Podcast: One Cure For All Snake Venoms Via AI

In this episode, we discuss a new anti-venom therapy designed with the assistance of AI led by a superstar team of researchers which includes a Nobel laureate! 

This podcast is sponsored by Mouser Electronics. 

Episode Notes

(2:30) - AI-designed proteins neutralise snake toxins

This episode was brought to you by Mouser, our favorite place to get electronics parts for any project, whether it be a hobby at home or a prototype for work. Click HERE to learn more about how AI is being used to expedite the drug development discovery phase! 

Hint: it’s a great primer for this episode :)

Become a founding reader of our newsletter: http://read.thenextbyte.com/

Transcript

Hey everyone, in today's episode we're talking about a safer, cheaper, and more effective alternative to conventional antivenoms. This is AI-designed proteins that are going to revolutionize the way that we treat snakebites.

What's up friends, this is The Next Byte Podcast where one gentleman and one scholar explore the secret sauce behind cool tech and make it easy to understand.

Daniel: What's up friends. Like we said today, we're talking all about drug design, AI development of a drug specifically as it relates to antivenoms for snake bites, which is a little bit of an interesting application I hadn't really thought of for AI drug development, but before we dive too deep into that topic around snake bites, I want to take a quick second to mention today's sponsor, Mouser Electronics. You've been rocking with us listening to the pod for a little while. You know that Mouser, we're longtime fans of Mouser. They're our favorite place to get electronics for our hobby projects as well as at work sometimes, and, but one of the favorite things that we experienced about Mouser as it relates to the podcast is they have an awesome connection to industry and to academia and they know kind of where the topics are headed. They know where they know where the ball's headed and they can help point you in the right direction with their technical resources. We've included a technical resource today that's talking about how AI is accelerating the drug development cycle. And it perfectly ties into today's topic. So, it's an awesome primer to understand kind of the premises of AI driven research and development. And in which steps in the process, AI might help accelerate drug development. The most promising one being that instead of having to do like a spray and pray type of approach where you take a bunch of random molecules and then you test a bunch of them. They talk about how AI can help identify promising molecule structures to test. So, you only test the most viable candidates, which essentially helps accelerate the testing process because you're not doing trial and error with a bunch of different ones. You're only testing the ones that are most viable, most likely to succeed. And that's a theme that we actually see in the main meat of today's topic later today.

Farbod: That's what I was gonna say. It's so fitting for today's episode. And again, that's a testament to how on the money Mouser is with these technical resources.

Daniel: Yeah, so we are including that as a link in the show notes. Check it out if you're interested. Again, we think it's an awesome technical primer to the topic we're talking about today.

Farbod: All right. Perfect segue.

Daniel: Now let's jump into it, right? Snake bites, just as a premise here, the venomous snake bites impact up to 3 million people per year. Traditionally it causes over a hundred thousand deaths and a large portion of the folks that end up getting a venomous snake bite also end up with a severe disability.

Farbod: About 300,000 a year.

Daniel: Yeah. And that's particularly present in Africa, Asia, and Latin America because traditional antivenoms have to be derived from animal plasma. I didn't know this.

Farbod: I didn't either.

Daniel: So basically, they try and immunize animals with small amounts of snake venom, and then they basically build up antibodies to be able to treat the snake venom. And then they pool the antibodies out of the animal's blood from their plasma. Essentially, you're, I don't know about how humane this is, but at least from an efficiency perspective as well, right? Super high costs, limited efficacy, it doesn't always work. And then there's sometimes crazy side effects when you put the anti-venom in humans. And because it's so hard to produce and so expensive, talking about remote areas in Africa, Asia, Latin America, this isn't a place where they're gonna have ready access, let's say to a lot of these antivenoms. And unfortunately, snake venom varies a lot by species. So, it requires region specific treatments as well. So, it's not like I can like take the antivenom for one specific type of snake and then go fly this to South America and you get bit by another snake there and the antivenom works. It's actually very specific to by species. So, it makes this whole antivenom development and antivenom a synthesization process really complex and really inefficient, which is basically why even though we know the premise of how to solve snake bites and how to treat venomous snake bites, it's still either disabling or killing almost 400,000 people every year.

Farbod: Absolutely. And what was interesting when we were doing this reading in the article, they mentioned that the venom of every snake is especially its own special cocktail that impacts people differently, which makes sense. If you watch the movies, I don't know, sometimes it makes your blood into jelly. I don't know, there's other impacts that you see on the various Indiana Jones movies.

Daniel: I've been bit by a snake, by the way. Fun fact.

Farbod: Okay, well, I feel like you should have left with that because I didn't know that. What's the story there?

Daniel: It was in high school. My brothers and I were building a cabin out in the woods and then at some point we decided we, it was on the lake and we wanted to build a dock for the cabin, which is cool. I was in the mud in the lake, like digging the holes for the posts for the dock and an adult copperhead came up and bit my knee and then swam away.

Farbod: Our copper heads. Yeah. That's what I was gonna say. Are they venomous?

Daniel: Yeah. Venomous snake bite. I survived, right? It fortunately adults know how to meter their venom output. And they would say it's just more like a warning bite. So, it was a small amount of venom. If it had been a young copperhead that bit me, guys, anyone who lives near copperheads, if you realize this, if it has a green tail, it means it's young and it doesn't know how to meter its venom. That thing probably like will release all of its venom in one bite and you have to go to the hospital right away. If it's a green tail, just think about it. If you ever get bit by a copperhead with a green tail, 911 go to the hospital right away, cause it's probably released enough to be lethal dose.

Farbod: Okay, that's good to know, because I didn't know about any of that. And two, by the way, so did you not go to the hospital? Were they, did you just know?

Daniel: We didn't go to the doctor right away, but that was a foolish decision, but it turned out to not come back and bite me in the butt. It only bit me in the knee.

Farbod: Folks, that's why you love The Next Byte Podcast, right? Not only do we talk about interesting tech, we talk about people directly impacted by it.

Daniel: Yeah, well, and that's not to make this all about me, but it's to say like, hey, this is not something that you think only affects people in the desert, in the jungle, with rattlesnakes and with cobras. This could be you in your suburban Northern Virginia area digging a hole to put a dock in the pond and you end up getting a warning bite from a copter head.

Farbod: Yeah, well, what I was gonna say is what was interesting for me to read in this article is that although every venom is different, it's its own cocktail, they all have essentially one thing in common, which is the three finger toxins. And it's this like protein family, which poses the biggest disadvantage in terms of using antivenom as a medicine, as a cure, because it completely evades the human immune system. So, like whatever fight you're trying to put up, it just gets around it and it becomes pretty lethal. And that's kind of what the folks here at the Technical University of Denmark were trying to tackle. So, like, we just talked about all these problems, there's all these variations. There is no one solution that kind of fits all. And they were like, but what if there is? What if there's some way that we can actually hit the foundation of whatever this venom is and then, I don't know, help it crumble that way?

Daniel: Well, and I think this is an important point to talk about who the cast of characters is working on this research, right?

Farbod: It's a star. It's all stars. Let's just say.

Daniel: Well, yeah. So, one of them is Dr. Timothy Patrick Jenkins. He's at the Technical University of Denmark. He's a bioengineering and data science expert. He's also a first connection for me on LinkedIn. So excited to talk about you, Dr. Jenkins.

Farbod: Humble brag.

Daniel: But the other one, University of Washington, is a star in his own right in the chemistry world, Dr. David Baker. He won the Nobel Prize in 2024 for computational protein design, the Nobel Prize in chemistry. So, we're talking about an expert here in bioengineering and data science meets together with a Nobel Prize winner. His research was in computational protein design, which means using computers to design proteins specifically for medical treatments. This is the all-star cast that's ready to tackle this problem, right? They look at the structures of the toxins inside snake venom, and they're trying to see if they can generate any proteins using computers, using AI, to bind to and deactivate these toxins to make sure that essentially, you're able to have an antivenom treatment that's protein-based without having to synthesize it using animal plasma and using AI and using the fundamental structure of the toxins, generating a protein that you can inject as a treatment and use that to disable the neurotoxins found in snake venom.

Farbod: Yeah. And it's, we were just talking about the Mouser article at the beginning, right? How you can use AI models to kind of limit how much spraying and praying you're doing in the lab. The team explicitly calls that out saying, hey, this deep learning model that we're using, we only had to test a few molecules, which significantly reduced the discovery phase and that significantly reduces the cost, which was one of the biggest cons of these anti-venoms that you had mentioned earlier. The next part, you briefly went over it, but the anti-venoms use plasma. This, just because you're growing proteins, which can be done with the average microbe that is largely available in any lab, again, reduces the cost significantly. So now that you've saved money during the development process, now you're saving money on the production side. And then at the very end of it, you're no longer confined to developing a treatment that's species-specific. This is supposed to be as general as possible.

Daniel: And so, let's talk about where they were able to find some general themes, right? So, they looked and studied mainly Cobra venom. So, I'm not sure how applicable this is to other types of snakes, but specifically for the Cobra family, they, I guess, first to understand the secret sauce of how they solved it, they had to first understand the secret sauce of Cobra venom. So, so they looked at Cobra venom and tried to understand what's the most major class of neurotoxins found in this venom. How does it work? What's the mechanism inside the body and why aren't current, current traditional antivenoms effective always at treating this. So, what they did is they found three finger toxins, which I'm guessing is alluding to the structure of the toxin. I don't think the toxin actually has hands with fingers, but they found three finger toxins, which are the most major class of neurotoxins found in cobra venoms, these three finger toxins are small, they're highly stable. So, they don't break down that easily inside the body and they're small enough that they often evade the immune system. So even if your immune system recognizes that there's a snake bite, or you even get an anti-venom injection with antibodies, your immune system may not always be able to fight these three finger toxins because they're small enough your immune system can't detect them. And so, what they did is they basically tried to design, I kind of alluded to it earlier, we tried to design proteins that will bond to these three finger toxins and directly neutralize them and basically create a more effective treatment against these three finger toxins than even the traditional antivenoms have.

Farbod: Yeah. And it's important to note, they called it out in the article, so I'm just re-raising their FYI, that what this treatment in its current state allows the researchers to do is increase the efficiency, I guess efficacy, of the antivenoms that are already on the market. So, it's meant to be used as like almost a booster or a supplement to using what's already done. But what I found to be interesting is that it also is the first baby step towards a new method of developing antivenoms and no longer being confined to...

Daniel: Well, and I think that's-…

Farbod: I don't know if your take was different, but that's like my main take away from it.

Daniel: Well, I'm with you. And that's, I think right now they're identifying a gap in where traditional antivenoms aren't fully effective at protecting against lethal doses of these three finger toxins. But like you said, this is a proof of concept of them being able to identify a toxin structure, design, test several candidates. But again, they were only testing the most viable candidates. They didn't have to do a ton of trial and error. They were only testing the most viable candidates. But they tested candidate proteins that could bind to and neutralize these toxins from cobras. And I know that snake venom has a bunch of different types of toxins. So maybe this is their first step at neutralizing the three-finger toxin. But if they are able to break down more cobra venom structures, right? And understand the different toxin groups, they could just rinse and repeat this a couple of times and then create an anti-venom that's fully derived from proteins, from lab synthesized proteins, not from immunizing animals and then stealing their plasma. And then you're able to use a completely lab designed anti-venom to treat against snake bites. In this case, right, they're only targeting one toxin that's a part of this cocktail of toxins that Cobra bites contain.

Farbod: Right. And their lab results are also pretty interesting. In testing with mice across different toxins, different doses of this medicine that they've developed and the type of protein that they use that was spit out from the deep learning model, the survival rate was 80 to 100%, which those odds are looking pretty good for the mice.

Daniel: Yeah. And not just like I mentioned, you can get a snake bite. That's a not a lethal dose. Right? So, what they did is in these mice, they injected every single one of them with a lethal dose and 80 to 100% of them, depending on the trial, were able to survive, which is really impressive. And again, unlike traditional antivenoms, this was specifically related around, we're going to introduce a lethal dose of three finger toxin to mice, and then only introduce our lab generated protein to treat that. And so, there was no animal immunization required in this treatment process. They put the toxins straight in and then they put the proteins in as well. And up to a 100% of these mice were able to survive a lethal dose, which is pretty impressive. And I think kind of thematically speaking here, as we mentioned, right? This is a, and they put it in their abstract, I think, which is interesting. And that's linked to the article that we included in the show notes is they said, not only is this something that we can lab design and tailor toward the toxins, it's also low-cost production compared to the animal immunization method. And it's also highly manufacturable. So, they're saying they're able to produce these for less dollars per treatment and they're also able to produce them faster and easier. So, they're saying this is the potential to be more widely applicable, especially in low-income regions where currently the animal immunization and plasma stealing method doesn't necessarily work. What's also interesting is the potential for broader applications, right? If you can use the same AI-based protein design techniques to treat other diseases, including certain viral infections, et cetera, like, are we able to use this same principle that we used against toxins and snake bites to go identify how to neutralize certain viral infections, et cetera. And I think it's an awesome proof of concept, let's say, or another feather in the cap, let's say, for AI drug discovery. As we mentioned in the Mouser article at the beginning, right, this project essentially is a microcosm for the rest of AI drug design, which is AI's rapidly accelerating how fast we're able to develop new therapeutics, reducing both the time that it takes us to find new treatments and then also the cost and testing and proving them out.

Farbod: Well, that gives me a whole lot of confidence in case one of us gets bit by a snake the next time that folks like this are doing research that should be not too sensitive about the age of the copperhead biting us.

Daniel: Yeah, I mean, this is definitely, like we said, it's traditional antivenoms still have to be used for the foreseeable future. But these AI-designed proteins have already proven that they're effective. They can either enhance the existing treatments or potentially completely replace them at some point in the future.

Farbod: I'm excited, dude. I think it's about time you give us a little wrap up and then we got it in the episode.

Daniel: Let's wrap it up. All right, folks, every year snake bites kill thousands of people and current treatments are expensive, they're risky, and they actually don't always work. But what if AI could change that? Scientists have designed AI powered proteins that naturalize in the deadly venom inside the human body. It achieves up to a hundred percent survival in lab tests with mice. They're cheaper, they're faster, and they're safer than traditional antivenoms and this breakthrough is not just the start for antivenoms. It's also the start for drug discovery far beyond snake bites. I think the future of medicine is here with AI, but it's interesting that they're starting first with snake bites.

Farbod: Money. I love it.

Daniel: I'm in. I think that's the pod.

Farbod: Yeah.

Daniel: Peace.

Did you like this episode? Great. Make sure you share it with a friend because friends don't let friends miss an episode of The Next Byte. Thanks for listening and we'll catch you in the next one.

As always, you can find these and other interesting & impactful engineering articles on Wevolver.com.

To learn more about this show, please visit our shows page. By following the page, you will get automatic updates by email when a new show is published. Be sure to give us a follow and review on Apple podcasts, Spotify, and most of your favorite podcast platforms!

--

The Next Byte: We're two engineers on a mission to simplify complex science & technology, making it easy to understand. In each episode of our show, we dive into world-changing tech (such as AI, robotics, 3D printing, IoT, & much more), all while keeping it entertaining & engaging along the way.