecular biology. and the advent of this tool enabling us to do recombinant work with coronavirus really spurred this. And so one of the things we thought we could do was basically to make recombinant viruses to measure the recombination rate between coronaviruses. It's something we know happens, but there aren't really good tight measurements of how frequently coronavirus genomes recombine.

7:04So at the time, SARS-CoV-2 was still a BSL-3 pathogen. So we built our first recombinant coronavirus system based on a common cold coronavirus, OC43. And the first virus we made, other than the wild-type virus and simple viruses carrying reporter genes, were basically viruses that had an inactivated GFP reporter gene, inactivated in such a way that they had frameshift mutations introduced near the amino terminus and near the carboxy terminus. So that made viruses that made pieces of GFP, but not a functional GFP. But if you co-infected cells with those two viruses, if they recombined, they'd make a virus that made a functional green GFP.

7:57So that was sort of our first piece of coronavirus genetics. We We haven't actually published that yet. And just for the general audience, GFP is green fluorescent protein, a very handy fluorescent assay. So when the GFP works, they glow. Right. Yes. It's been a while since I've been on TWIF. No, that's all right. I'm still calibrating. Continue. I have to pitch this. So you've got this frame shift. Right. So one of the things we found, actually, early on, was that if we made these two viruses that signified recombination by acquiring green fluorescence, we found that if you change the nucleotide sequence of one of the GFPs, we expected the recombination rate to go down.

8:50but actually the opposite happened it went up I probably won't get into the details of why we think that might be the case we only have hypotheses we don't have data but we figured that maybe one of the reasons might be because what's happening is that double-stranded RNA is forming and during replication the sort of strand transfer event that accompanies recombination might be facilitated if there was breathing of the duplex RNA and the virus was trying to recombine with double-stranded RNA and that would be facilitated if the duplex could breathe. Another idea was that perhaps when you have double-stranded RNA that's imperfectly matched it might be more nuclease susceptible.

9:47So we did an experiment where we tried to measure the recombination rate between these two viruses that had the ability of the duplex to breathe, so we supposed, but we also made viruses that had wild-type mutant NSP15, the hypothesis being that perhaps and nuclease was nicking the duplex and facilitating the combination. That hypothesis proved to be completely erroneous. But it gave us some of the reagents that initiated the paper that we're going to talk about today.

10:31So one of the other things that we thought would be interesting was to actually try and measure the rates of other types of evolutionary events. So there's a long history of measuring nucleotide misincorporation into viruses as they replicate. But what's been much more difficult to study are rates of insertion and deletion. Because they are, at least were thought to be, much rarer events. We didn't know much about their frequency or much about the mechanisms accompanying them. Sequencing tools until quite recently weren't sufficiently developed to measure those rates. And so because we were building these tools with coronavirus OC43, what we then built was a tool that we thought could measure these insertion and deletion rates.

11:29and the way we did it was related to the the tools that i just explained to you but um what in fact we did was to put a gfp gene into the viral genome and then just knocked it out of frame so it wouldn't make a green fluorescent protein it would it's the gfp gene is preceded by a peptide with a start codon that's out of frame with the bulk of the GFP gene. So it doesn't make green fluorescence. But if that virus acquires insertions or deletions in that particular spot and of a particular length, it would put the GFP back into frame and we get a green virus. And in so doing, we could both measure the rate at which those events happen.

12:19And then by isolating those GFP positive viruses, we could see what actually happened and perhaps learn something about the mechanism. And that's basically the genesis of figure one. And the elaboration of that experiment was to include a virus that did or didn't have NSP-15, which we'd actually made for another experiment. And basically, because we had that virus, we incorporated it into this experiment. And that's what led to the finding, the central finding that's depicted in the first figure. That is basically that having a nuclease in the viral genome greatly increases the rate at which the virus acquires insertions that put that GFP-reporter gene back into frame.

13:13Were you surprised? We were actually, yes, quite surprised. I was the the hypothesis was in fact and this had some basis in a previous publication that it would perhaps affect the rate of deletion. So when you think nuclease you think deletion you don't really think insertion. But in fact the opposite occurred. So while we don't really have enough numbers to quantify the deletion rate in this experiment, it's basically a black and white result when it comes to insertions. So we were actually very surprised. We weren't expecting it at all. But that's what kicked off this entire line of investigation, that kind of serendipitous result.

14:04So why did you choose OC43? This is one of the common cold coronaviruses, right? Yes. So it's a virus that grows very well in cell culture. BSL-2 pathogen, so we could work with it without too many restrictions. We did have access to BSL-3, and we have more access to BSL-3 now. And now we can actually work on SARS-CoV-2 at BSL-2. But such was our impatience to become coronavirus molecular biologists. And didn't have unlimited access to BSL-3, we started with OC43. It's an interesting virus in its own right, actually. And we found some other things that are quite fascinating about it, but that's for another day.

14:54So NSP-15, it was known to be a nuclease. And so what was known about it before? So this is in large part the work of Susan Weiss, who's a longtime contributor to coronavirus biology, working with Volker Thiel, I think, initially, but then subsequently on her own. The canonical role of NSP15 is basically to reduce the amount of double-stranded RNA in cells. So double-stranded RNA, as you know, is something that's sensed, causes cells to make interferon, bad for virus, sometimes bad for pathogenesis. And what was found is when you deleted or inactivate, you can't actually delete NSP15, you have to mutationally inactivate it from the virus.

15:51The virus makes more double-stranded DNA, a more aggressive innate immune response. Not much effect on replication in cell culture, but very dramatic effects on replication in animals in a way that's really dependent on their innate immune response. I should say very clearly, we don't have any problems or any reason to doubt that as the canonical role of NSP15. What we're describing here is an additional role, not anything that supplants what's previously been reported for NSP-15. There's a more recent paper from Vinny Menachery and Andrew Routh, who did some deep sequencing of SARS-CoV-2 wild type and NSP-15 mutant.

16:46They found something not very different to what we found. and in fact we used their publicly available sequence data sets in our paper, they found that NSP15 affects the deletion rate and the recombination rate. The mechanisms underlying those effects aren't really worked out, but the effect is clear. I wouldn't feel very comfortable speculating on what's going on. And I don't really have a clear idea how to explain their data, but it's clear that that happens too. So NSP15 clearly chops up viral RNA, perhaps not equally all viral RNAs, but that has somewhat pervasive effects, right? Innate immune activation, change in the balance of viral RNA deletions.

17:45Also, subgenomic to full-length RNA ratios change a little bit when you complete NSP15. So it's probably doing multiple things. I mean, this is a question I want to ask later, not now. You know, you have this canonical function and something new you discovered, but I want to wait a bit, okay? But I will because it's an obvious question, right? Anyway, Kathy, did you want to ask something, your purple question there? Oh, well, yeah. So you've described why you use the NSP-15 mutant. It was sort of something that you already had. But would there have been – and you were kind of surprised by the result.

18:26If it had not turned out to be that NSP-15 was having the effect that it does, did you have another candidate in mind like next in line we will try some other gene or are there other possible candidates actually no for the for the phenotype we found we we weren't even thinking about that as something to look for right we chris as i recall said well well, let's put this NSP15 mutant in because there's this paper saying it affects the amount of deletion and recombination. Let's see what effect it has on our assay. And you had it already. Yeah. And not only does it affect deletion, but it affects insertion, and that's what was surprising.

19:15We had no reason to think that it would affect insertions. So these experiments were first done using GFP, visual GFP, and you find that NSP-15 knocks down the amount of insertion. Then you switch to deep sequencing. Can you describe that process and why you did that? Right. So, I'm perhaps not quite sure in what order to say these things because the way they're presented in the paper isn't exactly the same as our thought process as we went through. That's usually the case. I want to hear what you were thinking, yeah. So, you know, in this GFP virus experiment, it's actually incredibly laborious to go from one green cell to isolate a GFP-expressing virus from that because you have a well that has 100 different viral species in it and you have to go through multiple rounds of limiting dilution to actually isolate the GFP plus virus and sequence it and characterize what actually has gone on.

20:31That was the real motivation for switching to the deep sequencing method. But before doing the deep sequencing, we did have a piece of information that when we sequenced these half dozen or so viruses that had inserts in them, the sequences of the inserts matched the viral genome. Five out of six of the inserts. For one out of six, 15 out of the 20 nucleotides matches the viral genome, and then there's another six nucleotides. Sorry. Another four nucleotides that match somewhere else in the viral genome. But the four could be a match by chance. We'll come on to all those calculations a bit later.

21:24But before doing the deep sequencing, we kind of knew that the inserts that this virus was picking up came from the viral genome. And we also knew that they happened when NSP-15 was present.

21:44But Chris said, you know, we've kind of had enough of isolating green viruses. Couldn't we do this a bit more efficiently if we just sequenced the hell out of a virus stock? And I said, yeah, OK, well, give it a try. So he grew up two virus stocks, wild type virus, exactly the same viruses that we've been using with GFP reporters and wild type and mutant virus stocks. generated virus populations of rather normal size that you generate in cell culture. This virus grows really quite well, up to 10 to the 9 or so black forming units per mil. And so he took, I think, about 10 mils of virus, concentrated it to one mil and took a fraction of that and sequenced it.

22:39but sequenced it at a depth that's kind of unprecedented. People look to get coverage of the entire viral genome. And, you know, if you get 10x to 100x coverage of your viral genome, you can be pretty confident that that's the sequence of the viral genome. What we want to do is characterize all of the variants or as many of the variances that we could find in that population. so we we um put it on a sequencing chip that would give hundreds of millions of reads so basically we sequenced the virus at a depth um such that each each individual nucleotide in the genome was sequenced um at least a million times right so the total amount of sequencing data we get from each virus stock is i think it's approximately equivalent to 30 human genomes um or 3 million viral genomes okay so when you're looking for low frequency events right which we could do with the gfp virus because you can put you can put 100 million viruses on a monolayer of cells and pick out those individual cells or wells that had a green virus that's you know you don't get to pick and choose what you sequence you just have to sequence by brute force and that and so that's what we did so essentially essentially what's happening with deep sequencing is what you want to do is sequence every single virus in your sample.

24:21And the way you accomplish that is to sequence the whole sample a million times, basically. Not exactly. You want to try, at least know what the population size of the virus is in your sample and have that somewhat coherent to the amount of sequencing you're doing. Right. So if you have a tube with a thousand virus particles in it, absolutely pointless sequencing that hundreds of millions of times because you just get the same sequence over and over again. Right. We sort of calibrated it so that we would be sequencing. A few percent ish of the viruses in the sample. So, but millions of viruses in a sample that had tens of millions, sorry, millions of sequences in a sample that had tens of millions of viruses.

25:13So we weren't going to be saturating. Right. You're limited by, I mean, millions of sequences. Sequence reads is a heck of a lot of sequencing and you're limited by the technology to a certain extent. But the goal is to get as deeply into the individual viral particles as possible. Effectively, we have 3 million-ish viral sequences. Right. They're not 3 million genomes completely sequenced. It's the equivalent of unlinked breeds from 3 million viruses. And the other thing to bear in mind is people don't usually sequence this deep because what people are mostly looking for are point mutations. And the error rate of sequencing when you're considering point mutations is about 10 to the minus 4-ish.

26:03So it's kind of pointless to go much deeper than that because you're just sequencing the error rate of your technique. But of course, that error rate doesn't apply to insertion mutations, right? The error rates there are very different, and perhaps we'll get onto the sources of artifacts later. But bottom line is we could sequence these two virus populations and we saw basically a result that confirmed what we found with the GFP viruses. You see many, many more insertions in the viral genome when you have NSP15 present. And when you inspect those insertions, and now it's all sort of bioinformatics, you find that a large fraction of those inserts come themselves from the viral genome.

26:57so at this point i have to sort of mention how do you how are you sure that it comes from the viral genome um first thing you have to do is constrain yourself by length because um say inserts of three four five six nucleotides random sequence because of those length occur by chance in the viral genome, we kind of, while we document them, we kind of ignore them for the process of analysis. We can't be sure about where they come from. The sort of the cutoff where you begin to get confident about the origin of your inserts is around 11 nucleotides. So we did a fairly complicated analysis that shows that if you have a sequence of 11 nucleotides and it perfectly matches the viral genome, the chances of that occurring by chance are less than 5%.

27:54So we basically constructed a slightly more complicated algorithm than that to confine our analysis to inserts of 11 nucleotides or greater, and a huge fraction, probably 90 % or more, have a perfect or near-perfect match to sequences in the viral genome. And they are, in the case of OC43, 87 times more abundant in the wild-type virus compared to the NSP15 mutant virus. And they're mostly from the plus strand also, right? Right. They are. So the implications for that are a little complicated. So we know that plus strands are more abundant in the viral genome, but it also suggests that they're incorporated during minor strand synthesis, right?

28:58Because, sorry, getting a little ahead of ourselves. If we met, we don't know exactly how they are inserted, right? Whether it's a ligation process, which is possible but unlikely, or a sort of a recombination-like process where the virus is copying, making a copy of the viral genome, and then jumping to a short oligonucleotide, which we suspect is the product of NSP15 cleavage of the viral genome, copying that oligonucleotide and then going back to copying the genome. That only makes sense for the inserts to be of plus strand origin if it happens during minus strand synthesis. Because during minus strand synthesis, you're copying a plus strand.

29:53So to get a plus strand insert, you have to copy another plus strand. If it was the other way around, it would look like you had an excess of minus strand inserts. I wonder if at this point it might be good, at least my understanding is that the NSP15 digestion of double-stranded RNA that you're talking about is predominantly digesting up viral double-stranded RNA that's being produced. so I actually think that NSP15 is reducing double-stranded RNA indirectly actually by digesting single-stranded RNA just reducing the amount of RNA that's available to make double-strand but the net result is that then there's not as much double-stranded or it could be double-stranded RNA that's breathing typically nucleases of this type cutting, unpaired...

30:55So this is not a double-stranded RNA nuclease, right? We know this. Not really, no. Okay. But then, I guess I was trying to make the point that you're talking about digestion of viral sequences as opposed to host cellular sequences of RNA. Yes. Okay. Yes. So we don't know for sure, but we think this is probably all going on inside the double membrane vesicle in which coronavirus replication is occurring and from which host RNAs would be largely but perhaps not entirely excluded. So next you wanted to go to SARS-CoV-2, right? So tell us how you did that. Okay. There's one other thing we noticed about the inserts before we jump to SARS-CoV-2.

31:50And that is, if you look at the sequences of the ends of the inserts, right? Actually, look at the sequences of the inserts. There's two things of interest. First thing is that the inserts come from all over the viral genome, and they can end up at kind of any place in the viral genome, okay? Okay, so if you plot the distance, because when you confine yourself to 11 nucleotides, you can know what the destination of the insertion is, but you can also know what the origin of the insertion is. And if you calculate those distances, it's as if they're random positions on the viral genome. It's an unlinked process, the generation of the insert and the insertion of the insert.

32:39The other thing about the inserts is if you look at them in their origin context, so from where the inserts come, we know something about NSP15. It likes to cut RNA 3' to U nucleotides. And if you look at the inserts in their origin configuration, they are highly enriched to having U at the ends. So they look like they're the products of NSP-4D digestion, and they're inserted quasi-randomly around the genome. So then we knew all this before we even began to think about SARS-CoV-2. So we first looked at this publicly available sequence data set from the Ralph and Minituri labs. they used a somewhat different method for acquiring their sequences but they're the only group that had really published anything that was remotely deep enough for us to do this analysis of inserts it still wasn't as deep as we had gone but it it was sufficiently deep that we could we could see a difference between their wild type virus and nsp15 virus in terms of the the number of inserts.

33:57It wasn't quite deep enough to do the types of analysis that I just explained about the nucleotides at the end of the inserts. So we generated another sequencing data set on SARS-CoV-2, just wild-type virus, this time sequencing even deeper. So we did a billion reads from SARS-CoV-2, equivalent to about 40 million viral genomes. And when you look at those inserts, you find essentially the same characteristics. The inserts look like they are the products of NSP15 digestion, the distances between the insert origin and destination, are like two randomly chosen points on the viral genome. and then when we looked at both sequencing data sets again this goes back to the the origin and destination of those inserts we found that a subset of the inserts right not only had one origin but actually looked like two oligonucleotides from two different parts of the genome had come together and been inserted into one spot in either the OC43 or the SARS-CoV-2 genome.

35:22So the potential for generating genetic diversity is quite remarkable in that respect. So there are some differences between SARS-CoV-2 and OC43. The length of the inserts tends to be a little shorter. and the number of inserts is a little less. We don't really know why that's the case. Perhaps just differences in the inherent properties of the two enzymes, the amount of enzyme, its activity. But there's this subtle difference. But in the important criteria, the fraction, the distance between inserts and their origin and destination, and the fact that they're NSP15 substrates is very, very clear in both viruses.

36:21And in both viruses, you see what seems to me, having been told for years that inserts are rare, that this is not rare. This seems to be happening all the time, right? Right. So rare is a relative term. Of course. I've learned to use those words carefully because, you know, one in a million is rare. But if you have 10 million, it's common. Right. So when we talk about these, we try to use numbers. That's not in any way intended to be a reproach. But basically, the insertion frequency is you get a few inserts per thousand genomes. So if you have a thousand virions, you're going to have a handful, probably around five or ten genomes that have an insert.

37:19That's less rare than I expected. Right.

37:24I was going to say, and compare that to the point mutations that you would see in that same number of genomes. Right. So the point mutations would be a bit more frequent, but not massively more frequent. So why do you think over time we were misled to think that insertions or deletions are rare? Right. So they are far more likely to be deleterious or even lethal to the virus. So the vast majority of these inserts will be purged when low multiplicity infection happens. Which means you won't isolate them then. Yeah. Yeah. Yeah. So we... Sorry. You must know what fraction are deleterious just by looking at the sequence, right?

38:19So we don't exactly. What I can say is that we should qualify this because we're doing this experiment under pretty defined conditions. So we're starting with a cloned viral genome. We're transfecting cells, have a very small founder population of virus, and we expand it over a small number of days with no immune selection, no other selection other than replication competence to a population of 10 to the 9-ish. right remember though that um in an individual coronavirus infected cell right there's not one genome there's many genomes okay so these defect if if the if say 10 out of a thousand genomes have a lethal insertion in them then there's many many complementing wild type genomes that would maintain this in the population so when we sequence we're getting you know provided they're not present at extremely high frequency lethal mutants can be carried in a in a viral population and probably a newly generated all the time so when when we try to answer vincent's question kind of just by counting what fraction of the inserts are multiples of three and don't have a stop code on right because they would clearly be protein truncating, only about a third of them are in frame.

39:53So most of them are probably lethal mutants. They're not particularly enriched in non-essential genes. So this really looks like if you did a random mutagenesis experiment and then sampled right away, this is probably what you get. There's not really been no evidence actually of any selection pressure having been applied to these mutants under these experimental conditions. So it's probably a reasonably good estimate of the mutation frequency, not an estimate of what fraction actually would go on to survive. Of course, the inserts, even if they're in frame, they may be deleterious. Absolutely.

40:38Right. A large fraction of those would certainly be deleterious. But this certainly represents a process that is generating a lot more insertion diversity, if you will, than naive folks like me might have expected. Right. So this is the food on which evolutionary pressure feeds, right? Right. So it probably is evolutionarily selected for. So then the NSP-15, it's got the immune suppression function. is probably also have this function involved in evolution, right? Probably, yes. I've spent some time thinking about this, but I kind of tie myself up in knots in thinking about how much foresight can we endow viruses with in engineering their own evolutionary rates, right?

41:34Or how often do you have to have a selection pressure applied that selects four insertion mutants for a virus to maintain the ability to generate insertion mutants for that to be maintained. Yeah. Well, the problem here is that you make an active site amino acid change to knock out the nuclease activity and you're going to affect both phenotypes. You can't separate the two. Right. Absolutely. So in that case, it becomes rather difficult to say that's the activity that evolution is selecting for. Right. Because you select one, you're going to select the other. So, I mean, you could do all the thought experiments, as you say, as you want, and you're never going to get an answer.

42:19And this is kind of a longstanding problem in evolutionary biology. How much utility does a gene or protein have to provide in order to be selected for? And I think just intuitively, I mean, helping immune evasion is a huge benefit to the virus already. And then this generation of extra viral diversity is maybe a bonus. Yeah. I think that is a totally reasonable way to think about this thing. I mean, you can imagine that the insertion is not as important as point mutations. Right. But the other thing is, I was wondering if you're going to do experiments where you co-infect with two different coronaviruses and see if the insertions are swapped across genomes.

43:12Yes, we are thinking about how to do those experiments in a way that doesn't get us into trouble. Yeah, exactly. But it also is informed. I'm pretty sure we could do this with a virus carrying a GFP and look for the appearance of GFP inserts in a virus that doesn't have GFP or something. Well, and you could do it with two other, like OC43 and one of the other common cold coronaviruses that people are not so wigged out about. Yeah. I think experiment one is to do it with OC43 and another OC43 carrying an innocuous gene. Yeah. It actually gets a little tricky when you start mixing coronaviruses.

43:55They are quite good at super infection exclusion. It might be that the first one in there just kicks off enough of an innate response to keep the other one away or sucks up all of an essential cofactor that stops the other. Because we've been trying for unrelated reasons to generate double infected cells and it's more difficult than you might imagine. I wanted to ask you about this experiment where you do ultra-deep sequencing on libraries from PCR amplicons. Why did you do that? So everything we've talked about so far is with purified virus, virus grown in cell culture. We can isolate the virus particles.

44:43We can put a lot of stuff into the sequencing reaction. but we also wanted to know whether the same process happens when you for example have infected animals or infected humans right and there when you harvest virus from a person or an animal most of the stuff that you get out is not viral RNA and so the population size that you need to generate in order to get the sequencing depth to do this type of analysis becomes much more challenging. And you kind of need to do a PCR step to amplify that nucleic acid up to a level that you can sequence. And that comes with a set of other problems. So interestingly, you did a control.

45:43to sample with two RNA molecules. Can you explain what that does? Right. So the issue is when you start PCR amplifying things, you've got polymerases copying nucleic acids, and they have a mutation rate of their own. They, in principle, can also generate insertions. Obviously, there's no NSP15 there, so we're not worried about that. But it is possible for all kinds of jiggery, pokery, polymerases jumping here at low frequencies, right? But remember, we're looking for low frequency rearrangements to occur. We reasoned that if there was variation in a starting population of, let's say we took a sample from a mouse lung, which is one of the things we did that had say a million RNA molecules in and we wanted to find the frequency with which there were inserts in those million molecules we still have to PCR amplify how do we distinguish what was there in the original sample with what we accidentally introduced when we PCR amplified it So the way we addressed that was to go to our purified virus, where we kind of know what the insertion mutant frequency is, and reason that if we just dilute those 10 to the 9 molecules down to one or two molecules, all that variation in there will be erased.

47:24Plus, by PCR amplifying those one or two, the two is a calculated number. we can't know the difference between one, two, or three molecules in a PCR reaction. We know there's at least one because we've got a PCR product, but we don't know if there's one, two, three, or maybe even four. But because our insertion mutant frequency is sort of in the one in the hundreds, right, if we go down to those numbers, we shouldn't find insertion mutants when we go down to that level. So we basically did side-by-side experiment. PCR amplify are complex and complex 10 to the 9 molecules and PCR amplify about two molecules.

48:10And what you find is when you amplify those two molecules, you do in fact find some rearrangements, right? But most, actually all of those rearrangements come from within the sequence you're amplifying, right? So if we're looking for real insertions, we look for insertions into the PCR amplicon that come from viral genome regions that are outside the bit you're amplifying. Right, because your PCR is only amplifying a piece of the viral genome. Right, exactly. And so when we do that, all the artifacts that you generate in your amplicon melt away. The insertions don't come from elsewhere. where the artifactual insertions come from where you're PCR amplifying.

48:59All right. So then you use that to look at virus from mice and humans, right? Yeah. And what did you find there?

49:11Basically, same thing. Lots of insertion mutations. It doesn't really matter the context where you're getting the viral genomes from. It does become a little more difficult because when we're harvesting from a mouse, we know the number of viral RNA molecules in the samples that we generate. But then when we start looking at sequence databases, so SARS-CoV-2 sequences that have been deposited in sequence read archives, you're basically pooling data from many individual people. You don't have any real idea what the viral population size is in those individuals. but by applying those criteria that I've just talked about and then looking at the inserts, looking for the same biases that we found associated with NSP-15 driven mutagenesis in our cell culture experiments, you can find the same things both in the mouse and in the human.

50:20So it's not something that's just happening in cell culture. And of course, you should do this in bats as well, right?

50:31reviewer number three sitting over here um yeah we would like to do that but the logistical challenges are even more more of a problem and we would not have really any idea of the population size in a bat guano sample there's no it could be measured it could be there are no databases from bats existing? Not of the depth that we would need to do this type of analysis. But obviously, we're going to get to why that's important, but that would be something interesting to do, right? Right. But to be frank, though, given that we found this process occurs equivalently in OC43 and SARS-CoV-2 in mice and in humans, I'm not particularly worried that it doesn't happen everywhere.

51:23Neither am I, but you know you get pushback unless you show it. Because we're going to make some claims about this and that you can imagine anyway. So let's talk about, so then you find furin cleavage sites inserted. And did they just jump out at you or did you look for them? No, we looked for them. So, you know, we haven't been living under a rock. We're not unaware of the controversy that's been associated with this fur and cleavage site. And, you know, the first thing that I've been thinking about this for a long time, actually, because, you know, it's been a kind of an argument that's largely been dominated by people just insulting each other.

52:17Yes. How can you be such an idiot to think that that arose naturally? How can you be such an idiot to think that it didn't arise naturally? And so in trying to sort of bring some quantitation, some rational basis for arguing rather than just having a go at each other, the first piece of math I did was what are the chances of a furin cleavage site just appearing by chance? Can I just jump in with a quick backgrounder for people who might have been living under a rock, as you say? So the significance of the furin cleavage site is this is an enzyme cleavage site that is found in SARS-CoV-2 that is not found in other closely related coronaviruses.

53:08And a major argument of the folks claiming that this originated in a lab is you have to make this significant change to insert this furin cleavage site that had to have been done in a lab. You couldn't have had that arise naturally with point mutations or what have you. And so that could have arisen if you got an insertion, but then there's this whole thing about, well, but insertions are so darn rare, how could that happen? So this is what we're getting to now. Yeah. So I apologize. I guess I assumed your listeners would be on top of this. Most of our listeners are down with this, but we always like to be sure.

53:48Yeah, fair enough. Fair enough. So the other point, which is sort of amplifies what you said, is when people look at an alignment of sorbicoviruses that include SARS-CoV-2 and this furin cleavage site sticks out like a sore thumb, right? It's a little bit of a raw slash test, the phrase I used in my ASV talk, right? So if you're conspiratorially minded, you think that all these other Salbeck viruses don't have it, looks like it's engineered. But on the other hand, if you're not inclined to think in that way, you think, well, viruses, they make insertion mutations. They're pretty rare, but the population size is pretty large, so maybe it's not that surprising.

54:40And if there's a selective advantage. Right. So totally reasonable to look at that Rosash test and think something else. Basically, what this paper does is bring some numbers to this debate. And so the first thing to, not the first thing, but one thing to bear in mind is that the generation of a furin cleavage site is not difficult, right? So the minimally, actually, this isn't a very good guide to a furin cleavage site, but canonically, minimally, all you need is two arginines, R-X-X-R. That's thought to be a requirement for a furin cleavage site. Actually, you can get furin cleavage with only one if the other amino acids cooperate.

55:36And not all RXXRs are, in fact, furin cleavage sites. Where XX is any amino acid. Any amino acid. Right. But any amino acid isn't exactly right. It's EFRI. Right. There's some scoring algorithms that help you predict. But I don't, you know, without having tested them exhaustively. The point is, this is not some massive engineering challenge. So you can actually do math and figure out how often an RxxR motif would appear in random sequence, right? And you can know there's 64 codons and six of them encode arginine and just multiply those two fractions together. And about 1 % of random 12-base sequences include RxxR motifs or predicted for in-clearance sites.

56:31So you can use numbers like that, coupled with the fact there are approximately 30 ,000 different places in the coronavirus genome where an assertion mutation can arise. we can know how often insertions arise from the measurements that we've made and it is possible to to do a calculation for how frequently on based on purely theoretical grounds it would be for a furin cleavage site to occur at one particular position in the coronavirus genome and the number is not shockingly large when you consider the population size of coronaviruses. And the number is actually somewhat consistent with what we find when we deeply sequence these coronavirus populations.

57:27If you sequence, you know, several million viral genomes, you will, by chance, find the furing cleavage site approximately at the spot where SARS-CoV-2 differs from other sarbecovirus ancestors or other sarbecovirus cousins, let's put it that way. And so just by looking at our deep sequencing data sets, which don't sample the entire population, we found a couple of new furin cleavage sites precisely at the place where the existing furin cleavage site exists. So if you think furin cleavage site is lightning striking in one particular spot, we've got it struck there at least three times before looking any more closely.

58:27But the bottom line is, yes, furin cleavage sites are not difficult to make. Insertions are not difficult to make. There are only about 30 ,000 different places in the genome where insertions can occur. And so just by chance, it's sort of inescapable that in viral populations of a particular size or greater, that they contain viruses with new furin cleavage sites. at that spot. When we go back to our SARS-CoV-2 viral stock and sequence even deeper, and the way we do this is to not just put the viral RNA on the sequencer, but do a set of PCR reactions where we're putting multiple aliquots of 10 to the 8 or 10 to the 9 RNA molecules, PCR those, and then just focusing on that region of the spike protein, and then sequence those, we find many more furing-cleving sites.

59:32And actually, the ones in the paper are very likely an underestimate because we imposed on those sites the requirement that we found a perfect match in the viral genome. And because of the steps we're going to rigorously exclude PCR artifacts, we absolutely think we're also probably excluding some naturally occurring furin cleavage sites. We can't be sure that they're not PCR artifacts, but we strongly suspect that they're way more than we report in the paper. So basically you show that acquiring a furin cleavage site is not rate limiting. No. So what does this do for the origins? Does it tell us how SARS-CoV-2 FCS arose, or does it just generally remove it as evidence for engineering, in your opinion?

1:00:33so I think the latter. You know, any individual event, right, must by definition be extraordinarily rare. But there are many different ways to generate a furium cleavage site. And it happens with such frequency in unselected viral populations that we shouldn't be surprised to find a furin cleavage site there, provided it endows the virus with a selective advantage. Clearly, if you do the math, right, in a sample of, you know, a sample of the size that you would find in a swab from a single infected individual, you would expect, based on these findings, to find multiple viruses with new furin cleavage sites there.

1:01:43So the only thing that's stopping them from rising to prominence is whether there's a selective pressure that enriches for them. Right. Right. Which would explain why so many other coronaviruses don't have that because they don't need it to complete their life cycle. Right. So there's another point we need to bear in mind here, though, that it's not always advantageous for SARS-CoV-2 to have a furium cleavage site. I can show you many cell culture experiments where the virus was selected to lose it. There's very likely sort of a balancing act between the stability of the spike trimer and its fusogenicity.

1:02:26And whether it's more important to have spike stability or more important to have fusogenicity could absolutely be on a knife edge in terms of what's advantageous. Now, if you look for a few iron cleavage sites arising in, say, OC43, would you expect to see the same thing? So we haven't specifically looked at that particular spot in the OC43 spike protein, but they occur randomly across the OC43 genome in our sequencing data set at actually somewhat higher frequency than in SARS-CoV-2 because OC43 tends to generate slightly longer inserts.

1:03:23So we haven't done the specific analysis you asked for, but there's no reason to think it would be really any different. Right. So Eddie Holmes has always argued that the absence of a furin cleavage site in the known Sarbeco viruses other than SARS-CoV-2 is a sampling artifact. So do your results bear on that? It depends what you mean by a sampling artifact. He means we haven't sampled enough. We only have a few hundred SARB echoes. Yes. I mean, and they're all consensus sequences as well. Each one of those hundred is probably generated from a, probably from an Illumina sequencing data set that has a depth of, way less than the depth that we analyze.

1:04:15So that their consensus sequences is one thing. Um, now if we, if we searched hard enough, could we find a sarbeco virus in a particular host for which having a furin cleavage site is advantageous? And so that's the dominant species. Absolutely. That's possible. I don't know enough about what the environment of a bat gut is, um, what the immune selective pressure, what the advantage of having a stable spike trimer versus one that's more fusogenic, how that plays out in bats. Or in raccoon dogs or in whatever other species you choose to analyze, yeah. But yeah, I am pretty sure that if you took those hundreds of sarbecoviruses in bats and sequenced all of them at the depth that we have sequenced, you would find furin cleavage sites in all of them.

1:05:27Okay. I don't think there's any doubt about that. Right. Just from the, if you make the minimal assumption that the mutation insertion frequency that we find in OC43 and SARS-CoV-2 is basically the same in that SARS-CoV-2. viruses. So the reason this is significant is because the argument has been from some people, the FCS couldn't have arisen naturally, and now you have said it can, and that's the bottom line. And if you put it another way, as you said at the end of your ASV talk, it's because those people are saying it hasn't been found when you look at hundreds of sarbecovirus genomes, but come back when you've looked at a billion genomes and see what you yeah so so it's a it's important to recognize what those people don't say when they say it's never been found in a hundred 800 salbacoviruses that that statement is is loaded with the implication that bat salbacoviruses don't or can't generate furin cleavage sites right right I think, you know, without actually sequencing bat sarbecoviruses, I think we've falsified that premise.

1:06:56That doesn't negate the likelihood that having a furin cleavage site is probably negative in the majority of circumstances in bat coronaviruses. right but you know we we don't have to speculate for very long to think that um you know a transmission event from a bat to a raccoon dog minimal circulation in some intermediate species or even in humans that that mutant that generated a furin cleavage site and it doesn't really matter what the furin cleavage site what the precise sequence of the furin cleavage site was which Whichever sequence it was, it was going to be a vanishingly rare individual event, right?

1:07:49Yeah. But if it has a selective advantage, it will rise to the top. The other thing we should bear in mind is every dominant variant of SARS-CoV-2, Delta, Omicron, that came to be virtually every sequence in the world at some point subsequently, began its existence as one molecule of RNA in one person's lung at one time point. And that single event, vanishingly improbable as a single event, seeded an entire global spread of an individual virus. so you know arguments based on the probability of something happening being extremely remote just don't really hold any water because the there's every everything that happens in biology is extremely rare so at the end of your asv talk which i had a chance to listen to the recording of or watch the recording of, you said, human intuition generally fails in accounting for phenomena impacted by large populations or long periods of time.

1:09:09And so, and that rare events like one in 10 million become common when large populations of greater than that are sampled. So, and that's the amount of virus that would be in one individual. Yeah, absolutely. it. So I never thought the urine cleavage site was the best evidence for lab, but many people thought so. And the reason is that no one had something to put it in that was very close to SARS-CoV-2. But I do think this is like a nail in the coffin now because I think this removes it completely. But I'm really interested when this is published to see how people... So Richard Ebright is incredibly negative on the FCS.

1:09:55He says it's clearly a mark of lab engineering. I want to see what his response is. Right. So, obviously, so I've experienced some of the pushback from the lab leak fraternity who are still pretty active on social media. And basically what they say is that, okay, so you've shown that furin cleavage sites can arise naturally. We never doubted that. But what you haven't shown is that furin cleavage site arose naturally. This is classic denialist strategy. Oh, no, we already agreed to that. What we're now going to fall back to is, you know, you can't disprove what we're arguing, which of course you can't because you can't prove a negative.

1:10:49Right. But as I've just argued, any individual event in biology is impossibly rare. This is a kind of cool example. There's a paper by Harrison and Sachs in PNAS, and it argues about the lab origin of SARS-CoV-2. and they talk about the furin cleavage site and how it's not found in hundreds of sarbecoviruses and oh look at the sequences of these furin cleavage sites that you find in the epithelial sodium transporters that are being worked on at unc um yes right and could have been a top candidate for choice for insertion into into SARS-CoV-2 um they kind of gloss over the fact that half of the sequence that they identify was already present in other sarbecoviruses by the by but one of their favorite candidates is the epithelial sodium transporter furin cleavage site from the mouse epithelial sodium transporter and if you look at the amino acid sequence of that furin cleavage site it spells out something very interesting it spells out SARS-AS

1:12:12what's the probability of that happening yeah vanished it's impossible that that happened but it happened years ago i saw a bioinformatics analysis of of searching for elvis in the genome and they determined that elvis is actually quite commonly distributed so so some of your insertions are from the host right right maybe so it becomes so we have a category of insertions that are difficult to classify right so i i talked at length about how you need 11 nucleotides to um to be pretty confident that it came from the viral genome now when you go to the human genome that length becomes quite a bit longer right so we do have inserts of 11 nucleotides or more that match the human genome but don't match the sarbecovirus genome we're still actually working on those a little bit but it's possible to have chance matches to the human genome with sequences that are quite that are a bit longer sort of 15 16 or so basically how long do you need to make a pcr primer in order to only amplify one bit of human dna so i i you know in the preprint i don't want to be necessarily held to the the frequency that that that's the number of inserts that come from the human genome what we suspect there are there's a large category of inserts where we're it's ambiguous where they come from and what we suspect is happening there is that they're composed of inserts of less than 11 nucleotides that are actually chimeras probably come from the viral genome and are inserted inserted as chimeras but it's difficult it's nearly impossible to prove that statistically because each component of the chimera could occur by chance multiple places so there will always be some ambiguity nevertheless it is absolutely possible that a small fraction less than 1 % come from the human genome because I think some people have published on this already saying that there are some bits that look like human sequences right, so the for example the 12 nucleotides that compose the furin cleavage site.

1:14:58There are multiple matches to that sequence in the human genome, but there would be matches by chance as well. That's the issue there. So other viruses encode nucleases, right? Influenza, bunya viruses. Do you think they might contribute to insertions as well? absolutely possible um we we're thinking about hiv actually which carries a nuclease right along with its polymerase and chops up the rna into oligonucleotides right after it's reverse transcribed rnash right yeah so that might be our next uh port of call in terms of uh trying to extend this to other viral systems, but yes. And perhaps, so, you know, getting into the realms of speculation, it might be that chopping up the genome into oligonucleotides creates a sort of a freely diffusible pool of sequences that can more easily invade and elongating polymerase, as opposed to a longer viral genome that might be more difficult to insert itself into an active site.

1:16:20So, yes, you're absolutely right. It might not stop at coronaviruses. Hey, why limit a good thing to one virus? Yeah. Is there anything that we missed we should talk about, Paul? I don't think so. So I think we've given this a pretty good thrashing. Well, I wouldn't say we thrashed it. We looked at it pretty closely. Pretty closely, yes. I think the conclusions are justified from the data. How about that? Well, yeah, I think that's right. There's probably one sentence in the discussion that the LabLeak community took exception to. And it's basically where we're talking about a Furing-Cleavage site versus the Furing-Cleavage site.

1:17:11And so the final published version will probably make it clearer that we're talking about a Furing-Cleavage site rather than the Furing-Cleavage site. But if that's the worst thing about this paper, then I'm happy with that. Okay, Paul, thank you so much. Thank you. For joining us. Yeah, it was really great. Rockefeller University. I hope you get published soon. Me too. As I said, looking forward to more. All right. Very nice talking to you. Thank you. See you around. Take care. Bye-bye. Well, as you know, it won't satisfy anyone because if you've made up your mind already, you're not going to be convinced by data.

1:17:56No, right. You can't reason somebody out of a position they didn't reason themselves into. Let's do a couple of emails and then we can do our picks.

1:18:10I'll take the first one because it's short. Anthony writes, is saying I'm making this up now forbidden on Twiv? So just to give you context, Rich likes to say I'm making this up now when he's not sure about something, right? And someone wrote and said he really shouldn't say that because, you know, scientists have enough. With certain people, scientists are already not trusted. So and so I talked about it. And so Anthony writes, fabricating, receiving, question mark, an idea peg is the shared sacred spring of both science and art. A scientist searches reality for a whole that fits the peg. The artist uses her talent and their talent and skill to craft the setting for the inspiration.

1:18:58I would not say it's forbidden. I would just say, if you want to use the expression, you just say, I'm just not sure the facts here. I'm just not sure of the situation. That's all. Or I'm hypothesizing. I'm hypothesizing. You can say, Rich can say I'm making it up and then you can say I'm hypothesizing. Yeah, I'm postulating that this is maybe how it goes. Yeah. Alan, can you take the next one? Okay, so this is a long email that I will partially summarize. I'll give the intro and then I think I get the gist of it. Dallas writes, Dear wise men, presumably, ladies and gentlemen, you happen to read a letter that pushed the idea of using legumes and more vegan diets to avoid the massive amounts of energy required to synthesize fertilizers.

1:19:48It sounds good in theory as long as you don't note that humanity is already using most of the best land area and fresh water supplies to feed ourselves on this finite planet. That makes the efficiency of converting sunlight into protein very important, which puts it in my thermodynamic wheelhouse. The probable activists who prepared the letter that ignored the land area probably wouldn't believe someone like me saying he is missing the 800-pound gorilla in the court of reality. So I asked Google AI. And so there's a long discussion of the amount of energy required for the Haber-Bosch process to fix nitrogen, which is used to generate, I think, half of the nitrogen that is now consumed by people in crops.

1:20:32and gives a table, table one here, calculating protein yield per year of sun and pointing out that if you use solar panels and electrolysis for the Haber-Bosch process, you generate more protein yield for an amount of sun versus natural biological nitrogen fixation. And then goes into some depth explaining why the solar industrial pathway wins. Now, the bottom line, because solar panels capture sunlight with far greater efficiency and because artificial photosynthesis systems coupled to electrochemical reactors do not lose energy maintaining living plant tissue, the solar-powered electrochemical route produces multifold higher yields of ammonia and subsequent protein per unit of solar radiation compared to biological systems.

1:21:29If you'd like to explore this further, I can break down the exact chemical formulas and energy calculations for both pathways or show you how synthetic biology is trying to bridge this gap. Concludes, to handle the coming 2.5 plus billion more mouths to feed as people get older, we can't let the activists disrupt progress like Greenpeace did with their anti-GMO, anti-Golden Rice campaign that killed and blinded millions of children. Over 100 Nobel Prize winners suggested that Greenpeace created crimes against humanity. And Dallas is from Huntington Beach, California. Thank you for that deep dive into a topic that I'm sure Dixon would have found fascinating.

1:22:10Well, Dixon, when he wrote that most of the land is being used already, that's exactly Dixon's wheelhouse, right? Yeah. And his argument was most of the land that we could farm is now being farmed, and the problem is that it is being farmed badly in a lot of cases. And so this, you know, the idea of increasing the efficiency with which we turn sunlight into protein certainly was something he was very interested in. Yeah, I didn't know that you could use solar in electrolysis for Haber-Bosch. You know, I thought you had to use all that heat, et cetera. But we also had a letter somewhere else about this as well.

1:22:56Who was it? Haber-Bosch process came up on an earlier episode, and I forget the context. So I made it a pick. I made it a pick. So this paper with the nitroplast, right, and nitrofixation. And then I started talking about Haber-Bosch, and we did it on TWIM. But someone said that, you know, the number I quoted, half of your nitrogen molecules come from Haber-Bosch. Someone wrote a long email, you know, suggesting that that wasn't exactly correct. And I don't know where that email was written to because it wasn't written here anyway. So we've gotten two on that already. Kathy, can you take the next one?

1:23:45James writes,

1:24:15object to the government's actions, but the reader must notice these afterthoughts. The science article is at URL, and he gives a URL which you're going to have to fix before you post it. Anyway, as a retired engineer, I have relied on TWIV since 2019 for scientifically well-vetted information on virology, immunology, epidemiology, and the like. Would you guys and gals please devote a good chunk of time on one of the shows to talk about the actions against Drs. Barrick and Gary, and other similar government actions of Project 2025-inspired defunding actions. Please identify both the politicians and scientists who are colluding in these activities and talk about the impacts on basic and applied research, as well as where or if government funding might be replaced with research and development funding from non-government organizations.

1:25:10Thank you very much for all the work you do, both in the research and development of science and technologies, and the communication of same to us life science challenge viewers. The availability of your subject matter expertise has been critical during these past six years. Jim. so to having spending a lot of time talking about the mistreatment of Bob Gary or Ralph Baric or others I think it is adequately covered in articles in science and nature and so forth so So, and then trying to identify the politicians and scientists who are colluding in these activities, I think that is going to be somewhat difficult, at least with respect to both categories, politicians and scientists.

1:26:18I mean, so, and talk about their impacts on basic and applied research and where it's going to be replaced. And to be totally pessimistic and Debbie Downer about it, I think the impacts on basic and applied research are already huge. And where or if NGOs are going to be able to replace that funding, I don't think that's possible. So one analogy I make for people is, okay, you can think about the Kennedy Center and the damage that's been done to the management and organization of the Kennedy Center. And if you were going to say, well, how long would it take to fix that in an ideal world? And you might say, oh, I don't know, maybe five years or maybe that's being too generous, maybe 10 years.

1:27:16Well, personally, I think the damage to research in not just virology or biomedical science, but everything, ecology, the climate, energy, etc., is way beyond that. And so I think we're looking at 50 or more years, and that would be if you had government funding coming in to save the day. So I think what you're asking is not possible for us to do, but maybe Vincent or Alan are less pessimistic. I am not less pessimistic. The problem with funding, with pulling the rug out from underfunding in basic research, which is what we're talking about here. I mean, this has just been slashed arbitrarily and capriciously and deeply.

1:28:22And there was the whole doge nonsense, which shed thousands of employees from the federal government, many of them in research and public health. This is a generational destruction of American science, which formerly led the world for generations. I mean, going back to World War II, Van Iver Bush designed this research enterprise and basically told us how to use it. And it's been going great and has fed, I mean, that's why America owned the 20th century was our undisputed lead in science. And that has been destroyed. And it's not, you know, you can say, well, maybe NGOs, maybe a couple of philanthropists could come in, maybe private industry could.

1:29:15None of that is going to replace the structure, the depth of expertise, the sheer quantity of funding. I mean, it is not really the issue. Yes. To say nothing of the destroyed morale. The destroyed morale. And the reason this is generational is, yeah, I mean, Elon Musk is a trillionaire. He could write a check for the entire NIH budget for the next 10 years and not feel it. that's not the problem. The problem is that the people doing the research have been jerked around enough and they're not stupid that they are going on to other things or retiring early. And the next generation of scientists, which we need to continue this, is already...

1:30:03I'm the parent of a junior in college, okay? And she's in math and computer science and she's looking at her job prospects and trying to figure out, well, do I want to go to finance or do I want to go to biotechnology? Government-funded research is not really on her radar because that's not necessarily a good thing to go into now. So you lose a whole generation of scientists, and then you're going to have to rebuild the whole thing from scratch. And I think Kathy's 50-year estimate is completely reasonable. I think that's the best we can do. We can tell you what's been cut. It's hard to predict, but as Alan said, we know areas are already suffering.

1:30:47And that assumes that in 2028, responsible management comes in and we don't do this stupid crap again for another 50 years, which I don't know how reasonable that assumption is. I feel that even if we do change, we're going to get them back again at some point in the future because now they are embedded in maybe U.S. I don't know if we can ever get rid of those. That those individuals, I would say, James, you should take a look at the article on Bob Gary. It's open access. It's in Bloomberg. It's a very long, comprehensive article, and it tells you exactly what's going on with him. And really similar things are going on with Dr.

1:31:31Barrack. There's no case against either. It is simply vindictive, as is the case with Tony Fauci. They have nothing on any of these individuals, but they will financially destroy them by having a government investigate them. They're going to have to hire many lawyers and their life will be ruined, basically. Already the Gary lab is shut down. The Barrick Lab is shut down. They don't have any more grants. And so going litigation, having a lawsuit and anyway, so they're in big trouble. And it's not going to end at them, unfortunately, because we have vindictive individuals in power right now. And, you know, prosecution shouldn't be just for vindictiveness, but that's what's happening.

1:32:22All right, let's do our picks of the week so we can end on a lighter note. Kathy, what do you have for us? Okay, let me get to my pick. I picked the Jurassic Museum of Asturias, which I was lucky enough to go to in August. Asturias is one of the geographic regions of northern Spain on the coast. and this Jurassic Museum is made in the shape of a dinosaur footprint so it has three big lobes to it and I put in the link to it from Wikipedia and so that's in English and so I was looking at the exhibits I think it was mostly in one of the lobes of the footprint and then kind of turned the corner and I should tell you that many of the dinosaur skeletons and so forth in the museum are reconstructions.

1:33:19They're not actual dinosaur bones. And so I turned the corner and there's skeleton casts of a male and female T-Rex doing something together that I've never seen before in a museum. and I looked at my friend Isabel Novea who was a virologist at the University of Toledo and that's how I know her and she's moved back to her home in Oviedo which is in Asturias and she took me to this and I looked at her and she looked at me and kind of smiled and laughed like yeah Kathy you figured it out and then there's a big mural on the wall showing them and if you go to this wikipedia site and you scroll down far enough you'll you'll see them the other thing that was even more exciting and part of the reason why this museum is where it is is that that whole region of spain is rich in dinosaur abelia i mean memorabilia i don't know what you would call it but dinosaur stuff.

1:34:26And so just a little over three kilometers away, then we drove to this site on the beach where there are dinosaur footprints. And I got to stand in dinosaur footprints. I never, I just got goosebumps again. I never in my life would have dreamed that that was possible. And as we walked there, there were all these ferns and flowers and stuff that made it just seem kind of Jurassic. And it is right on the beach. And several times a year, there's a super, super high tide and the water washes over these footprints. And so that means that there's not really anything that can be done to protect them.

1:35:12and they've lasted for thousands of years. And so you go down some steps and you're now at beach level and you can stand in the footprints. So if you go to Asturias, which has lots of things that Isabel took me to, I went to one of the places that has cave paintings. So I got to see some of those famous ancient cave paintings. But go to this dinosaur museum. I'm happy to tell you more about it, how to find the footprints. You've got to go to the beach and see them. That's very cool. Yeah. Jurassic porn. Yes. And by the way, if you ever want to stand in dinosaur footprints again without getting on a plane, well, maybe you would have to get on a plane.

1:36:02If you ever find yourself in western Massachusetts, we have some just up the road from me along the Connecticut River. there's preserved dinosaur footprints. Well, not preserved. They just happened to be there next to the Connecticut River. Right, right, right. Alan, what do you have for us? I have a book I read recently. This is nonfiction. It's called Grand Central Winter by Lee Stringer with a foreword by Kurt Vonnegut. Lee Stringer, in the mid-1980s, he was homeless and addicted to crack cocaine and huddling under a platform in Grand Central Station to have a place to stay. He had a little hovel he'd hollowed out there.

1:36:46And he was looking for something to tamp down his crack pipe with, and he found an old pencil. And then he started carrying that around behind his ear, and he found a composition book a little while later, and he happened to start writing. And then he kept writing, and he really liked it. And he essentially, over a period of years and with a couple of setbacks along the way, he wrote himself out of his addiction and into a job. And you can see why when you read this book. He is a really excellent observer and essayist. And there's just, you know, these poignant insights on life on the street. and some of it is stuff you expect, and some of it is stuff you really didn't expect.

1:37:36And it's all fascinating, and it's taking place—so this came out several years ago, and actually probably about 20 years ago now. But the events he's recounting in these essays took place in New York City in the late 80s to early 1990s. So if you were in New York City in the late 80s to early 90s, you may see a lot of familiar stuff. He was the editor of Street News for a while, if you remember that, the paper sold on the subway by homeless people. So a lot of really cool stories and very, very well written. Wow, that's cool. And they're essays, so you can read one in 15 minutes and set it aside and pick it up again.

1:38:25I read this at the gym.

1:38:30I passed a guy on the street this morning. He was sitting in front of Penn Station. He was just sitting on the ground. It's still warm out, so you can see that. He was all hunched over, and he had a sign next to him that said, I'm losing hope. That's it, just that, nothing else. And I just, I walked by and I just wanted to go back and do something, right? Because I've never seen just that kind of despair. I don't know if it was real or not. I assume it was, right? But this is, this guy. This guy lived that, right? Yeah. Yeah.

1:39:11But it's really, it's an uplifting book. Yeah, it sounds like it. Yeah. Yeah. I have two videos. I've been watching these videos on olive oil because apparently a lot of olive oil is fake. So I give you two videos which are not made by AI because a lot of them are made by AI and it drives me crazy. And so one, how to choose a real extra virgin olive oil. This is by a guy who's an MD-PhD and he says, I present you with real data. He reads papers about this. And so he has a scientific approach to it. He says the end. So a lot of these oils are like canola oil with a little bit of olive oil in it, right?

1:40:00They're mixed. And so he says, you can tell a real olive oil, the key is when you swallow it in the back of your throat, you get this spicy taste. and that's telling you that it's real. But even he says extra virgin, which is supposed to be the first pressing, you know, and all that. He said that gets faked all the time. Anyway, so that's a good one. Then there's a short one by a cook. How to tell if your olive oil is fake. You basically put a little bit in a bowl and you stick it in the fridge and if it's real olive oil, the next day it's going to be hard. And if it's not, it's going to be still liquid.

1:40:37So that seemed pretty good. In the video description, she does say this isn't a perfectly reliable test. Yeah. Apparently. Does it depend on the temperature of your refrigerator? No. Well, I'm going to try it with olive oil. Refrigerator or freezer? I forgot what you said. Okay. I don't want to play it now. But I will try it with olive oil and canola oil. I don't have any canola oil. Maybe I have to try something else. You can report back to us. Tell us the data. I tell you the data. Yes. Anyway, I thought it was funny that an MB PhD is doing the other one, too. I mean, he's got very corny jokes, but I think it's interesting.

1:41:18And I just think it's too bad if olive oil is faked. I mean, oh, my gosh. The consumerist thing about this world is getting out of control, right? I mean, I've been watching a lot of videos about the crash of the Amazon plane in Miami, right? And they had a cargo of 32 ,000 pounds of contact lenses. And I'm like, so your contact lens is just another commodity, you know, like everything else. It's kind of sad, I think. We have a listener pick from Kathy. I heard your intro on episode 1355 about the derision of the scientist Ada Yonath's work during her lifetime. He's the woman who just died, and she had solved the ribosome structure, got a Nobel Prize.

1:42:07I recently read a great book called I Told You So by Matt Kaplan. It looks at modern scientists like Katalin Carrico and her struggles in historical doctors like Inez Semmelweis revealing the professional and personal attacks they endured. It was eye-opening and well-written. Also, Kaplan makes suggestions for science as a whole to reduce the vitriol in modern research. I hope you enjoy the read. Yeah, that's what Paul was talking about. The vitriol in the fur and cleavage site. People can just calmly talk about it. They have to say you're stupid to think that. Come on. That's not what it's about.

1:42:46Just have calm, measured discussions. People, I guess, think if you're screaming that you have more of an impact. But it doesn't work that way. All right. That's TWIV 1357. You can find the show notes at microbe.tv slash TWIV. You can send your questions and comments and picks of the week to TWIV at microbe.tv. And if you enjoy these programs, we'd love your support. Microbe.tv slash contribute. Kathy Spindler is Professor Emerita at the University of Michigan in Ann Arbor. Thank you, Kathy. Thanks. This is a lot of fun. And you told me to point out that the episode number is also 1357, which is the first four odd numbers.

1:43:35Nice. Pretty cool. Alan Dove is at alandove.com, turbidblack.com. Thank you, Alan. Thank you. It's always a pleasure. I'm Vincent Racaniello. You can find me at microbe.tv. I'd like to thank the American Society for Virology and the American Society for Microbiology for their support of TWIV, Ronald Jenkes for the music, and Jolene Ramsey for the timestamps. You've been listening to This Week in Virology. Thanks for joining us. We'll be back next week. Another TWIV is viral.