a16z Podcast - All about the Coronavirus

Episode Date: January 30, 2020

This episode of 16 Minutes on the news from a16z is all about the recent coronavirus outbreak -- or rather, a new type of coronavirus called 2019-nCoV for 2019 novel coronavirus. Since it's an ongoing and fast-developing news cycle, we take a quick snapshot for where we are, what we know, and what we don't know, and discuss the vantage point of where tech comes in. Topics covered include:definition of a virus, categories of coronavirusesorigins and spreadhow this stacks up so far against SARS and MERSspeed of sequencing, implications of genomic infospeed of information sharingR0 ("r-naught"/"nought") and what it measuresdifferent ways to think about how bad a given epidemic iscurrent moves and treatmentsOur a16z guest is Judy Savitskaya on the bio team, in conversation with Sonal Chokshi.Link sources or background readings for this episode:Centers for Disease Control and Prevention (in the U.S. Department of Health and Human Services) + typesWorld Health Organization (in the United Nations) -- situation report #6, January 26Other background readings / pieces mentioned in this episode: "Scientists are moving at record speed to create new coronavirus vaccines--but they may come too late", Jon Cohen, Science (AAAS), January 27"Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China", The Lancet, January 24"Discovery of a novel coronavirus associated with the recent pneumonia outbreak in humans and its potential bat origin", bioRxiv, January 2 *note - preprint, NOT peer reviewed*"The deceptively simple number sparking coronavirus fears", Ed Yong, The Atlantic, January 28 *this appeared AFTER this episode was recorded, so sharing here as additional reading only*image: CDC

Hi everyone. Welcome to this week's episode of 16 minutes where we cover what's going on, what's in the news from our vantage point in tech. In this episode, we're going to go deep on one topic, which is the coronavirus and is a very fast developing news cycle, so we're going to take a snapshot for where we are right now. And since the show is all about teasing apart what's hype, what's real, given all the buzz and headlines out there, we're going to try to focus on what we know and what we don't know. I've tried drawing wherever possible from primary sources. So CDC reports, World Health Organization reports, etc. instead of only looking at news headlines and derivative reports. And our A6 and the expert, who I'll introduce in a moment, will be bringing in the vantage point coming from bioengineering and that aspect as well. So first of all, let me quickly summarize the news. People are referring to this outbreak as the coronavirus, but it's actually a new type of coronavirus because coronavirus is actually the general term for a more common category of viruses. And this current strain is called 2019 NCOV for 2019 N-Novel coronavirus. It's a rapidly developing situation, but as of January 26, according to the situation update on the World Health Organization website,

there's a total of 2014 confirmed cases that have been reported globally. Of these 98% were reported from China, including Hong Kong, Macau, and Taipei. 324 of 1,975 cases have been reported as severely ill, with 56 debts reported to that date. Finally, 29 confirmed cases have been reported out of China in 10 countries. And in the table that the World Health Organization provided, there's two cases listed in the U.S., but there's more. Again, this is from the six situation report, which comes out every few days, and this one came out on Sunday, January 26. That's a very high-level summary. Now let me welcome Judy Savitskaya on the A6 and Z Bio Deal team. First of all, really quickly, what is it? What is the coronavirus? Yeah, so let's discuss

what even is a virus. So a virus is basically a bunch of DNA or RNA, some sort of nucleic acid surrounded by a protein shell called the capsid of the virus. That is the entire organism. And a lot of people actually don't even call this an organism because it's not quite alive and it's not quite dead. It's something in between. There's kind of a debate in the scientific community and the philosophical community about what is a living thing. And the place where most people have come down is that you need two conditions to be alive. You need to metabolize, which means you're taking some chemicals, transforming them into other chemicals, and pulling out energy in the process and using that energy for something. And the second requirement for a living thing is to multiply, to reproduce.

So viruses really only satisfy the second condition. They don't do anything on their own. They don't, outside of a human host, they are non-living. So that's why there's such a debate. The bottom line is that they don't metabolize, but they do multiply. Exactly. So tell me now more about the coronavirus category. So the reason it's called corona is because on electron microscopy images, it actually looks like there's a little crown around the virus. The capsid for the coronavirus has these proteins on it that are spikes. And a lot of the ways that we're developing vaccines against this virus and a lot of the ways that we're identifying different types of these viruses,

is by characterizing those spike proteins. And as with most things in biology, we use Greek symbols to denote different versions of the coronavirus. So there's the alpha, the beta, the delta, and the gamma, which are kind of four of the main categories of coronavirus. A lot of the common viruses are either alpha or beta types, but SARS and mares, they're all betas. Okay, so that's kind of scientifically what it is. Now let's practically break down the symptoms. According to the CDC, the symptoms can include fever, cough, shortness of breath, or other respiratory symptoms. And they believe that at this time, that symptoms of this virus may appear in as

few as two days or as long as 14 days after exposure. And this is actually similar to what's been seen with the previous incubation period of mares viruses. And I'll get to what that is in a minute. But unlike those viruses, this particular one rarely produces obvious like runny noses or intestinal symptoms necessarily, just according to one report. And that was recently published in Lancet. So I guess the question is that all of these things, are on a continuum. It's not very discreet. Like, this is all symptoms that can describe, frankly, any common cold. Right. It's the web-empty problem. Right. Basically, you can Google it and associate yourself with anything. So the question I have is, how does this stack up against

Mears and SARS? And just really quickly to summarize, SARS was the acronym for severe acute respiratory syndrome. There was a big outbreak of it in the early 2000s. And then Mares is Middle East respiratory syndrome. And that is new as of 2012. So coronavirus has caused about 10 to 30, percent of colds, just your common colds. And those are not nearly as serious as this disease. And some of the differences between these epidemic-causing coronaviruses versus your common cold is just the severity of the infection, the likelihood that you are to die or to have really serious complications from the infection. And in all of these cases, SARS, mares, and this current coronavirus, it's because the coronavirus is infecting the lower part of the respiratory tract

versus just staying around your upper respiratory tract area. Right. So not just your just your mouth, nose, and sinuses, but by going into your lungs, so just to give a quick summary on where the fatality rates are, SARS apparently claimed about 10% of people, and Mares was much worse, claiming 30% of the people it infected. It's also interesting because I've been reading a lot of papers, but none of them are peer reviewed. In fact, one of the papers between Friday and today was already updated with V2, but the authors of the local institutes of virology, Chinese Academy of Sciences, local hospitals, and the provincial CDCs in China, within this area, supposedly analyzed full-length genome sequences from five patients at the early stage of the outbreak.

And they, A, found that almost all of those were identical to each other, so it's the same virus. And B, that about 79.5%, and again, this is the current paper still being updated, identify to SARS-coronavirus. So that's actually the most interesting thing from a bioengineering perspective about this particular epidemic is how incredibly quickly we have sequenced this virus. For past epidemics, it's taken time for us to really understand the genome and the molecular nature of a given virus that's causing an epidemic. In this case, within two weeks, people had already published draft versions of the genome sequence for this virus. And the science is happening in the sort of really live way that doesn't happen very often, where people are commenting

literally in the Gen Bank, on the Gen Bank website. Right. The CDC uploaded the entire genome of the virus from the first reported case in the United States to Gen Bank. And it's also interesting because in the age of social media, which cuts both ways, virally, it's also spreading information much faster. Coronavirus was first detected in Wuhan City in the Hubei province in China, beginning with 44 patients who had, quote, pneumonia of unknown etiology or unknown cause between New Year's Eve and the first couple of days of 2020, and then was identified as a new type of virus isolated by Chinese authorities on January 7th. And then on January 11th and 12th, the World Health Organization received detailed information from the National Health Commission

in China that the outbreak is associated with exposures in one seafood market in Wuhan City. And basically it's showing the spread of information, whereas with the SARS crisis, journalist Helen Brandswell at Stat News was commenting because she had covered the SARS crisis in 2003 that one might be tempted to say SARS start was worse. It spread faster. Not sure that's true. SARS was well underway for at least 4.5 months before the world knew there was a new virus spreading. This current coronavirus seems to have been spotted much, much sooner after its emergence. But what's really interesting is not just that it's been spotted sooner, but that the genetic information, we have is moving much faster. So can you talk to me about what that tells us and why that matters? So there's a couple of areas where you get benefits from having all this genetic information so quickly. The first is just diagnosis. So if tomorrow somebody in San Francisco was to go into an urgent care clinic and say

that they have a cold, we could really quickly identify whether or not that actually belongs to this epidemic. Another advantage we have once we have the genome sequence is that in this age of genomic medicine that we're entering, where we're actually creating vaccines that are based on genome sequences. The third implication is for figuring out treatments and also predicting some of the features of the epidemics. So we know in this example that this coronavirus looks really similar to SARS. So we can look back at the SARS epidemic, understand how quickly it spread, in which populations. And from the genomic information, you can actually see, do, for example, the spikes on the

corona, which are involved in getting into cells, do those look similar? to what the SARS spikes look like and maybe the treatments that work in this case as well. So the high level summary is that having the genomic information, which we didn't have then, when we do have now, but then I mean the SARS outbreak, which happened about 2002 to 2004, the peak was 2003, is that you can classify things much more easily figure out where it belongs, doesn't belong, kind of isolate that, that you can develop things faster based on it, although there isn't a vaccine yet, and that you can figure out treatment protocols based on the similarities and differences. Was there anything else on the connection between mares and SARS from a genomics

perspective? Yeah, so because the science is happening live, you're seeing a lot of pretty quick modifications to what people have already said. So the first paper analyzing the genome of this virus that I saw, at least, described that the spike proteins that are used to enter into lung cells are different enough between SARS and this new coronavirus that they thought it wouldn't be as bad as SARS. And then within like literally two days, I think, I saw a paper that corrected that and said that actually the protein is quite similar at the protein level. Are there any bioengineering implications of that? This goes to me that the eternal question of how DNA expresses itself practically in the complexity of disease.

That's a really good question. This mirrors everything in genomics where we thought that once we have the sequence, we'll know all the answers. And that's definitely not the case. We can know from the DNA what the difference is going to be in the protein. That's one to one. There's no guessing there. But once we know what's different about the protein, we don't yet know quite what that means for how it will behave. So we don't know if that difference means it's stronger or it's weaker or if it will infect different cells or what that's going to mean. So now let's move on to more of the details of how it spreads and the measurements of that spread. So first note is that generally the coronavirus is spread through air. They're known as zoonotic in that they originate in animals and

only sometimes leap to humans, and there's been some speculation that this one seemed to originate in bats, but it's usually indirect mechanisms. So in mares, it went from bats to camels before going to humans. And a couple of papers have commented on the similarity of this new virus to bat DNA. Like one found that it's 96%. Identical, another journal of medical virology also observed similar components, but suggested snakes. And many other people are skeptical of that. The long story short is, no one really knows, despite having some genomic information. around it. But it's interesting to note that important epidemics come from animal sources and some speculation about the reason for that is that the viruses have time to evolve in their animal hosts

and they're evolving away from what human hosts have seen before and what their immune system has been able to recognize before. So then let's talk about the spread. A lot of the articles are talking about R0 or R not. What does it measure? So R not is essentially the number of people that you would expect to get infected from any single case of infection. So if the R not is say, what does that mean? That means that you should expect that on average, five people will get sick from one single person that they come into contact with. So for every person that has the disease, five additional people get the disease. And interestingly, these are reported as ranges. So like measles has an R not of 12 to 18. Yeah. So measles is super infectious. It's known as sort of the highest or one of the highest infectiousness variables, which is why it was so important to have vaccines and herd immunity for measles specifically. And the reason that it's reported as ranges is because the they depend on the particular population and the particular moment in time. So then let's just talk about some of the facts of the spread. So of the 29 exported cases reported by the World Health Organization, 26 had a travel history

from Wuhan City and China. And then for two of the three cases that were identified in countries outside of China, one in Australia had direct contact with the confirmed case from Wuhan while in China. And one in Vietnam had no travel history, but was in contact with a confirmed case. His father had a travel history to Wuhan. So what we do know is that human-to-human transmission is occurring. The preliminary R-Not estimate that was presented at the International Health Regulations Emergency Committee was a range of 1.4 to 2.5, which relatively to the measles example

is not as crazy bad. SARS had an R-Not between 2 to 5, so that kind of puts those numbers in perspective. It might be a little premature to set these numbers just because the number of cases has still been relatively low, and they haven't had time to play themselves out. So we don't know if there's many, many cases out there who have not presented symptoms, and therefore we don't have clear stats on those people. The big question here is how bad is this epidemic? That's the question with every epidemic. And it's actually so much more complicated than that because there's a number of different variables that go into determining how bad something is going to be.

It's very tempting to put sort of a single number on how bad on a scale of one to ten, for example. Right. But it doesn't really take into account all of the nuance in each particular epidemic. So R not is calculated from the data is actually an aggregate measure. There's a lot of different variables that go into R not. So to break them all down into their individual components and then we can build them back up into R&T, there's a lot of variables that are going to matter here. One is how well does the virus transmit itself?

So if it's airborne, it's able to multiply or transmit itself substantially more easily. Right. Versus like exchanging bodily fluids and which requires a lot of specific contact. Exactly. Another piece is how is it actually getting into the cells. is it good at infecting cells? Is it good at its job, essentially? Another question is, what is the population that it's occurring inside of?

Is that population moving a lot? Are people coming into very close contact with each other? So that's also going to factor in. A particularly interesting feature of this is that it happened during Chinese New Year, which is a period of time when a lot of people in China travel. So what I think is really interesting to this, and having worked at the Gates Foundation and seen how we think about epidemics on a global scale, is that there's different, there's two orthogonal ways to think about an

epidemic, which is how much does it spread, and then how bad is it once you get it? And you're saying it's orthogonal or contradictory? Why? So there's a notion of case fatality, which is for each infection, with what likelihood will the person die from that infection or have like very serious complications? And that's actually completely orthogonal to all of the other variables that we talked about before. One is that if the virus is actually not that deadly or if it expresses itself in a person after a substantial incubation time, it might end up creating a bigger epidemic because that additional time will allow the patient to infect additional individuals. The R not for that particular virus might be substantially higher, even though its fatality rate is lower. So it's an interesting paradox that you could actually have a virus that is less bad once you get it, but is more bad on the population scale.

So you have to define the metric by which you're saying whether or not an epidemic is bad. Is it the number of people who die? Is it the number of people who are infected? Is it the extent of the spread geographically? There's a lot of different ways to think about how bad an epidemic is. Now let's just talk about treatments and concrete things that are happening right now. So just to quickly summarize what's happening, the CDC is conducting entry screenings at five major airports, Atlanta, Chicago, Los Angeles, L.A.X, New York City, JFK, and San Francisco. doctors are treating symptoms. There's no vaccine yet. The CDC has developed a real-time

reverse transcription polymerase chain reaction or R-R-T-PCR test that can diagnose this virus in respiratory and serum samples. So what that is is basically a way of seeing if the nucleic acids, if the patient sample contains the same sequence as the sequence that we know to be involved with the virus, that we know as a part of the virus. So you don't have to sequence every single patient. Okay. And then on January 24th, just a few days ago, the CDC publicly posted the assay protocol for this test. And the quote, they said, currently testing for this virus must take place at the CDC, but in the coming days and weeks, they will share these tests with domestic and international partners. And they're also growing the virus and cell culture, which is necessary

for further studies, including for additional characterization. So concretely, now from a practical technology point of view, because obviously this is a serious crisis, there's a lot to be done, a lot of different players. What are you sort of seeing as some of the things that might happen that where tech can help? It's really interesting about this moment is that because we have this increase in genomic medicine, just this past weekend, the Coalition for Epidemic Preparedness Innovations, CEPI, gave out grants to three different pharmaceutical companies of a total of $12.5 million. And they're currently engaged in a race. These companies are targeting dates for releasing their vaccine of between four to 16 weeks from now, which is just totally.

unheard of. For a company to spin up a vaccine for the SARS epidemic in 2003 would have taken months to years for them to develop the new drug and actually get it approved. In this case, two of these companies made these types of vaccines for other sequences. They're able to take what they've already built in-house for their other programs and they can just very quickly adapt it to this particular virus. All they have to do is create a drop in replacement of the sequence that they've already worked on with the sequence of this new coronavirus. Is it fair to say that it's not dissimilar to sort of engineering in terms of semiconductor and manufacturing lines that when you say drop in sequence, does that mean that it's basically a matter of using existing

scaling methods and you're just changing the actual code, the actual code, quite literally the biological code? I think that's actually a really good way of describing it. It's kind of like changing the code, but you already have the manufacturing line set up. You already know exactly how you're going to make this thing. And there may be differences, but they will be minimal compared to the differences that would have existed with a completely different medicine. having to bespoke or custom make it yourself.

Yeah, exactly. So bottom line it for me, Judy, how should we think about this news about the coronavirus still developing? Just to be very clear. And specifically, this is for NCOV 2019. We've covered through the high level of what we know and what we don't know. What would your bottom line be from the perspective of a bioengineer? So the bottom line is that we really need to think about how we are interacting with people, how we're traveling and how we are protecting ourselves from the virus.

From my perspective, as a bioengineer, we've been talking about how sequencing and synthesis of DNA is becoming faster and cheaper every single day. And this is an example of that in action. This is not something that would have been possible even a couple of years ago. So I think that what we're seeing is the beginning of how quickly and how efficiently we're going to be able to get to vaccines in the future as we continue to decrease the cost of sequencing and synthesis. This is a really interesting time because we're able to figure out the diagnostics piece, the vaccine piece, and the treatment piece. It's still in progress, but we have a huge head start. Thank you, Judy. For those of you who would like more information,

please visit www.c.c.gov slash coronavirus. I've also included the link sources for this episode in the show notes, which you can find at a6.c.com slash 16 minutes. And as a reminder, if you haven't already subscribed to this separate show in your podcast feed, please do so now, and thank you for listening.