Capital Markets Day Transcript - TWST 2026-05-21

Emily Marine Leproust, CEO

I'm glad to welcome our CTO, Si-Wan Chen, who's going to tell us some of our strategy. Thank you, Si-Wan.

Si-Wan Chen, CTO

Hey, good morning, everyone. My name is Si-Wan Chen. I am the CTO on Twisted Bio. I'm an oligo chemist by training. Actually, I was the first employee on R&D side on Twisted. So, like, I've been working in this company for 13 years, like, you know, a long time. I'm really excited to be here because I don't think I've ever talked to a large group of investors in this setting. So I'm a little bit nervous, but I'm very excited to show you what we're able to do. So I think you guys all just went through the tour, the production tour. I'm sure you have seen a lot of really interesting, a lot of amazing stuff there. I want to say a lot of magic we do really started with the silicon chip. So I think you have all seen the silicon chip, which is like a 96-wall plate size. instead of having 96 wells we actually have 10 000 wells on it we call them clusters and like you know within each cluster we can make 121 oligos so that's enough oligo to make one gene and on the chip we're able to make 10 000 genes and like you know on the total chip we're also able to make about a million oligos right and we have multiple riders the riders can run actually really really fast i'll talk a little bit more about it later that give us a capacity of making 32 million articles per day capacity enables a lot of different applications and like you know we also like you know do the dna synthesis base by base like using phosphoramidide we have done so much optimization like you know to really improve the efficiency and every single layer so like right now we're actually able to do synthesis all the way to 500 base pair so we call them 500 mer which is actually quite amazing if you really think about it because as I said I was an oligochemist I remember when I was doing work in graduate school trying to make a 30-mer on a chip that's enough lens for hybridization for DNA microarray that's quite amazing already you're able to do 30-mer and quality good enough to do hybridization that's good if you're able to do 80-mer, 100-mer everybody look up to you saying you can do 80-mer in a consistent basis So that's quite amazing. But right now, we actually do 500 mer in a consistent basis on a production setting. So, like, you know, and also an error rate that's, like, industry-leading. I think we're able to get to one in 2,500 base pair, one in 3,000 base pair. Every rate, that's, like, you know, really unheard of. Like, you know, compared to the traditional synthesizer, you get a fine 500 base pair. That's pretty much as good as you can go. So that's our foundational platform. And on the platform, we really build up a lot of DNA product, right? Oligos, oligopool, gene fragments, clonal genes. And also, like, you know, in the last couple of years, instead of printing A, T, C, G, four different bases, we're actually able to add in modified bases to, like, you know, make some sRNA antisense oligonucleotide. You're actually going to hear a talk from Dylan from GSK later today talking about how they use our platform to enable high-throughput sRNA screening. And on top of it, actually, on top of the DNA layer, We also built up a protein layer where we can do different type of antibody discovery work and, like, you know, antibody production and antibody characterization. So, like, you know, so we have this amazing platform where we can make, you know, DNA and a scale and a very fast speed. We really built up a very impressive MPI machine on top of it. So I want to show you where we were back in 2021. Like, you know, this is essentially the product we had back in five years ago. So we have SimBio product. We have NGS product. On the SimBio side, we have OligoPool. We have Variance Library. We can make gene fragments. We can make clonal genes. And on the NGS side, we had exon panels. We have some custom panels. We had a couple of library appropriations. They're all good product. Like, you know, they're actually, like, we're still selling those products very strong these days. But at the same time, I want to say the footprint, the product length is somewhat limited. And I want to show you in the last five years what we have done in terms of expanding our product roadmap here, our product line. So this is where we are right now in 2026. So I'm not going to go over all the product here, but I want to highlight a few things we can do. So on the gene side, we can do, like we had clonal genes two and a half years ago. We launched the ExpressGene, which offers industry-leading turnaround time for clonal gene product. And a few weeks ago, we actually announced ultra-complex gene in a symbiote beta. That's really leveraging the 500 base pair synthesis capability I mentioned a couple of slides ago, so that we don't have to really worry too much about DNA secondary structure. We're just really going to make the DNA chemically and stitch only a few pieces together to make ultra-complex gene. That serves as a backbone for a very complex mRNA and help us get in the world of nucleic acid therapeutics. You're going to hear more from Patty later today talking about that part. At the same time, on the simbio side, actually we built up a protein layer on top of it where we can do antibody discovery in vitro, in vivo. We also use AI tools to do in-silico discovery work. At the same time, we built up a very impressive antibody expression system and antibody characterization, generate lots of data to help our customers to do AI model training, like, you know, refinement, and help their AI-enabled drug discovery work. And Colby is going to talk more about it, share more insights on the AI side. And the NGS side, like, we expanded our panel product, like, lots of panel product. We put a lot of emphasis on MRD, molecular residue disease, which we really think that area is going to ramp up very, very quickly. So, like, you know, Jimmy is going to cover that part later today. And on the library preparation side, Like, we had a couple of library preparation. We added quite a few more new products, you know, found some generic library preparation, some standard fragmentation, and also some more specific ones focusing on, like, you know, CFDNA, focusing on, like, you know, FlexPrep, like, to really enable microwave conversion from, like, you know, to NGS. That's actually really, like, you know, enabled by the enzyme engineering capability we developed in the last few years, leveraging our synthetic DNA, like, you know, DNA synthesis capability. So it's a really nice product. What makes it even more exciting, I think that's actually just really the beginning of what we can do. Because if you really look into all the applications, it can be enabled by oligos, by DNA, like in terms of oligo number, mass, lens, like in different areas. We can see there's a lot of things we can do. We just need to choose which one we want to go as we continue to move forward. so yeah this is the the table just to reference like you know the way how we perceive like you know what dna can do for different applications so so that's enough about mpi machine we're going to talk a lot more about it throughout the day what actually i want to spend like you know the rest of my talk here to share you is you a little bit more about how we do operational excellence because that's actually a really important part we spend a lot of time on it we're really excited about it and that's also a part might be a little bit underappreciated by people like you know from the outside world so like you know i want to share a little bit more what we do so we always come back to the silicon chip right so that's really our foundational platform we do pretty much everything on top of it and i want to say the chip itself is actually not static it's actually like you know it's a we continue to iterate and improve this platform so like you know a few things I want to share. Number one, we're able to maintain really, really good error rate in the last couple of years, like 1 in 3,000 base pair. Sometimes the best run we've ever seen was 1 in 4,000 base pair, which is super, super amazing. And we also improved the consistency of the DNA synthesis quite a bit in the last couple of years. So believe it or not, actually the DNA synthesis can be pretty variable, can fluctuate quite a lot. There's even a lot of seasonality in DNA synthesis. So really think about it, like, you know, just even like if you're in a rainy day, the moisture can be high, like, you know, that's kind of affected your DNA coupling. Like, even the rush hour, the ozone from the rush hour can help, can generate, like, you know, can actually affect quality of the oligos. And sometimes, like, for the vendor, when you buy all the bulk chemical from the vendor, they might store the chemical outside in the hot summer. Like, the quality is still going to be good, but it's actually going to have an impact on the oligo quality. So we have done a lot of work actually to just really tease out all the details, like trying to improve the performance. So I want to say right now, if we're looking at longitudinal data, for the oligosynthesis has been very flat, like in a month by month, and across all the machines that we have here. So which is something probably cannot be said by other competitors when they have hundreds of machines running plates over plates. Like one common comment we hear from our customers, is, like, they struggle with the current provider because, like, you know, they get lots of batch-to-batch reproducibility issues, things like that. That's not something we have to worry about because of all the work we have done on the synthesis improvements. And at the same time, like, we continue to reduce the cost of the synthesis. As you see, like, you know, in the last three years, from 2023 to 2026, we're able to reduce the synthesis cost by 60%. And a lot of cost reduction comes from the use of less solvents. So as you see in the middle, back in 2023, it took us about 51 liters of chemical to make a million 100 mer. That's quite amazing already, because that's a million different sequences, 51 liter of chemical. And we did a calculation back in the days. It's actually like 99.8% reduction compared to the column-based synthesizer. So which is like, you know, we're talking about three orders magnitude lower than what people normally use. Yet, like, you know, in 2026, we're able to reduce that number to 14 liters for a million oligos. Like, you know, what does that mean? 14 liters, if you really think about it. So, like, you know, look at this bottle. This is a 500-mil bottle. So we're able, a bottle of solvent like this can be used to make 35,000 oligos. I think that's the scale we're talking about here. I think for the people who are around DNA synthesis, a bottle of chemical probably normally used to make a few oligos, but here we're able to pack in 35,000 oligos into a small bottle of solvents here. And more impressively, I think we have done a lot of work to reduce the turnaround time. So, like, you know, in the first half of 2023, it took us 26 hours to make 1,100,000 nucleotide oligos, which was, like, really fast at that time. And when we launched the express gene, we tried to really look into everywhere, every place we can shrink the time. So on the right side, we're able to reduce it from 26 hours to 13 hours. And at that time, like I was joking, oh, that's super fast already. I don't think there's much we can do to make it even faster. Yet in 2026, we're able to shrink it down to seven hours to make 100 mer. And that's actually a 73% reduction in turnaround time. and not only we get kind of 37% reduction turnaround time, what actually means we have four times increase in capacity so we have the same number riders compared to three years ago but our oligosensor capacity actually increased by four times with the same number riders that actually enabled a lot of new applications, like new product for example, I think I talked about the ultra complex gene which need 500 nucleotide oligos is if we took the 2023 chemistry it would take us 5.5 days to make 500 mer on the rider and that's going to take up a lot of capacity and also like the product is not going to be competitive because like in a front get-go you spend almost a week just making the oligos but nowadays we can actually do it in less than 30 hours for the 500 mer which like you know really enabled ultra complex gene product and at the same time like you know we built up enough like enormous amount of capacity 32 million articles per day capacity really support the needs for MRD customers because they all come in with the personalized panels another thing we do it's automation we actually like you know take automation very seriously we do we try to do automation like you know from start to finish so when you go through the production tour like I'm sure you see a lot of automation like we actually use a lot of Hamilton right that's the liquid handler to move liquid or pipette liquid move plates around so you know a lot of applications like it seems like Hamilton works well enough right just to put them load the load set of deck load the tips load the example and like you know in an hour or so like come back with what you need to do so that works well for like a lot of applications but not good enough for the gene production because the gene production is actually very like you know has many steps has 20 plus steps in the process there if we were to use like you know standard long workstations like Hamilton the issue we're run into is like take operator to time to set up the deck, load the tips, load the plates, and then run it. And once the run is done, have to move the plate from one machine to the other machine, do it again. So it's actually quite labor intensive. There's created a lot of bottlenecks in our manufacturing process. So what we did here is we actually we identified all the key bottlenecks in our production process. We built up integrated automation, integrated system to enable all the bottleneck steps here. So those are four major systems we built, from oligo fragmentation system to take the rider oligos to gene fragments, and then the second one to, then we clone the fragments, then we play them, we pick all the colonies. We have to pick millions of colonies to support our demand. And then once we pick the colonies, all the colonies go through the next-generation sequencing sample prep system. We actually process about 10 million samples per year on gene facility, which I believe were probably more than the highest volume in terms of sample volume for NGS. I don't think anyone else in the world can match the scale, the number of samples we process on sequencing every year. And once you identify the perfect clone, we're going through a plasmid preparation to get purified plasmid, and they're ready to send to customers. And that's essentially how we enable the scaling with more and more integrated systems. As you see, right now we have a total of 20 integrated systems on the production floor, and some of them you already see on the production, like to enable genes and proteins, and we have quite a few systems in South San Francisco as well to support our NGS product. And that number was actually much, much lower just a couple of years ago. We only have seven systems. So we continue to develop the tool, implement the tool, continue to improve our production by having more and more integrated solutions. One thing I really want to highlight is when we build tools like this, we always think through scalability and trying to make it future-proof as much as we can. So when we brought a new product, new process, we really want to make sure multiple process, multiple product line can run on the same machines. So back to the ultra-complex gene, which the workflow is somewhat different from what we normally do for standard genes, but we're able to leverage all the systems you see on the left side and with minimal addition for equipment. That's how we want to make it future-proof and future-compatible. At the same time, with automation, we're able to improve our capacity, improve our throughput, and also with lower footprint. So when you walk into the GeneLab, the GeneLab 1, that's actually the footprint of the GeneLab 1. So we used to have two assembly lines, like, you know, as you see on the right side, those two green blocks. So that's all the, like, you know, individual automation back in the days. So that's operators basically do work on automation one, finish the work, move to the station two. As they work down the aisle, they go from oligos to gene fragments. That's what we used to do. So it's automated, but we still need, like, you know, three to four operators running through the process. It takes about nine to 10 hours to go through the process because it's automated, but like you do need like you know pretty much all hands on time to make it happen and over 10 hours you can make 12 plates then in the last year we actually put in two integrated system the two solid like in a green box on the right side the footprint is one-fifth of what we have for the gene lab for the original workstations but we're able to like you know in these cases operators only have to come in like you know a couple of people loading the plate loading the tips they come back in six to eight hours like you know with 16 plates ready to go they don't have to worry too much about it they can actually work through the other work like you know during the day so now like with the integrated system we're able to manufacture two times the fragments with one-fifth of the space now like actually we open up the gene lab one we're doing more specialty work we can use it for cell-free work we can use it for mrna work in the future that's how we ensure scalability and sustainability as we continue to grow. So, in twist, we always go crazy about speed. I feel like we're always in pursuit of speed here. I think I already talked about the story about how we reduce the time for oligosynthesis, right, from 26 hours to 7 hours in the last three years. And we did something pretty similar for gene product. I remember many, many years ago when we launched a gene product. took us like 30 plus days to make a gene. It's a very, very slow process. And we refined the process. We're able to make standard genes up to 5KB in 10 to 15 business days. And when we launched ExpressGene, we're able to really shrink it from 10 to 15 days to four to seven business days, which is faster than anyone else in the world. And if you want things to be even faster, we can go with gene fragments, like in the two to four business days, super high quality. I think James just gave an example, like, you know, the one-day fragrance we made for Mike Wiley, actually someone we worked with, like, a few years ago during the Ebola outbreak, like, you know, on the NGSI. So that we're able to deliver genes in less than one day, actually that gene spent more time on the FedEx truck than, like, in our facility. That's just how fast we are, how crazy we are when it comes to, you know, keep continuing to improve the turnaround time. And we're doing something pretty similar on the IgG side as well, right? when we launch the product 20 to 25 business days good but not fast enough and able to reduce it like to 10 to 15 business as of today and if you want things to be even faster we can do them in cell free manner and then you can get your antibodies in five business days so so like you know i think on the operational excellence side like there we look at a speed like because we always want to go faster we also trying to reduce the turnaround times that with the same number of people we can do more work we always look for opportunity where we can like you know save money like reduce the cost and also ways for us to improve the capacity like improve the throughput so like you know this basically show you like you know what we have done in the last like you know three years actually even in the last year and a half we have 46 tractor project to improve like you know tractor cpi continuous process improvements to improve our operational excellence resulting in tens of million dollars saving in the last two to three years. I'm not going to be able to go over all the details about this 46 project. That's just too many. But I do want to highlight one project we did on the panel side. So this is something like, you know, so we always make, like on the NGIS side, we make panels. We're always the best when it comes to making panels. The highest quality, faster turnaround time, we can make panels in two to three weeks, which is like much faster than other people, which could be easily double our time. And starting a couple of years ago, actually, we saw like, you know, MRD. We definitely see that personalized MRD is going to be a big trend. People are going to do more panels, like it's going to be all personalized. There's going to be lots of them, right? So just to find out understanding, like we're like, hey, customers really need the panels to be incredibly fast because MRD patients cannot wait. The The window for the test is very short. We want to make the panel as fast as possible, ideally less than a week. And we also need to build a very highly efficient process because we need to be able to make hundreds, thousands, millions of panels because that's how many cancer patients we're talking about. And we cannot compromise the quality. We still have to maintain the highest possible quality because all the patients deserve the best panel we can make for the cancer detection. so we know that's what we need to do then like you know we planned a couple of years ahead of time saying hey we see the surge coming it's not there yet but we need to be ready so like we did a lot of planning like you know a lot of preparation we know we need to put in like fully walk away automation system to make the panels we need to build in like you know the software to really make this happen like you know the integration of the software mes and and the hardware and the hardware and like you know then it's all up to execution like you know I think like one thing we do really really well and this company is execution and end of the day like with the fixed headcount like you know for the for the panel production we're able to increase the capacity by 10 times and like you know actually reduce the turnaround time very significantly from like you know two to three weeks to make a panel to like less than five days to make a panel at the same time like we're actually able to like you know take the learning we have to make MRD panels, like an expanded for all the panels, actually resulting in more than $2 million saving, like, you know, in the consumables when it comes to panels. That's actually the year one saving. The number is only getting bigger and bigger as we continue to scale our operation. So I think this is my last slide here. I think just to want to give you a taste of, like, you know, how we see the world, like, you know, like when it comes to MPI, it comes to, like, operational excellence. I'm incredibly excited about what we're able to do. Like, you know, I'm happy to talk to all of you, like, you know, offline. I'm going to pass the mic back to Emily here.

Emily Marine Leproust, CEO

Thank you very much, Siwon. Much appreciated. So next, we are going to have a fireside chat with Dr. Francis Arnault. It's my great pleasure to bring Francis here. I have my note cards. I feel like Alex Trebek on a game show. and we have a runner so there's someone with a microphone in the room so I have my own questions but if you have a burning question please do not hesitate to get in so to start thank you Frances I remember when we were talking about a twist when we were a baby startup company we were thinking about who can you bring in the SAB? We need the best. I told the board, you don't know that, but I'll tell you now, we need Frances before she wins a Nobel Prize because we'll never get her after. And so you graciously accepted it. And maybe my first question, I will start there, is we had a number of SAB meetings just at the beginning. of startups failed. Did you think we were going to make it? did you think where crazy was never going to work?

Speaker 8

I thought it was crazy, but that you were going to make it. Because the need was so great and the ideas were clearly the right ones, right? To miniaturize and go with the silicon. I'm an engineer by training so this is deep engineering and not appreciated by many biologists. So I thought you were exactly the right person to make that happen. And you did it.

Emily Marine Leproust, CEO

We did it. Well, if it's good, it's the team. If it's bad, it's me. So at Swiss, I'm only in charge of the bad stuff. You know, we just did the tour. What do you think? What surprised you? Good or bad?

Speaker 8

Well, I have to say I'm blown away because I moved away from the SAB in the early days. I had to deal with a whole bunch of other things at that time. So I didn't follow in detail. And this is the first time I've actually visited the manufacturing. So I love it. I have to say I love it. The attention to execution, although you were always attentive from the very beginning. You put in quality control as number one. Careful engineering is number one. So I'm not surprised. But it's fantastic because I use these products all the time in my research. And we love it. We absolutely love it. But I'm not surprised to see how well you've been able to scale this.

Emily Marine Leproust, CEO

That's great. So you're an expert in enzyme. I think a lot of your career has been focused on enzyme. I know you have to catch a plane after this, so you'll miss our part in enzyme engineering. But just tell us why enzymes are so important from your perspective.

Speaker 8

I've been working on enzymes forever. And these are those remarkable molecular machines that convert really cheap materials like carbon dioxide and sunlight into complex chemicals like trees or you and me. All the chemistry of the biological world is done using these machines. And now we're coming into a period of synthetic biology. It's not just all about health care, folks. It's about everything. And biology can make virtually anything. It's just bringing the cost down, just. It's bringing the cost down and opening up, expanding the possibilities for biology. Because biology makes you and does a pretty limited set of chemistries. But evolution is an algorithm that can go out well beyond. So I see now enabled by what you're doing and AI, a combination, as enabling biology to do any chemistry you want. to make any pharmaceutical, to make any fuel, to make textiles, to make all sorts of things we need in our daily lives that are done using dirty chemistry today, or not even done at all. So I think we're at an inflection point, really, that's more exciting even than it was 12 years ago.

Emily Marine Leproust, CEO

I totally agree, yeah. And so enzymes are proteins, right? So there are non-therapeutic proteins in the two big categories. There's also the therapeutic protein side of it. You're an academic. You're in the DSAB. You're founder of a number of companies. You're on the board of Alphabet. So you have, I think, a unique view into protein engineering and also as it relates to therapeutics. And AI is changing that field. And one of the questions, and I'm sure a lot of investors in the room have that question, is around AI-driven drug discovery, is it a flash in the pan? Or is it here to stay?

Speaker 8

Well, what do you see? What are you seeing? Because you're seeing the same thing I'm seeing. And I want to hear it from your mouth.

Emily Marine Leproust, CEO

Well, I'm going to say what I see. What I see is like, you know, we're going to need a bigger boat in terms of capacity. I think that what we see is there's a first path of model building where people need a lot of sequences expressed in protein, tested against a number of targets to build a model, then we find a model, and then you have to turn the crank. And the turn the crank needs more DNA. And I think what we're seeing is AI is going to be the first path. for drug discovery in vivo and in virtual are still going to be needed but AI has a huge advantage is that it's fast it's much faster than having to either enuculate a mouse you have to wait for the mouse to be your drug company and or you have to do face display or use display and that takes days so that's what we're seeing is AI as the first pass does that

Speaker 8

jive is what you're saying? Well, of course, because enzymes are more complicated than the binding proteins used in therapeutics. It's the next generation, right? So it's still, AI is still not good enough unless you combine it with these optimization methods for which I won the Nobel Prize. So imagine that you can combine AI to get you a good starting point for chemistry, but then you combine that with iterative optimization, all of which uses synthetic DNA in an active learning cycle. You just press the button, and I think in the next five years, we'll be able to genetically encode any chemistry. So I'm super excited about this, and I know it's not a flash in the pan. I write the checks for that alphabet, and they're really big checks you know it's not a flash in the pan because whatever you can't do today that won't be true tomorrow and we're just going to get better at it so then do you think it's gonna make dna

Emily Marine Leproust, CEO

obsolete or do you think it's gonna create more demand for dna because you're gonna have to try

Speaker 8

more things so that's a really good question if we get perfect at ai right you just make one gene and you're done. That is completely unrealistic, right? Because there's so many specifications that we don't even know to make on the DNA. How it cures a disease, how it catalyzes a reaction in a particular laundry detergent formulation. I mean, there's all sorts of things that we can't write down as specifications that only become real when you translate the DNA into the sequence that you compose using AI into the real world. And I think it was said, you said it this morning, DNA is the point at which you translate your computation into the real world. That is the physical manifestation of your computation. And I think it always will be, because it's not efficient just to go and make proteins synthetically. Nature does it by DNA, and she does it for sugar.

Emily Marine Leproust, CEO

I do it for sugar, too.

Speaker 8

A lot of sugar.

Emily Marine Leproust, CEO

A lot of sugar. So if there are any questions, raise your hand, and we have a runner that will bring you questions. So the last nine years, actually, we had a number of projects together. Early on, we did the membrane protein cameras, and then with the machine learning design charger optogenetics tools. And then very recently, we did a carbon transferases project. So as a user of DNA, how do you choose a provider? You know, you're using Twist a lot, but there are other providers. What's the criteria for you to make a decision of where to go or do you lead that to your team?

Speaker 8

Well, I come from academics. We never have enough money. So price is very, very important. That's why I yank on your chain and say, hey, I want a better price, but I'll give something in return. So, yes, price matters a lot for the academics. But I also sit on the industrial side, and their time and quality are the key things. So there are different metrics, and I have to say I'm pretty impressed that you can meet all of those metrics.

Emily Marine Leproust, CEO

Yeah, a lot of companies, between the speed, quality, or price, usually you have to choose two, any two. But with Swiss, we pride ourselves in you get all three.

Speaker 8

And that will be important because graduate students are relatively cheap. they used to be almost free and so time was not you know manipulations that wasn't important now they've become a little bit more expensive and so we'd actually do a little bit of calculation over the trade-off do we want the whole gene or do we you know do all that ourselves but in industry time is especially in this new AI era time is turnaround time is really critical and as we move into active learning test build design cycles the ability to turn that around and generate you know the key data will mark who wins in this race and believe me it is a big race there's a huge amount of interest in biology as a manifestation of ai and not just therapeutics as i say it's all of chemistry uh you know it's going to be a race who who can do this

Emily Marine Leproust, CEO

That's why we love speed. So, there's a question. Great.

Speaker 12

Puneet. Thanks for the question, and great to see you here. So, the question is really about there are a lot of assumptions being put forth in AI in terms of speed, speed to therapeutics, speed to drug discovery. can you talk a little bit about in terms of scaling where do you see as you talked about briefly about chemists role is going to change a bit too but can you talk about the scaling we're going to see or there's an assumption that we are hearing in the market that you know in 10 years we can see a number of therapeutics coming to market that can actually resolve a number of diseases. Now, that's a big assumption being put out there. Disease and biology is complex. Can you talk about the scale that needs to happen, the hiccups that can potentially happen, and then

Speaker 8

how long is this versus 10 years? So I sit on both sides. Demis Hassabis says we're going to cure all diseases, but he doesn't talk to the FDA. And you can't just go and test these things willy-nilly. So, yes, AI might cure some number of diseases, but it's all going to be at the rate at which drugs can be developed. And the pharma people are much more realistic, but maybe not quite as visionary. So what's going to happen? A lot of it depends on what happens on the regulatory side and on clinical trials. How do we understand the efficacy of these AI-designed drugs?

Speaker 14

Thank you. Any other questions in the room?

Emily Marine Leproust, CEO

I'll keep going, and if you have questions, please raise your hand. So as a scientist, I guess I talk to a lot of scientists, and I ask people about protein engineering or any topic almost. You ask three people how to do it, and you get five answers. And I'm always shocked that there's not more standardization, and we're trying to turn it into a competitive advantage by saying we'll do whatever you want. But why do you think that is that there's not more centralization? Like everybody wants to do something different.

Speaker 8

Hell if I know. Well, scientists are funny ducks. They like to put their own stamp on it. That's one thing. So there might be one great method out there. That doesn't mean everybody piles in to do it, Especially with a complex problem like protein engineering. Every protein is different. And every landscape, we call them landscapes, how you go and explore different sequences is different. How you measure everything is different. There's a great deal of bespoke engineering that goes into it, which really tends to push people into their own methodologies. That said, I do think, Emily, that there will be a push the button that maybe not be optimal for each problem, but will be optimal across most protein engineering tasks. And that will involve machine learning, active learning cycles with machine learning. So that project we did with you way back then with the optogenetics, that was the first time machine learning was applied. And, in fact, we developed those methods in 2012, 2013, and did that collaboration with you to demonstrate how you could engineer new properties. And that was well before the ChatGPT era.

Emily Marine Leproust, CEO

Well, speaking of ChatGPT, we have a personal question. So, when you use an LLM ChatGPT, it's all about the prompt, asking the right question. So, you know, obviously you're massively accomplished. How were you able to ask the right question? How do you think you got here? Like, what made you get here?

Speaker 8

I think it was desperation. Right? If you're pushed hard enough. I came out of an engineering background, and I jumped into protein engineering when it was a brand-new field. And the people who were doing it, and I'm sure you have had exactly the same experience, the people who were doing a little bit of protein engineering were from structural biology. And their whole mindset was you had to get a crystal structure of this very complex molecule first, which often people couldn't even get that, in order to go in and then rationally design everything. And I came in and I said, well, I won't get tenure or I'll die first before that happens. and I had to come up with some very different engineering mindset, which was look to the best engineer on the planet, and that's called evolution. So to me it was totally obvious, but to the field it was completely non-obvious. And I think in DNA synthesis you experience the same thing. Everybody was doing it the same terrible way, terrible way, and there's no way you could scale that. And you came in as a chemist engineer and said, And no, we have to completely rethink that. Isn't that the case?

Emily Marine Leproust, CEO

Yeah, being a contrient goes a long way, right? And I always say, one, to the team is like, because what we do is incredibly hard. And sometimes we're like, oh, my God, it's so hard. I'm like, good. Because if it was easy, every idiot would be doing it, right? And then number two, I always, you know, when there's a plan, okay, let's do this, I always ask why. Why this way? And the worst answer for me is like, oh, because everybody else is doing it. No, we're not doing that because everybody else, they have more resource than us. They have more experience. They have better channels, better capital access. And so how can we beat them if we do the same? So totally agree.

Speaker 8

But then you also have to have a vision of when you enable this capability, what becomes available? And how do you capture that and capture at least some of the value of that?

Emily Marine Leproust, CEO

Yeah, totally agree. And that's why I think it's very helpful to be technical. I mean, when I started Twist, I kind of was done playing my PhD because they're like, oh, you don't know business. But I think you can learn business easier than learning the technology. And when you talk to a customer, if you don't understand what they're doing, When you don't even know your product, how can you be effective? And so I think that has been very useful, being able to drop me in any accounts and I can sell any products that we have.

Speaker 8

But what I love is what the technology has enabled that we could not do before. And in my work, some of the same things happened, and that's how I chose problems, was what can I demonstrate that you just couldn't do with rational design? A good example is how do you make an enzyme work in a laundry machine? That was a big market for enzymes. How do you make it be happy? What self-respecting natural enzyme wants to work in your laundry machine with bleach and surfactants? There's no way you could design that. It had to come from directed evolution. And you can't go out to nature and find that. So you had to have some methodology. And you've done much of that, right? You've enabled people. My methods enabled people to do a whole bunch of much more important things than make laundry enzymes. Your technology has also enabled the whole synthetic biology industry, I would say, to do things well beyond what even you could imagine.

Emily Marine Leproust, CEO

Yeah, that's true. Any questions in the room? Okay, all right.

Speaker 3

Thank you for your time, Dr. Arnold. I'm curious, you know, there's a variety of models that exist today, AlphaFold, Bolt, and others that I think are adopted by pharma. I would be curious for your perspective on the utility of those models and drug design today versus what might be required from a new data generation standpoint to ultimately get to the point of, maybe not quite to the point, but a sort of push-button get drug or get closer to that point. rather than the models have been largely built on third-party existing data.

Speaker 8

So I'm very excited about the models. I love that Demas won and David Baker won the Nobel Prize for understanding protein folding and then design of proteins. But the bottom line is that structure is not function, right? Being able to predict the structure is a game. It's like winning at chess, and you have a good metric for that. But what we want in the synthetic biology community is something that does something. And the models just can't capture that right now, but I think they will. They'll start beginning to capture that as we get the right kind of data, which we don't have. So even though that problem has not been solved, we are getting close to getting things that can solve it through further experimentation. So this is why it's a beautiful time, right, for making DNA, because we're close enough that we just need a whole lot more experiments in order to learn what it really takes to make something that's useful.

Emily Marine Leproust, CEO

Question for them, yeah?

Speaker 7

Hey. To that point, I guess, if a lot of the open source tools are protein structure today and a lot of the drug developers are incentivized to have more siloed data sets in regards to function, how do you see this playing out over the next decade? Is that going to be the path forward that each developer will kind of keep some of that data in-house? Is that going to be sort of a rate-limiting factor on the field in general? Or do you think these open-source tools will move towards function and more downstream practicality?

Speaker 8

I think that's a really good question, and I don't know how to answer that, because I don't know how much data it will take. There are those who argue that enough data on your particular system is all you need, right? You don't need a world model that works across all modalities. And to me, as an experimentalist, that makes more sense because I know how bespoke every protein is. On the other hand, if Demis is right, you learn across all proteins and you just have a model that does anything. He doesn't know anything about proteins. So, I mean, to a first approximation, there's no quotes given here, right? This is Jenna Mouse rules. No, I love Demas. But so I don't know who's going to be right. But I do know. So I'm on the board of Generate Biomedicine. So it went public about six weeks ago. AI produced antibodies. We integrate it with a lot of experiments, right, to get to the right developable drugs.

Emily Marine Leproust, CEO

I think there was a question in the front. No questions? All right. Maybe the last question for me, what if you have to start again today? So suspend disbelief, you're back, going to universities. If you had to do it all again now with the AI tools, with the ability to have the wet lab outsourced, even maybe a twist, either at Caltech or not, if you had to start again today, what would you work on and how would you do it?

Speaker 8

Two answers. And I want to ask you that question. I love proteins. Protein engineering is not solved. So you could jump in at the place it is and still do really important work. Enzymes are not solved. But I am a student of evolution. Evolution works at all scales. And so why not apply some of the same ideas to tens of molecules, to whole cells, to organisms, to ecosystems? It's the same design process to design anything in biology. And so that means whole new scales of DNA, right? Not just one gene, but a whole ecosystem of organisms. And I do know that some of the most visionary people in the community are really thinking about that. How do we use AI, for example, to design whole genomes? That's happening. How do we use AI to design whole ecosystems? That's going to require a hell of a lot of DNA.

Emily Marine Leproust, CEO

I love it. What would you do? So I love DNA chemistry. I know how to write DNA, read DNA, sell DNA. So I think I was built to build TWIST. If we had to start today, frankly, I think it would be very hard. because when we started 2013 years ago, we knew that we had to raise a billion dollars to get to exit velocity. And back then in 2013, you could do it. You could raise the first $600,000 seed round and then a $9 million air round, all the way telling everybody, at the end, it's going to take a billion dollars. And we could do it because capital was available. And I don't know if we could do it today. I don't know if that...

Speaker 8

It's all being sucked up by SpaceX.

Emily Marine Leproust, CEO

And AI, yeah. So I'm glad we did it then because I think now it'll be tough.

Speaker 8

That's sad to hear because the Anthropics are raising lots of money and they're going to be buying all your DNA.

Emily Marine Leproust, CEO

Yeah, and we were able to do it, so it's good. And, you know, necessity is the most of invention.

Speaker 8

So do you think one of those companies will buy Twist?

Emily Marine Leproust, CEO

Well, now I'm being triggered. So hopefully they'll buy a lot of DNA. But I always say that that's not our goal. Our goal is to ramp our revenue, and eventually we'll get to buy Illumina and Thermo Fisher. That's not long-term guidance, by the way. But, you know, this is America. We are for sale every day. If there's enough zeros on the check, I will fly and wash your car. So on that, thank you again so much. I know you have to catch a plane. I very much appreciate the effort. I very much appreciate the partnership along the years. And we can't wait to continue being your DNA provider or protein or RNA to enable the great work you do.

Speaker 8

And thank you. And thanks to the whole team here. It's just marvelous what you've created.

Emily Marine Leproust, CEO

Thank you.

Speaker 8

All right.

Emily Marine Leproust, CEO

So next, oh, sorry, back. Next, we're going to hear from Colby Souders, our chief scientific officer. Colby is our own drug developer. And I believe that at least four drugs that he had his hands in the discovery are in the clinical trial stage and a few of them at stage three. So Colby, take it away.

Colby Souders, Other

Thank you. Thank you, Emily. Pleasure to be here. Thank you everybody for coming and those of you making the trip. Hopefully it's been a great day so far. Fitting to follow up that fireside chat with the AI topic, which I know is of particular interest for many. But we heard a lot of great comments in that. And so now what we're going to do here is dovetail that conversation into how we approach AI at Twist and how we solve this for other companies and how we enable AI companies to scale and to fulfill their promise to the market that they are making. So I'll dive into this in more detail. And hopefully by the end of this, you'll understand, you know, kind of our position on the market as well as what our solutions provide to solve that. So this is a slide you saw from C1 in the earlier slides. So you can see, first of all, a lot of products that we make touch that AI-enabled drug discovery. Now, I'll focus mostly on our IgG and antibody characterization, but also mention in the next few slides how all those other aspects that are touching that area provide those tools and what differentiates our platform in order to enable the AI drug discovery in IgG and antibody characterization. So first point to that being, Of course, we've heard a lot about DNA, but for AI-enabled companies, it's not just a DNA product that some of them need. Many of them need protein. Most of them need data. They don't necessarily need a physical product from us. They need that characterization data. But to enable that, we need to start from DNA. All biological material starts at that point. And so we've built out that speed, scale, and quality to enable that. Now, of course, we've talked about speed. We know with higher speed, we can evaluate more candidates in less time. That really accelerates that design-build-test cycle for our AI customers and enables them to develop more therapeutics within the same amount of time. Those are all huge advantages. And, of course, with scale, you need these very large data sets. I think what we've seen over the last two or three years is that folks would start with smaller data sets or they'd start with unstructured public data, and that wasn't good enough. They realized we need to do this at 10, 100, 1,000x what we were thinking two or three years ago. So we've made that scale. We've enabled not just our DNA, but also our protein solution side, our protein production, our data characterization delivery to match that scale because that's what those AI companies really care about. Now, of course, not only does this support more candidates, but it supports other modalities and different target classes as well. So I'll mention this more toward the end, but we think about things in terms of binders, and the field has gone from mini-binders to VHHs to IgGs, but there's a lot of other modalities that these AI companies are thinking about. And so we're just at the very early stages of scratching the surface of all those other capabilities. And now, of course, you can have speed and scale, but honestly, it means nothing if you don't have quality. Poor data, no matter how much you can make and how fast you can make it, will not be informative for a model, and it certainly will not develop a therapeutic. So you can think of quality in a number of different ways in a lot of the traditional methods that you would measure quality, but honestly, we think about it in a couple of other ways, too. So by providing flexibility in different formats or multiple production systems, now we're enabling our customers to develop their therapeutics in the context that they want. We heard it a little bit on the fireside chat, but one of the important things is there's so many different types of biologics. Each of them needs a different system, a different format. If you don't provide customers with that flexibility to order it, the DNA, the protein, the data, in the format that you need to fill your model or fill your therapeutic pipeline, then it means nothing. Again, we've got several different product lines. I'll talk a little bit about how our unique multiplex gene fragment and gene pool systems really enable new library development. And that isn't just for scale, but again, for quality, because now you can start making combinations of different products in really unique ways that no other company can enable. And so that's a huge piece, not just of a product offering, but a quality, because the design and the flexibility that a company has to make that design and then actually fulfill that is unparalleled with those products. And then finally, that fit-for-purpose downstream use. Again, I mentioned this, but the end point and the starting point is very important for these AI companies. They don't all want to follow this linear gene to protein to data. Maybe they want to come in at the protein. Maybe they want to just get data from us. And so we provide that flexibility. So somebody can start and end on that train, if you want to say, at any point. And so that's very important. So we've built this foundational system for DNA synthesis and protein production and data collection. Now, the important thing, though, is you can't attack it from one side. So saying, okay, I can build scale, I can build speed, I can make product really fast. But you also need that expertise. So you need that expertise in the characterization to fulfill the AI ML dream of data sets. And so we've attacked it from both angles. So by this I mean we've been developing for over a decade now different in vivo, in vitro antibody drug discovery platforms. To fulfill these and to deliver the hundreds, maybe thousand therapeutic programs that we've done for partners, We need to establish all of those advanced characterization methods that all these AIML companies want on the back end of their data production systems. And we've already built these out by having all the deep expertise in that full end-to-end antibody discovery platform. And so this has required a number of tools where we have things that support all of these in vivo systems, all the in vitro library discovery that we've done. And so we've developed these tools, some being shown here, and all we have to do is now apply that to the scale and the speed that we've developed on the front end. And so that's been a very seamless process, so now we can deliver the binding evaluation, the affinity measurements, the developability characterization, those functional assays, the things that are fundamental and that these AI ML companies want, but they want it in high quality but scalable data. data. And so that's really what differentiates us, is that many companies try to approach this from one side or the other. Maybe a company only has expertise in scaling, maybe a company only has expertise in deep characterization and antibody drug discovery, but we do both, and we merge those two things. So on the one side, we've got the deep expertise in antibody characterization, protein binding, things of that nature, but we're able to apply our scalable automation and operational excellence that C1 was talking about earlier. So that's how we apply it. Now I'm going to get into a bit more detail on exactly the different types of workflows that these AI ML companies are essentially ordering from us and crave in order to fill the pipeline of data that they need. So two main modalities here. So one being the library-based workflow that is very wide, the other being the clonal sequence workflow that goes very deep. So what do I mean by that? The library wide workflow means a customer might come to us with tens of thousands or hundreds of thousands or maybe even more individual designs, and they say, I want to test all of these designs. I'm not sure how well my algorithm does, so I want to test all of those from at least a very basic standpoint. So now I can narrow it down. Now at the end of that, you can see a lot of those actually feed into that clonal sequence workflow. But what the clonal sequence workflow does is goes very deep. So we apply all of those characterization assays that I was talking about to the proteins that we produce on the scale of not just hundreds, not just thousands, but maybe tens of thousands. So now, between one side or the other, you can tackle a problem from one antibody, if somebody orders it, all the way up to hundreds of thousands or millions. So being able to utilize each of these workflows, apply them at the right time and in the right sequence, is very critical to the success of these AI ML companies. Now I'll provide a couple deeper scientific examples of where we've applied this. First being on the library side, I'll have two examples here. One is where we had a customer who actually just wanted the DNA delivered. They wanted to do the screening in their lab, but they required our library technologies to enable them to even screen that in the first place. And then in the second example, it'll be a full end-to-end solution where we not only did the library production, but as I'm showing here, all of that panning and NGS output, the screening, the lead candidate selection. So here in the first example, going from one Nobel laureate to the other, this one's with David Baker, and where we collaborated and did some work for his lab, and in this particular example, the challenge was using the algorithms that his lab has developed, the protein MPNN tools and the RF diffusion algorithms to design binders, VHHs, nanobodies, however you might know them, to four different unique binding sites on proteins, and wanted to generate 9,000 unique sequences that were targeting those different sites. So we used our multiplexed gene fragment technology to make all of those libraries, put them together, and then send them to the Baker Lab, where he was able to then validate and pick out the leads using cryo-EM and different SPR techniques to measure and validate that those models were working. And so it was very critical for us to be able to use our precise printing library technology and print those libraries, the exact sequence that he wants, to be able to deliver that to multiple targets with over 9,000 designs per target. Now, in the second example, this was a collaboration with Amazon Biodiscovery and Memorial Sloan Kettering Cancer Center. And so in this particular challenge, Memorial Sloan Kettering designed over 300,000 unique sequences. And now this was very large because this was actually to an undrugged target. Very complex protein, something very complex that had never been targeted before in biology. So we needed to use a very large library to interrogate this and figure out if we had any valuable binders. So we did all the screening. We looked at 12 different populations, millions, tens of millions of NGS reads. We found hundreds of individual clones that we then selected. But we did all of this work with the library screening, not just the library production, but put them in yeast, did the protein selection, the screening, and we found a number of different very interesting and very valuable targets. Now, the great thing here is that they're coming back. So just like many AI ML companies, they realize we're probably not going to solve it on the first round. So they're using the data from this first round of output to come back, optimize it, and do a second round. now we look on the clonal sequence side there's a number of different applications here but again think people want a lot of data now we need to go very deep to characterize this to train the models very precisely the first example here being a study where we partnered with charlotte dean's lab now she's a world-renowned expert as well who designs open source bioinformatic tools that are basically used universally across the antibody industry, so very impressive. Now, the aim here was to develop a tool to predict nanobody structure and properties. Now, her lab had done this for antibodies before and had published those tools and methods, but nobody had ever done this for nanobodies or VHHs before. So we worked with her lab in order to develop the wet lab data that validated and fed back into these algorithms to generate, and we found a number of different really unique properties that were very interesting that hadn't really been realized before for nanobodies, and those have been now incorporated into that model. now in the second example here for the clonal workflow the this is again with amazon biodiscovery this time with the gray lab and so here what the idea was is we wanted to characterize all of the different models that were being published in the amazon biodiscovery website what we were able to do was take 5,000 different designs across 50 different targets. It's a lot. It's a big study. We generated 70,000 data points across seven assays for this particular one. So again, went very, very deep. This enabled them to learn which of these algorithms were valuable to predict different properties. Not every algorithm was perfect for every property, but now we know which tools to apply to which problem within that platform and we can learn and fine-tune those models now in multiple iterations of that cycle. So very valuable data set, one that's being used for benchmarking most of the Amazon biodiscovery tools and that other folks that it's open access and people able to use. And so suffice to say we're fully aligned with Amazon's mission there to build this ecosystem of AI agents and be able to empower these scientific capabilities and make them accessible to many researchers, Not just the largest companies or the well-funded AI ML companies that can design their own algorithms, but making them available to all. Now, those were some very detailed, specific scientific examples, but I think what might be most valuable here in this setting is talking about the customer journey. In particular, what do most of these companies do when they come to us with a problem, when they come to us with an AI ML workflow that they need to execute on? Now, most of these companies start at what we would consider the model building stage. So in this particular case, they might have a model that they've already developed, similar to some of the other examples that I just gave. and they'll say, well, first, we want to do a pilot study with you. We want to know that the data we're getting is going to be valuable, and it's going to fit into our models, and it's structured the way that we need. So that'll usually be on the order of tens to maybe hundreds of sequences, a fairly small study. We'll do all the production and all the characterization and data delivery for those known sequences from the partner where they already maybe have data on that, and they're benchmarking us to say, okay, how accurate is your data compared to what we expect. It's usually completed in just a few weeks and for a matter of $10,000 to $100,000. Now, once we pass that pilot stage, now we go into the first round of training. And so this is where we've got thousands of sequences, maybe tens of thousands of sequences for a single round for a single target. And again, we will go through the full make and test cycle to deliver data. And then this is completed, again, in about the same time frame. Slightly higher cost, but usually less than a million dollars to be able to feed into the model. So now this is real-world data that they're using in their design algorithm. But, again, now they need to learn. So this will feed back in. Almost never are these models perfect the first time. So usually you'll see two to maybe five additional rounds that will provide fine-tuned training for these models. So by that, we mean that it's predicting particular properties of the proteins that the company is interested in. Maybe they're assessing how well it fits also into other parameters. Once we go through this, typically we've established a really good close long-term relationship with the partner. And more importantly, we've become embedded essentially into their make test cycle. Now, the interesting thing is that then most of these companies realize, okay, I need to build more foundational models. So they'll say, well, I had this original model I came to you with. I learned a lot from it. But now I want to predict a different property. Or maybe I say, well, that went really well. I need more data. I learned from that process that I need more data to solve additional problems. And so here's when we get into that library build process. So we have tens of thousands or hundreds of thousands of sequences. And again, those are designed across a number of different proteins now. So we're looking to build generalizable models. So we want to say, okay, I'm not solving a problem for just this one property or this one protein. Now I'm solving a problem that I can apply to a wide variety. These are much larger studies, but usually completed still in weeks and for hundreds of thousands to millions of dollars sometimes. Sorry. Then on the validation of this, that's very similar to what we saw. Again, we're going into that clonal sequence process where we're making, testing, all of those. And that's so very similar to what we just saw. And again, we go through multiple cycles of that to learn and feed that back in to build these new foundational models. Now the company has a great platform. Now they have foundational models. They've fine-tuned them. They've tested them. They've learned from them. And now they say, okay, we're ready for therapeutic discovery. So now they'll apply this to a set of targets, usually not just one at a time, usually multiple. And there'll be hundreds to thousands of candidates that they want to make and test. So it's not just making tens or dozens like has been published in some. When you're making a therapeutic, you don't want to take that much risk. So it's better to make hundreds or thousands and overshoot it rather than undershoot it. And so we'll do the production. We have a second layer of characterization we'll do here. We'll get into functional characterization so we can really tell if this was an effective therapeutic for that particular application. again completed within a matter of less than a month for under a million dollars very effective drug discovery campaign we definitely will take top hits into optimization models again zero shot discovery of these therapeutic candidates is is not where it is today maybe in the future but we're still a ways away from that and personally i think we'll always want to do some optimization tinkering models whenever you're developing a final therapeutic. And so again, we'll do hundreds to thousands of these in a similar cycle. And then finally, once we're done with this optimization process, then the partner will nominate that lead candidate. They'll move those into efficacy testing and animal models and tox models at different CDMOs. And then once those tests, then they will enter the partner's therapeutic pipeline. So the great thing is that these aren't just theoretical. We've been doing this, and we've been doing it for a little while. Just recently, though, we've completed over 200,000 proteins expressed. Over 130,000 of those were assayed. This has generated 7 million data points for dozens of customers. So very impressive scale. Now, what you see here is one example of that in the data output. So this is, you know, millions of dollars on a slide, basically, in data. So a very impressive throughput and is critical for that training and process. And so the, thank you, the amount of biophysical data and information that is available for this particular data set has enabled not only model building, but also therapeutic development. And so that's the key here, is this is what we're talking about when we say we generate these large data sets, not just to help enable them to build more models, but to enable therapeutic development in the future. And here's another great example of a multi-million dollar project where, again, we're looking at 50 different targets here. So again, binding candidates, 50 different proteins. We're also benchmarking against control antibodies. So those are in red that you see here. And so we're measuring the candidates that we are testing, the AI design candidates against those benchmark candidates to find out which ones would be the best properties. Now we compare all of this data together. So again, it's not just binders. It's not just biophysical properties, but it's all of this combined so that now you can select lead candidates. So here we see in this green box, you put all of this data together for somebody to, again, train a model because the model will depend on good data and bad data, the data in the yellow and the red. But then when you're selecting the therapeutic candidate, now you can select from that green box. So that really allows people to use the most out of all of this data that we're delivering. And the last example that I'll end on here is a really interesting one because this kind of illustrates the way the market goes. So every, in the biology, every 20 years or so, a new model, a new method will emerge. So here what we're looking at is actually a campaign where we ran in vivo, in vitro, and AIML all together. So in this particular example, we have hits from each method. We took those all the way through functional characterization to find the best hits, and we found that hits from every method were valuable. So really the message here is that it's not that one method supplants and replaces the last. The traditional drug discovery methods are still very viable and very useful. AI is now just a third additional tool that we can add to the mix. It's a very unique one and the newest in this series of in vivo, you know, hybridoma discovery originally, in vitro phage display discovery, and now AIML. The interesting thing from this one is actually the partner has told us that they've nominated the lead. They're coming back for more optimization on that, but the lead is actually from the AIML library, interestingly. So that was their lead candidate, came from that design. It passed very well through the animal efficacy studies, very useful. So we know this is working. We know this is a tool that people are going to continue to use, and we're very pleased to see that, that all three of these methods can work together in concert to find lead candidates. And then, again, kind of teased it a little bit at the beginning, but it doesn't end here. So far, we've talked a lot about monoclonal discovery, single candidates, but we see a wide future in different modalities, one of those being bispecifics. So we put a lot of effort into this recently as well. So you can see here, we don't just end at the multiplex gene fragments like I showed for the Baker Lab study, but we also have the gene pools. So now being able to make a single gene of that length means now you can really uniquely pair different bispecifics together. It's a very unique thing to be able to do so that you can start, again, AI designing how these should come together. And now we can do that in that library method, but then the question is, okay, now when you want to go deep, when you want to have that characterization workflow, how do you do that in high throughput? Because bispecifics are traditionally very hard to work with. And that's where the B-Body platform comes in. So we had the licensing of the B-Body platform from Invenra, and this works extremely well in that high-throughput method to make hundreds or thousands of candidates very quickly and then characterize them very deeply and keep them in that format for your downstream manufacturing in CMC. So this is where we think that the AI ML market is going next, or at least one of the ways that it's going, one of many that we will continue to support. And so we've been very proud to support it and provide the data, the genes, the protein, wherever somebody needs to start and stop along the way. And we're very excited about the future of the industry as well and the different modalities that it's going to enable. And with that, I'll turn it back over to Emily.

Emily Marine Leproust, CEO

Thank you, Colby. So what does that mean in terms of business? I'll take a click here. Thank you. So in terms of business, we mentioned that we were able to deliver a large number of antibodies, dataware assays, a large number of data points. And to help illustrate, we wanted to share the revenues that we got from five different customers. So it's not an exhaustive list. There is a mix of a frontier AI lab, a large pharma, AI native biotech, a dry lab biotech. And what you can see is that we have to meet customers where they are. They all have different colors because we have a full money of product. And at different times, even the same company needs different product. And that has been our strategy to become a leader in drug discovery. We had to, one, have the full menu. If you want a hybrid domain, we'll do an hybrid domain. If you want a single-cell workflow, we have that. And so on and so on. So not only we have the full menu, but we have the best implementation of that full menu. And that is being seen here. And AI is turbocharging this. And so you can see generally it's up and to the right, and customers are doing things differently. And that is back to the key of what we provide is industrialization of what you want, very high throughput, but we're going to be flexible and we're going to customize our solution to exactly what you want. In terms of the market, what does that mean? So we are updating our belief in terms of what the market size are going to be. and please note that this is for 2030. And so we think that the DNA market is going to be flat, it remains at $2 billion. We're thinking that the antibody drug discovery is going to grow, and on top of it, in antibody discovery service, we have an additional half a billion dollar that is driven by AI-driven drug discovery. And then we think that the protein expression market is also going to grow. And so we think there is a big opportunity. Of note, the assumption that we have here for AI is that there is not one dominant model. This is an assumption that many models are going to win. And then the second assumption is that AI drug discovery is going to work. And we heard from Colby, we think it's going to work. We have examples of it going to work, but if it didn't work, then the market will not be as big. And last, we are very happy to share that the growth between 4825 and 4826 so far is in the triple-digit order growth. So we've talked about AI in the context of drug discovery, but as a reminder, AI is broader than just drug discovery. And here are just three examples of work that has been published by customers. On the left, we have using AI and synthetic DNA to discover, engineer, develop CRISPR tools. In the middle, the same leveraging AI for mRNA expression in cell engineering therapy, in therapy with mRNA as a modality. The promoter, the enhancers, the terminal regions are very, very important. And you have to engineer those regions. And nothing better than the combination of AI plus Swiss DNA to be able to engineer that. And then last, on the right, you can leverage AI for protein engineering. We heard from Frances that she's doing it, others are doing it, and you'll hear from Patty this afternoon, even our own experience in engineering our own enzymes and protein leveraging AI. So, we absolutely love AI-driven drug discovery. We think it's going to grow our market. It's going to be a great catalyst for twist. And we should not forget that actually the AI as a tool is useful in a much broader fashion. So with that, now we're going to hear from our customers, and I will let Angela introduce those customers. We have a number of customers. I think you're going to like it.

Angela Bitting, Head of Investor Relations

Fantastic. Thank you, Emily. We've had some great discussion about our internal platform. And now we are going to hear directly from customers. We have some prerecorded videos, and we have some customers here with us to present live and in person. Our first is a video coming from Josh Meyer, who is the co-founder of Chai Discovery, where he leads the development of AI-driven technologies to accelerate drug discovery and molecular engineering. If we could cue that video, that would be great.

Joshua Meyer, CEO

Hello, everyone. I'm Joshua Meyer, co-founder and CEO of Chai Discovery. And today I'm excited to tell you a little bit about what we're working on at Chai and how we're leveraging experimental capabilities at Twist in order to fuel our journey. A bit of background about myself. I started my career at OpenAI back on the early team in the nonprofit days. We worked on GPT-1 and GPT-2 back then, showed some of the early scaling laws, and I realized that if the models were able to learn how to speak English and speak French, why wouldn't they someday be able to learn how to speak DNA and proteins? So I went to Facebook of all places where I trained the first language models, transformer language models for protein sequences and then before starting chai was chief ai officer at a company called absi another antibody pipeline company where we also did a lot of great work with twist um another fun fact is i was actually one of the twist beta users uh uh back uh in my my academic days i was working at the broad institute uh working on gene editing and we actually tested some of the first oligol pools uh coming out of twist uh so i've been working with the twist family for a very long time and very fortunate that you know we're a continued partner of them now at chai at chai discovery we're building a computer-aided design suite for molecules and the big vision of what we're trying to do in this space is to generate antibodies that are ready to go as therapeutics if you think about the process of making a drug there's many many steps that go into the process even all the way from the pre-clinical stage you need to get get a molecule that binds to a target efficiently. So for instance, what you can see here is we might take a drug target in purple. We'll have a specific portion on that drug target where we need the antibody to bind. And then actually what our models are doing here is these are diffusion models that are placing the atoms in 3D space so that they actually bind the targets effectively. So we need to do this in a specific way. We need to do this in a high affinity way. But even if we can get a binder to these targets, there's a big difference between making a binder and making a drug. So these molecules also need to have a whole host of drug developability properties. When we produce the antibody, we need it to be specific to the target. We need it to be expressed with high yield. We need it to have low viscosity, so we can't have the antibodies self-interacting and binding to one another. They just need to bind the target. And then we also need them to be stable. Even if we can make antibodies that bind these targets and have these drug developability properties, we also need them to actually have therapeutic function. So one of the really nice things about these models is that they reason at the atomic layer. These aren't like the NLP models that I was working on at OpenAI that reason in English or reason in French. These models actually reason in atoms. And that's allowed us to come up with very specific designs. So for instance, you can see here a peptide MHC complex where we would like to get a design that is specific to a single point mutant on a cancer. And it turns out that now the models are able to do things like this. And if we can bring all these challenges together, then the outcome will be generating drug-like antibodies, including to hard targets, which are difficult to go after with traditional methods. So really opening up the surface area for the kinds of molecules that we can make and the kinds of targets where we can apply them. So how does TWIST fit into this picture? Well, if we look at how the field has evolved over literally the past year, the success rates of AI methods in molecular design have really gone through the roof. So literally a year ago, back in May of 2025, the state-of-the-art for antibody design was a 0.1% success rate, meaning one in a thousand of the antibodies that you would make would actually bind in the lab. Well, one of the really exciting things that we've seen with CHI is that if you look at the CHI-2 model that we published in June of last year, we were able to go from that 0.1% success rate to about a 16% success rate, meaning if I make 1,000 antibodies, now 160 of them are going to bind. So really, you know, an over 100-fold improvement in the number of designs that are binding the target. In order to make this happen, there's both a ton of compute power that we need in order to train our models, but then, of course, also the data sets in order to train the models as well. So this is where a company like Twist can really come in. I think Twist really has this potential to provide a ton of the data needed in this area. The technology is, again, a great fit where you can start from the beginning of actually synthesizing the DNA. If you think about what happens when we're building out a training data set at CHI, you might have a certain sequence that you want to design. We might use our model, for instance, to design an antibody molecule, we can take that protein and we can think about what the DNA is. So TWIST, of course, great technology in DNA synthesis. We can either make gene fragments, so each one of those sequences can actually be, you know, we can order those independently, or we can actually even bring this together into larger oligopools and then synthesize many, many molecules at the same time. So we can scale data sets that way. The other thing that we would like to do is if you look at the way that antibodies are conventionally discovered, you might take a target and put it into an animal to immunize it and basically have the model create antibody, the animal create antibodies against the target, which we would then extract from the animal. Or we might run a phage or a yeast display library where you might have a fixed library. So a bunch of random sequences that have been fixed that we we try to latch onto a target. But what's exciting about AI is actually now if I want to cut down those timelines and also go after those harder targets, I might take a target, generate those molecules with the AI model, and then go straight into the lab and order DNA for each of these. So whereas before I might have used the same library each of the times, or I might not have even used a library at all, you're now for every target actually having AI-driven repertoires or AI-driven sequences that we're going straight to testing in the lab. So I can imagine that this is again another place where we have a close partnership with Twist where we can very quickly send sequences to Twist on a Monday, for instance, and then a couple of days later actually have DNA synthesized so that we can go and confirm whether these designs actually work in the lab. So to give you an example of a case study of how we might use a method with these success rates in order to either create data sets or validate our models. Let's talk about that drug developability challenge. So one of the exciting things we're seeing with the latest models is that we can generate an antibody that doesn't just bind the target like I showed on the last slide, but actually has drug developability properties like thermostability, polyreactivity, self-interaction, hydrophobicity, as well as a host of manufacturing properties. And now again, and let's think about how we actually build this data set, we would take a bunch of targets for this benchmark, we generate antibodies against each of them, and now we have to, again, we go to, you know, Twist for instance, and we generate those sequences, and then we turn them into antibodies, and we measure developability properties. And one of the great things about working with a company like Twist is that there's actually expertise in all these areas. So you could actually run this entire study at Twist, right? So everything from the gene fragment synthesis to the antibody production to the measure and the developability properties, that can all be done over here. And one of the things that we've loved about the partnership with Twist is just how responsive Twist is to some of this feedback from a customer like ours. And they have been since the early days of working together. And I think that's really important if you think about where this industry is headed. The workflows that we are running today in the lab look pretty different than they were even just a year ago, right? Because if you're going from testing, you know, a system where one in a thousand of your antibodies bind versus where, you know, a hundred and a thousand or two hundred and a thousand bind, it really changes the kind of experiments that you might want to run and the feedback loops that you'd like to run. So if it only takes now 24 hours to design an antibody on the computer, the next bottleneck is actually how fast can actually synthesize that molecule and then run these downstream experiments. So we're really happy that TWIST is continuing to invest in this and continuing to make those timelines faster because that's something that can directly benefit the feedback loops through which we can validate our models and then also continue to build their training sets. Maybe lastly, the thing I'll say is that this is a really exciting area to be building in right now. I think both the advancement and the experimental methods of how fast we can turn things around and then also just the pace at which the models are developing really creates an interesting flywheel where as the models get better, there's more DNA that we need to order. And then as we can synthesize DNA faster and we can run these experiments faster, the models update faster. So then this whole thing goes into a really nice flywheel. And as a company like Chai, we've skipped over this slide earlier. The company's raised about a quarter billion dollars of capital in order to go and kind of push the frontier on these models and then take those models and deploy them in some of the largest pharma companies in the world. But in order to do all of that that quickly, we've needed to work with partners like Twist who've been able to iterate really quickly in our needs, able to push the bounds on how many of these antibodies we can actually produce, the amount of sequences we can measure. And this has been extremely productive for us because it means that we haven't had to go and develop all these things in-house. We can instead rely on a partner like Twist who has the expertise to really deliver on the scale that we need in order to push forward the bounds of AI. Thank you very much for having us, and hope the rest of the semester day goes great.

Angela Bitting, Head of Investor Relations

All right, fantastic. So Chai Discovery, they do not have a wet lab, right? You heard from them directly as to how they leverage our services, and they use a wide range of services. We've worked with Joshua for many years in different capacities, and so it was a great example of a very satisfied customer who continues to push the bounds of research. Now, I'm pleased to introduce a real live person in the room, Dylan Flood, who is co-founder and scientific director of LCBio, a wholly owned subsidiary of GSK, where he leads the development in next-generation oligonucleotide therapeutics, and RNA engineering technologies. Our objective is to show a lot of different customers doing a lot of different things. So, Dylan, over to you. Thanks so much for being here.

Dylan Flood, Other

Thank you. Yeah, thanks for having me today, guys, and I'm excited to show you a little bit about what we're doing at LCBio. Like Angela mentioned, we were a small, scrappy San Diego biotech company that started in about 2021 and in 2024 became a part of a much larger company called GSK. It's been a wild ride. It's been super fun. But I'll tell you today about our really interesting technology that we built that allows us to really increase the efficiency of oligonucleotide screening and selection to make better drugs. We did that with a lot of collaboration with C1 and his chemistry team here. It's been, as myself as a trained chemist It's been wonderful to be able to ask To pitch these crazy types of requests To see one on their team And have them come back and say You know what, that's wild, but it might work Let's go for it So with that, we've opened up some really big doors And instead of using this technology To train models for antibody-type drug discovery I'll cut to the chase We're now using it to look at oligonucleotide drug discovery So kind of a different bend but it all comes back to DNA writing. So what are oligonucleotide therapeutics? These are short synthetic pieces of RNA or DNA that are used to target RNA or DNA in the cell. What we typically think of when we talk about RNA therapeutics are things called antisense oligonucleotides and short interfering RNAs. I'm sure you guys have heard of them. They're fantastic drug modalities. And for a long time, folks thought these things were extremely programmable. So, you know, when we started the company about five years ago, there was a lot of dogma in the field, and people thought that these things were super programmable. You could use Watson-Crick-Franklin base pairing to program your sequence. You slap on a few patterns, and you're good to go. That might have worked for some of the early targets, but now that these things are going from rare disease to common diseases, right, we need better therapeutics. And one of the things that makes this hard is therapeutic sequence prediction. There's a lot of, you know, you can very easily predict short-range sequence shape and folding on kind of shorter RNAs. But once you get to full target length, it gets really hard. And that's because these things are not a one-dimensional string of A's, T's, G's, and C's, right? They're a dynamic folded structure that has secondary structure, it has tertiary structure, and a really kind of underdefined protein RNA interactome. So what we did is developed a technology that we call Roslyn. It's a DNA-encoded library-based technology that allows us to rapidly screen massive amounts of chemical diversity through our platform. And this allows us to perform our selection techniques in a system that natively recapitulates the RNA structure and hopefully that binding interactome as well. So why did we develop Roslin? It was really to get at this, eliminating what we call the oligo design question. For any one target, there's tens of thousands of sequences you can make to that target. There's tens of thousands of patterns of chemical modifications you can pattern on that sequence. And there's tens of thousands of ways to stitch these things together. I'm not a mathematician, but I'm trusting the people who put that on the slide. There's a lot of ways to put these things together. So what we tried to do was come up with a way that we could take a much larger chunk of the pie to really look at the chemical diversity that we're seeing in this space. And that's because most folks, if you look at a lot of the drugs on the market now, were discovered after focused, kind of dogmatic type drug discovery efforts where people looked at 100 to maybe 500 different constructs. constructs. And we think there's just so much more out there to explore that we can find better molecules out there, even in crowded spaces. So what we did is we developed our Roslin platform. Again, it's a DNA-encoded discovery engine. That's all super fun and great. But what it actually allows us to do is increase the amount of screening and selection we can do at these various stages by orders of magnitude. So when we look for our sequence to any target, We're now, instead of screening a couple hundred, we are screening 10,000 to 100,000 constructs in a single tube all at once. And this is how we define our selectivity and our activity of our constructs. We then take one of those winner sequences, and we start to pattern on modifications to these things. DNA is a great format for storing information and doing all sorts of stuff, but it's not a really good drug, so we need to modify it. We take our best sequence and we start patterning in our modification patterns. We can do this on the scale of 10,000 to 50,000 constructs at a time, and this really helps us fine-tune our toxicity and our PKPD profile. And the last bar yet is not yet DNA encoded, but what we can do is fine-tune those properties through optimizing the linkages that link all these nucleotides together. So I'm going to show you some data. It's a little embarrassing because this is all from five years ago. It's the oldest data we have. It's all the lawyers let out the door. But even the first time we ran an oligo discovery campaign with our Roslyn platform, we were only able to do that because Siwon picked up my phone call and said, sure, I think we can put degenerate bases in there. And that was the start of a great partnership. But what we did here is we took a highly studied target transcript. This is for TTR. It was the benchmark case for all the RNA companies out there, all nylums, rionuses, so on and so forth. And we said, we're going to go find a better sequence to target this gene. What most of these companies have done is they've all landed in a similar locus, which is kind of over by the 3-prime UTR. What we did is ran our exhaustive screen, our exhaustive search of this transcript, and we found constructs that were distal to that area that were about 20 times more potent. So we thought that was proof positive, and increased screening would give better results. What we did then next is take our single sequence. We threw away the dogmatic GATMR-type approach and started patterning in modified nucleotides into these things. What we knew about that was that that would help us modulate the protein-binding effects of these strains of nucleic acids. And what we could see here is that we could keep our high-activity constructs while really modulating our toxicity profiles. What we could then do is take that kind of second-generation lead, and we can pattern in with some really cool chemistry that my other co-founder at LC developed, we could fine-tune the properties of these constructs. I won't spend too much time on that because I want to get to what I think is even cooler, which is where we are now. Back in the day, we were producing libraries with Twist with all DNA constructs. We were then brute-forcing our Gen 2 constructs with classical column-based synthesis approaches. But thanks to another phone call that C1 graciously picked up, So we've now been able to incorporate all sorts of interesting modified sugars into our synthesis process. And what this allows us to do is really change the activity and protein binding profiles of these types of molecules. So I'll talk kind of about these ones on the right here. It's 2' MOE and 2' LNA. These are kind of the base kind of state of play for antisensitoligonucleotides out there, as well as the PS bond. But after much back and forth with C1 and his great team, we were able to get some really cool methods that could allow us to incorporate PS bonds into these constructs at high scale. I'm not talking just in a couple sequences here and there. This is across the entire chip with complete control, with as good a control as you get with regular DNA. We were then able to incorporate two things, the Moe and LNA, like I mentioned. And really interestingly, we were able to incorporate Moe at very high fidelities. This is notorious for being a hard-to-incorporate base in oligochemistry. So the fact that we were able to do it on the silicon chip was amazing to us as chemists, super exciting as a scientist, but really started to enable what we're doing next. And what that is is building fit-for-purpose data to train our own AIML models to power our kind of the entire foundation of our oligo discovery platform now. Right now, we are using this data. Like today, we're using this data to start triaging compounds that come out of our selection and screening processes for possible tox effects. This is great. This is a wonderful use of AI. It takes our screening from hundreds of compounds down to maybe 800, which is a huge win for us in time and money. That said, where I see this going is as we can continue to increase this data corpus that we have, we're going to be able to start using this in a generative fashion to pre-design the constructs we want to move forward before we even get to the screening phase. And with that, you know, this doesn't happen alone. Our small little team of 20 is now a part of a massive team of about 20,000 researchers. So it's been great to integrate into GSK. Very different than the LC vibe, but it's fun. But with that, you know, that is the LC story, and I'm happy to answer any questions if anybody has any. Just shout it out.

Speaker 3

Thanks very much. The predictive toxicology and PKPD, it seems like that's a huge potential maybe for AI, but the models haven't made a ton of progress yet, and given most failures are kind of like the phase two tox phase. I'm just curious, as someone that's worked in the field for a while, what your perspective is on the progress we've made on predictive toxicology, and if you think there's potential to improve approval rates over time, if that's an area that AI can contribute to?

Dylan Flood, Other

Yeah, I think in a therapeutic modality like this, where there are strong class effects, class tox effects, and you can boil that down to interaction with a protein or a class of proteins, there's actually a lot of promise in being able to predict that, right? Because what we're functionally doing here is taking constructs that will fold into three-dimensional shapes, exposing them to these proteins, seeing how they bind those proteins in ultra-high throughput, and using that to make predictions. And if you can basically boil down what is causing the tox, and it's something you can address, I think we're going to be able to build the data to actually predict on it. But some of these other larger multisystem tox effects might be harder.

Speaker 10

Can you hear me?

Dylan Flood, Other

Well enough.

Speaker 10

Try it again. I'm kind of loud, but you probably want people to hear on the... You can hear? Thank you for doing this. So in your presentation and in the other presentations, it's been really exciting to hear how quickly AI is leading to development of a higher number of well-profiled drug candidates in much shorter duration. I'm going to kind of actually ask a little bit of a financial modeling question, which I know is a bit unfair, but you're generating a lot of data. You're probably better than me, so don't worry about it. So you're generating a lot of data up front. Is there a point where, you know, using you as an example of an important customer for Twist, where you've essentially generated enough data recognizing how much you have, your staffing constraints, your capital constraints where there's kind of like a bolus of activity with twist and then it drops or is it the opposite where there's kind of a consistent build of demand for twist because i think a lot of us in this room we're really excited about this but i think we're trying to figure out like our customers like you going to spend a lot on twist up front and then kind of pull back for a while or does it actually amplify over time yeah i think that's a really great question something

Dylan Flood, Other

that I have to talk to our higher-ups about all the time. So what I see is that kind of far-off future where kind of the screening dies off is something that's in the far unforeseeable future just yet. That's when AI takes all of our jobs and we don't have to work anymore, right? But in the near term, what I see is that we have just started to produce models that are useful and exciting to the internal stakeholders that we're working with. We've been able to deploy these on programs, not only to just reduce talks and triage, but also go the other way and tune in polypharmacologies that are interesting and elusive and hard to get at. So what I see in the next three to five to six to seven years is a steady ramp-up of needing to build data, and this platform relies on ultra-high throughput, parallelized synthesis to do that data build. i think there could be a time in the next you know short to medium term where for a single modality we have enough data that we can be predictive um but as a large pharma you have we have a handful of modalities that we want to go after that we can't even start to think about modeling yet because we don't have the data to do it right if we're putting our flag in the ground in asos and siRNAs. This is a rapidly expanding field. There are self-amplifying RNAs. There are up-regulating RNAs. There are RNAs that interact with regRNAs. This is a whole field of emerging RNA biology that we're still going to go after, ADARs, all these things. So I see while I'm still around, I don't see that dying out anytime soon. Super helpful. Thank you. Thank you so much. I am

Angela Bitting, Head of Investor Relations

continuously inspired by our customers and the endless creativity for the next problem, the harder problem, the deeper they go each and every time. And you just heard it directly from Dylan. It's not a problem that's going to be solved in our lifetimes. So with that, we're going to change directions a little bit. We're going to move away from drug discovery and we're going to focus on enzyme engineering for sustainability. So we're really going to change it up. Our next presenter is also joining us via video because she is in Australia. Vanessa Vongsathi is research founder and head of bioengineering and discovery research for Samsara Eco. She leads development of AI designed enzymes that enable infinite recycling of plastics and textiles through advanced circular biomanufacturing. Vanessa.

Vanessa Vongsathi, Other

Hello, everyone. Thank you for having me today as part of the Twist Bioscience Investor Day. My name is Vanessa Vonsuthi. I'm one of the research founders and head of bioengineering and discovery research here at Samsara Eco. Twist is one of our longest collaborators and enablers in our discovery and scale-up ecosystem here at Samsara. So it's a real pleasure to be here today, and I'm really looking forward to taking you through what we're building. So at Samsara, our mission is to create circularity for the world's most valuable materials. And we started this mission in 2021 by targeting a group of materials that is arguably one of the hardest to ignore, and that's plastic. The world has produced over 10 billion tons of virgin plastic from fossil fuels to date, and this really isn't just a problem of waste. So every kilogram or tonne of virgin plastic begins its life as oil or gas, extracted, refined, and then moved through a supply chain that is long, carbon intensive, and increasingly exposed to geopolitical risks. So it's no surprise then that the production of virgin plastic currently contributes to over 3% of global greenhouse gas emissions, and this is expected to reach 15% by 2050 if we stay on this trajectory. despite this only 10 percent of all the plastics we produce globally get recycled today and this rate is as low as 0.3 percent when it comes to textile to textile recycling the reality is that no matter how meticulously we sort the plastic waste at home most of the things that we consume actually don't make the cut for traditional mechanical recycling usually they're too contaminated contain dyes or are mixed with other plastics and textiles have an even slimmer chance of making it through in practice it's really only the cleanest clearest plastics that enter the mechanical recycling waste stream where they're melted and extruded into recycled plastic but with each of these cycles the plastic also loses some of its material quality and strength until it eventually ends up in landfill or has to be blended with increasing amounts of virgin materials and so this is really the problem that samsara was founded to solve and it's what brought us to leveraging enzymes to deliver material circularity so this slide gives you an overview of our technology platform it's an integrated system that takes end-of-life products and turns them into virgin identical circular materials at the heart of our platform is a machine learning driven enzyme design engine the designs we generate are brought to life using twists clonal genes they're delivered to our labs ready to experimentally screen in the 96-well plate format. And so rather than cloning and sequencing genes ourselves, we receive them sequence confirmed and ready to slot into our enzyme screening workflows. This means we can move really quickly from a computational design sequence to an experimental data point at a pace and scale that just really wouldn't be possible otherwise. From screening our engineered enzymes, we take the most promising candidates forward through characterization and process integration until they ultimately feed into our chemo enzymatic recycling process that you see here on the right. As input to our process we can take colored mixed or degraded plastics and textiles and then our enzymes get to work breaking them down into their original chemical building blocks also known as monomers. These monomers are equivalent to what we have to extract from petrochemicals today which means we can purify and re-polymerize them into virgin quality materials that can be manufactured into brand new products importantly this really enables infinite recycling and so we see no loss in the quality or the integrity of the material no matter how many cycles it goes through so at the core of everything we do are our enzymes and specifically these are new to nature enzymes engineered to break down plastics at speed and scale. These aren't just enzymes borrowed directly from nature and dropped into an industrial process. While some naturally occurring enzymes can degrade plastics, they rarely meet the demands of a real production environment. Often they're too slow, too unstable or unable to withstand the operating conditions that we require. And so we use natural enzymes only as a starting point and a foundation that we can then build on to create proteins that are optimized for speed, stability, and precision at scale. What makes this challenging but also exciting is the sheer scale of the protein design problem. So to give you a sense of this challenge, a typical enzyme with just under 300 amino acids has more possible sequence combinations than there are atoms in the observable universe. And so the question we face in enzyme design is how do you search that vast space efficiently to find the very few sequences with the properties you actually need for an industrial process as you can imagine searching through by random trial and error would be near impossible and so we need smarter more principled ways to navigate our search and this is what our platform is built to do it does turn out that one of the most powerful approaches we can take is to learn not just from the enzymes that exist in nature today but actually from their entire evolutionary history like species enzymes have a history that stretches back billions of years and along the way countless enzyme variants have existed and disappeared and so using a technique known as ancestral sequence reconstruction we can actually infer all of those earlier sequences and bring them back into view this matters enormously for our understanding of how different protein families work that also gives us very rich training data so rather than only having a few hundred natural sequences that are relevant leveraging this method gives us tens of thousands of ancestral sequences that allow our deep learning models to identify the patterns that better link things like protein sequence sequence to protein function activity and stability and so a really great illustration of this is our nylon 66 hydrolase which is estimated to sit at about 10 to 82 possible sequences away from the closest naturally occurring enzyme and to our knowledge it was the first ever enzyme that was characterized to be capable of degrading nylon 6-6 and so this is a break breakthrough that was really only reachable because of the richness of this evolutionary data and so if we compare mechanical chemical and enzymatic recycling approaches side by side the advantages are clear so our process enables a true closed loop circularity returning plastics all the way back to their original monomers with no loss of the material quality. It handles the mixed plastics and composites that other technologies reject and it operates at a low carbon footprint relative to virgin plastic production. And critically we've demonstrated that the monomers we produce can be re-polymerized into virgin identical end products that look feel and perform exactly the same way that fossil derived materials do. Our milestones to date reflect the real-world traction that our technology is generating on the product side with our partner lululemon we produced the world's first nylon 6-6 enzymatically recycled nylon 6-6 garment and launched a full retail collection with them we've also produced a clear recycled pet bottle these are consumer facing products that prove our materials are virgin identical in every sense that matters the industry response has also been equally strong so lululemon have committed to sourcing 20 of their fiber portfolio from samsara over the next 10 years and lskd another athleisure brand founded here in australia have also followed with a long-term agreement and we've also announced a polyurethane collaboration with the lycra company our first facility is now open here in australia as you can see in the top right and our commercial plant is on track for 2028 we've raised 107 million dollars to get there and so the technology works the market is ready and we now have the backing to build this but plastics are really just the beginning at samsara we believe biology scaled into industrial processes is one of the most powerful tools we have for addressing the material challenges of our time our platform is built on the marriage of protein design process chemistry and engineering and it's that combination that makes what we do differentiated we built this platform for plastics but it's designed to go much further the same approach of co-designing proteins and processes at scale applies equally to other polymers green chemicals critical minerals and and even carbon capture. We're not building a single product company, we're building the infrastructure to scale biology into industry. So thank you so much for your time and for the opportunity to present at the Twist Bioscience Investor Day. On a personal note, I've been working in this field for nearly a decade now, and the difference that Twist technology has made to the pace and discovery of innovation is something that we genuinely feel firsthand at Samsara every single day, and it's just fundamentally changed what's possible for companies like ours. So, if you'd like to continue the conversation, please don't hesitate to reach out, and you can also follow Samsara Eco on LinkedIn. Thank you. So, for those

Emily Marine Leproust, CEO

on the webcast, we are coming back from lunch, and for the second part of the day, for the afternoon, we are going to start by a focus on NGS strategy. That focus will be in two parts, and we will start first with our SVP of Product and Marketing, Jimmy Jean. Take it away.

Jimmy Jean, Other

Thank you, Emily. I might be measured by Angela, how many of you I put to sleep. So stay awake for me. I'll kick off this afternoon starting with NGS portfolio, and then I'll dive into MRD specifically. Coming back to the slide, I'm part of a commercial execution, and I'm also part of the NPI machine that I twist. So my focus, like I said, is NGS and plus MRD. So we are a 200-plus million NGS business. In the last five years, we grew on average about 30% per year. For those of you who follow the industry, post-pandemic, the industry grew low single digits. So how do we win and get to the 200-plus million today? is through all of these things on the screen. Quality, workflow, rapid customization, not going to bore you reading through every line, a full portfolio product. But I want to pick up where C1 left off that mentioned, right, on the bottom left corner. We sequence 20,000 samples per day for our gene production process using NGS. and you do the math six to seven million samples per year so we know NGS and we know what we're doing here and that's all you need to know on this slide where do we sit in our portfolio pre-sequencing we are compatible with all sequencers out there we are specializing from sample to that library ready to load on the sequencer and that's where we play our bread and butter start from target capture what is that that is you sequence only the things you're interested in and we enable our customers to do that over the years twist has built a very strong reputation twist panels and twist probes means high uniformity it's low sequencing cost and more importantly often, we enable our customers to fail less of their samples because of lower coverage. And for those who lab operations know, lower failure rates makes the world a difference. Maybe John can attest to that. But you might ask, why is this even relevant? Whole genome sequencing is taking over, no? Why target capture? So let's do that math together. So in 2021, per gigabase of sequencing cost $6, the highest throughput. Today, the number is $1 to $2. So cost of sequencing dropped up to six-fold in the last five years. Now, before we say everything goes to whole genome sequencing, you look at a typical cancer panel on the left in the top chart. From a typical panel to go to exome, it's a hundredth time step up of sequencing data quantity required. From exome to genome, it's another hundredfold of sequencing data required. And so if you take six-fold of reduction in sequencing cost and you say I'm moving from panel to genome, you are going to pay 1,500 times more in sequencing cost than you paid in 2021. And that math does not add up. Now, that is not to say whole genome sequencing. By the way, WGS stands for whole genome sequencing. sorry, I should point out that last slide, shouldn't say whole genome sequencing is not relevant. It's very relevant, and therefore, we're fully embracing it. In the past five years, we have built up a full library preparation portfolio that enables us to capture that whole genome sequencing market. In fact, this portfolio specifically currently grows 56% year-over-year for us. So why does that enable us to win? If all you got is a chip, why library prep? Well, look into what's in library prep. It's buffer, it's DNA, and it's enzyme. Buffer is just water and salt. Everybody has it. By now, I don't have to say DNA, and I don't even have to say enzymes and enzyme engineering, because speakers in the morning has told all the stories. Now you realize we have every reason to play and win in this space. One billion dollar market, a lot of room for us to continue to grow, gain, and win. Whole genome sequencing, we offer the highest throughput, meaning you can put 1,500 samples in a single flow cell or a lane to run whole genome sequencing on it. Going beyond whole genome sequencing, you can put 3,000 samples Again, industry highest on a single flow cell or a lane on it. And to really enable super high throughput sequencing, we did something that's not done by anyone before. And that is putting 1,100 samples in a single 96-well plate. What is that? Previously, you would need 12 liquid handlers filling out the entire lab. With this technology, you can go with one. Because we put 12 samples into a single well in the very first step of lab processing. What do you have to do? Two things you have to do in that one step. Adding DNA barcode or tagging so you know where that DNA is coming from when you do sequencing. Two, you normalize those 12 different samples to an even spread so that when you're sequencing them, those 12 samples get reasonably equivalent coverage. And that is very difficult to do and is enabled by what we call a normalization by ligation technology. Everything is done in one step, therefore 1,100 samples in a single 96-well plate. So, Patty is going to talk about our enzyme engineering and how that enables high performance, Just, again, this is an exciting part of NGS portfolio for us. As if you haven't heard enough about chips, I'm going to go there yet again. Largest, a lot of the largest clinical labs really trust Twist. Why is that? Do we just do better? But sad to say, or glad to say, the reason come back to the chip. So bottom left, up to 1 million oligos. Many thousands of clusters, what we call it. You can understand as wells, if you may. And in each well, 121 unique different oligos in a single well. Our competition, what do they do? They use 96 or 384 wells. to do the same thing. So I'll use a typical common scenario. You want to build a panel of 50,000 probes. Again, typical. How do you do it in the 96 or 384 web plate? Well, 384 at a time. It requires 130 synthesizer machine, or less machines, but different batches, adding up to 130 different production runs what do you hear when you hear that variation variation variation what is it for us you carve a small corner out of that chip the entire 50 000 in one machine on one chip in one batch all done together no variation no variation no variation and that's not the end of it. If you do it, 50,000 wells in 384 plates, guess how do they come together? And you're going to pipette 50,000 times to bring those 50,000 different wells together. For us, 121 unique oligos are already in one well, 120 times less pipetting. What do you hear? Consistency, consistency, consistency. From lot to lot, and that's what our customers counting on Twist to deliver over and over again. That is the unfair advantage. It's very, very difficult to overcome. So it's not just all about technologies. I want to go deeper into applications. We use liquid biopsy as a field. As many of you know, three key applications in liquid biopsy. And again, many of you know, we are a leading supplier in early cancer detection through our methylation detection system. We are a leading supplier in therapy selection for liquid biopsy through our custom panels for comprehensive genomic profiling. And we are a key supplier, not leading, yet not MRE. And so how we play in this space, in liquid biopsy, many companies, many labs are in this field, right? We are first and foremost making clear to our customers we do not compete with them. We are not in the horse race. And we don't bet on a single horse. But we work very, very hard for all the horses running on our platform for them to have a better chance to win. Because when they win, we win. And that's how we operate in this field. Going into MRD from this point, clearly their work for us to do in MRD space, as you saw from the pre-slide. I want to step back from the technology, from the application even, just talk through the patient angle for MRD. Linda, 63 years old, retired two years ago after close to 40 years of teaching. Two and a half a month ago, diagnosed with colon cancer. A month ago, had surgery, removed, confirmed stage two colon cancer. One month post-surgery today, she goes into the oncology office, oncologist's office, and asks the question, doc, what now? And the oncologist will wait for an RMRD results to tell Linda whether you can go home all done or more chemotherapy is needed. And that's the number one utility of MRD, which is a treatment decision at that point. My mother went through kidney cancer and breast cancer two years ago. Surgery, all done, fully recovered. But I can tell you that week five, you go into the doctor's office, you try to find after surgery, are you done or more things are coming. Every day matters. You can't sleep. And for this first utility, you are racing against time. And it's not easy, even though you think there's four weeks of window, for you to take action for Linda to know what next. Because these tissue samples need time to be put on the glass, shipped to the clinical lab, accession, doing exome or whole genome sequencing, do the analysis, do the panel design, order the MRD panels, processing MRD. Next thing you know, the window is closing quickly, and Linda will still work waiting, right? And that race to speed is a fundamental need for MRD. So the second, which is good news, Linda's good. The oncologist said, you can go home, just come back every three or every six months to monitor, making sure it does not come back. So month seven, she comes back, she does another MRD test. This is the second utility of MRD to really monitor disease recurrence. In this setting, it's no longer about speed. In fact, it has nothing to do with speed anymore. It is all about sensitivity. Because different sites metastasized require different levels of sensitivity to detect. And you want that test to detect every way of recurrence. And so with that patient angle in mind, we'll look into MRD Express from Twist. So to date, prior to this, it takes us a week to get an MRD panel produced and shipped to our customers. Not too bad. But we just went through with Linda. That's not good enough. And therefore, we completely redid how we make MRD panels. Pulled just about everything back to the single thing we're really, really good at. That is the chip. The entire MRD panel production now happens on the chip. And that enables us to do it in a single day without any compromise on sensitivity. We're known now to enable ultra-sensitivity without losing that, with no compromise, one single day. We solve that first problem. Now, moving on to the second problem. Many of you follow the space. There are many different tests out there. There's something about fragments to analyze that, using AI to analyze whole genome sequencing results. And they have fixed panels, methylation, multi-omics markers to detect MRD. In the personalized space, you have digital PCR to try to measure structural variations. You have Amplicon panels trying to do the same thing. And you even have technology to try to enrich patient-specific variants. So many technologies. I am so confused. Which one do we choose? This is all about sensitivity. And we want to have you see the way we see it. People often equate, look at biopsy, I'm already included, to looking for a needle in the haystack. Isn't exactly accurate, right? Because patients on cancer DNA and their normal DNA in the blood, 99% of the time, they are identical. which means you are actually looking for a needle in the stack of needles. And so the first question, forget about all those technology, you ask yourself to say, two scenarios. I tell you, there's one million needles. There's one that's different. Find that for me. It's not going to be easy for you. The second I tell you, same, one million stack of needles. Look for a needle that's five times bigger, and part of that needle is black. you now have a much better chance of finding that needle because you know exactly what you're looking for. And that is the difference between tumor-informed and tumor-agnostic, period. And so now you know which one is superior. Now, how do you further enhance sensitivity? Now I'll give you scenario number two. There are three stack of needles. In stack number one, one million needles, there are 16 special needles that look different In the second stack, there are 64 of them that look different. In the third stack, there are 2,000 that look different. Now, all you have to do is go up, choose one pile, and find two needles that look different. Which pile would you go to? Pile three, where there are 2,000 of them. I only have to find two. Congrats, you just figured it out. The size of the panel truly matters because that gives you the ultra-sensitivity and twist technology enable you to look for that 2,000 needles in that stack of needles. So, sensitivity and speed, as you here heard, is taken care of by our platform. I purposely skipped the other element that's very important, which is cost. But cost is a solved problem for twist. At the same time, the cost factor itself has very little to do with Linda. It has a lot to do with clinical labs, maybe margin of operations. We care about that, too. But speed and sensitivity really change Linda's life. And so, you know, we're great. Hopefully, you're convinced. But we want to take a step back to say this is not about twist. and for MRD field in general. Right now, as we say in this period, currently about 1 million MRD tests is done per year, just about. And it's growing about 50% year over year. So in two years, someone does the math, I did it for you, 2.3 million MRD tests. In an ideal world, every patient gets four tests potentially, but right now it sits around two or slightly over two. So that is 1.15 or more than 1 million MRD panels that needs to be produced. For us, every panel is how many targets? 2,000 probes. And you do the math, we have to produce 4.6 billion oligos just for the MRD field if we were to meet all MRD needs in the field in two years. 4.6 billion. Go to C1. C1 says, no problem. All I need is four riders. You've seen quite a few there. That's all it is for us. Now, if you look the standard oligo production way, which also is playing in this field, same math. They are not doing 2,000s. We say they can do 10 times less. So only 200 needles for them to do, which translate into, instead of 4.6 billion, they only have to do 460 million oligos. One-tenth of us, easy to do, right? But if you do the math, that translates into still 1.5 million oligos per day. What is a typical vendor for primer oligosynthesis capacity right now? 150,000 oligos. They can't do it. They can't do it until they grow 10 times. And that end picture is super important. We have that future-proof capacity figured out as of today, right now. And that's what we have. And for the Oligo vendor, we'll lend them a hand. We'll make sure this capacity issue for them in a couple of years, they don't have to worry about it. Thank you, everyone.

Emily Marine Leproust, CEO

That's right. They won't have to worry about it because we'll take it all. So next, thank you, Jimmy. That was great. Next, we're going to hear from Paddy. He's going to continue our strategy talk in NGS diagnostics with some details around our enzyme engineering and our nucleic acid therapies.

Speaker 14

I'm going to risk some new technology here. All right. It still makes me sound funny. Is it working? Really excited to present to you guys. Thank you. So I'm a DNA chemist by training. Lapsed, kind of. A number of different jobs I've had over time. I'm a somewhat grizzled veteran commercial leader educated in the school of life. I'm going to talk to you about what's cool in enzyme engineering, what's cool in sequencing, what could be coming in nucleic acid therapeutics. And for those in the room, they're going to understand this really well. I'm going to talk to you about commercial execution. This is my favorite slide deck ever and outside of my family, this is my favorite stuff to talk about. Second thickest accent in the company and all I can advise you to do is listen super fast because there's a lot of content coming. Enzyme engineering for NGS. Why do we care? We're a DNA synthesis company. But I think you've heard enough coming through today about why downstream of the chip matters. And I certainly can't talk with the level of intelligence that our customers have demonstrated around what's coming with the AI wave. But then we start to look at things like what our customers also do with our product. For example, in target enrichment and sequencing workflows, or how we use enzymes internally. On the tour, I'm sure you would have seen some enzymes being used to produce product. It's ever so important that we have a best in class enzyme portfolio. It matters for our customers. I'm going to explain why for the kits that we produce. It matters for us internally. Control of supply chain matters. Improved economics. My guess is not everybody in our competitive field loves our emergence in the space. So being in control of supply chain is incredibly important. We're fast, but also by being in control, it allows us to have a global strategy for commercializing product. Make your own stuff. The economics are more favorable. It allows you to sell all across the world through different distribution channels. And it allows us to truly enable the global community to get onto our platform. And the reason we like this space is we have an advantage. And I think you've seen it many, many times coming through the customer presentations. We can make a lot of DNA. We make it in parallel. It's incredibly high quality. The economics are enabling. And when you partner that with our application expertise, and I'm sure you saw on the tour, we do a lot of sequencing. Tens and tens of thousands of samples every single day being sequenced to make sure they're in good shape, to go in the right tube, to go to the right customer on the right day. So we've got this incredible throughput, incredible speed, the ability to make enzymes ourselves and screen into an application that we understand is incredibly enabling. And it's so important that we keep all of those pieces rolling together to drive great product out to the market and deliver good infrastructure for our own internal services. And if you go one step further, as we go through a cycle of design, build, test, learn, launch, we just continue to improve. We get faster, we get more effective, our knowledge continues to develop, and we launch really interesting products. And if you look at the workflow, you've seen it before, we're going to start with an enzyme target. We're going to use zero-shot design. We're going to lead off into the twist gene synthesis platform. I'm sure you've got deja vu all over again around this slide. we're going to crank out the sequences of interest we're going to go into high throughput protein expression purification into a screen where we know our application expertise is incredibly enabling we're going to pick out the enzymes of interest with features that we like we're probably going to turn the crank on this a few times because this does have some runway and so we're going to really optimize what we're building to create super products that when we've found the features that we like in screening we're going to oh, hang on, we're going to actually drive into making kits. If you think about our overall, how do we launch products? How do we drive our MPI machine? I'm sure you've seen so far, well, you've just had Jimmy present, who's full of ideas. Dr. Chen, wherever he is, is just an idea monster, and our tone at the top. We know what our customers are doing. We have another chemist that leads a company that has intimate knowledge of what's going on out in the field. We're constantly ideating. We're into our enzyme engineering platform to try and build out the enzymes of interest that then go into our product development pipeline, driven hard through MPI, and ultimately out into launched products. So pretty straightforward, right? You can see the molecular advantage we have and how we drive that into a new product pipeline. So I'll take just a couple of case studies. So for anyone that's a close follower of the sequencing world, There's a couple of important enzymes. We'll start with ligase. Now, I'm going to be a little bit honest. I'm kind of pleased that Dr. Arnold's gone. Talking about enzyme engineering in front of her, I was kind of nervous, afraid, and excited, all rolled into one. So whatever that is. But ligase matters in a sequencing workflow. You have to get the template that you want to analyze, modify, you have to ligate on adapters to get it to stick on a sequencer to allow you to characterize. Ligase is a crucial component of that workflow. And so you can see the methodologies we used up above, and we challenged C1 to say, right, let's make a ligase that's best in class. Let's get something that works with low-input templates, something that's good for cell-free, something that really drives conversion efficiency, and the way I think about it is no molecule left behind. There's nothing that's left in the tube. It all goes on to the sequencer. And also, an important point here, buffer sensitivity. Customers do the strangest things with your products. So building a system that's robust and can tolerate the variability in behavior is very, very important. And so again, LLM-based design, we're basically a high-through expression screening, and boy, did we find a really genuinely interesting ligase. And I'll save this for the exam that's coming at the end, so everyone better be paying attention. but the most important point here is we just draw the bottom left corner when you look at conversion. The twist is a big green bar. Big is good. What that means is the conversion rate of template to something that can be sequenced. That is lights out versus a competition. And then also importantly, if you go to the bottom right corner and you just look at how does the enzyme perform in a broad range of salt conditions, i.e. when the customer's got some pretty sketchy samples going into a workflow. Does it behave well? And again, conversion matters. No molecule gets left behind. Look at the behavior of the green line versus what's on the market today. That means you're going to yield a very, very competent, capable product. Flush with success and not wanting to let Dr. Chen rest on his laurel, we thought, okay, ligase is important. High-fidelity of polymerase is next. It's pretty obvious, right? Even I could build that product portfolio out. So we set off with the same challenge. GC bias is a problem. We know that. We use polymerase to assemble genes in the factory here, so on the other side of the business, but we also need to QC genes to say, okay, we can ship this and collect revenue. It's a similar strategy. Get out there. Let's get in there and use our LLMs to teach us what makes a good polymerase, high throughput synthesis, expression, characterization, or ACID to learn how it behaves in application, space-filling design of experiments to optimize the buffer. Again, same thing, broad range of inputs. You need something that works in the customer's hands. And lo and behold, out comes an exceptional polymerase. And I'll draw your attention to just a couple of points. You just look up in the top left corner, you can green for twist, or go, or yes, purchase order, yes, please, you can see improved performance at the edges of GC contents, both sides. You can see an error rate that's incredibly tight, incredibly high fidelity. And in general, for high-fidelity polymerases, it's hard to get them to behave in a product. You have to get it just right. That is a beautiful performance from our polymerase. And then most importantly, comparing to polymerase Q and I'm very pleased to be anonymized what that polymerase is but any enzyme fans out there may know you can see slight improvement and by slight improvement I mean when we use that in our gene factory, if I take that top result there, we can't ship that product right, that's the wrong gene so I've spent a bunch of money and I can't ship the product so for those of you that haven't worked with Emily that's a bad spot to be in really bad if you use the twist polymerase what it does do is it allows us to read through the complex sequence so then when we QC we say oh yeah the product's there so now I can ship and I'll still get the phone call so we're revenues but at least we ship this product if we just pause for a second it looks incremental I don't feel anybody getting the purchase orders out to buy the polymerase but let's just hold off and think that one through a little bit so we moved approximately a million genes, roughly, last year. Ballpark, plus or minus a little bit. If I can get a 5% improvement in terms of polymerase performance and acceptable performance, and yes, this is good to ship because I've got a polymerase that works and reads through difficult stuff, allowing me to QC and release product, that's quite useful. Not to mention the fact that we're also vertical now in supply. I don't need polymerase Q or polymerase K. This is our polymerase. And that is a very robust position to be in. I'm going to segue over to the other side of the screen here, and we're going to look at how our polymerase performs in sequencing experiments for customers. It allows them to amplify and access parts of the genome where others' polymerase will stutter. That's a problem. You sort of get constrained in terms of how good your whole genome sequencing experiment is. And that's something we're working hard to improve upon And core enzymology, novel differentiated features, are fundamental to driving that behavior and into those products. So I still don't feel like I've sold you anything. So let's go to the next slide. And we took these core enzymes and built these beautiful kits around them. The TrueAmp library prep kit and the PCR-free whole genome sequencing library prep kit. Really elegant products. Best-in-class ligase, high-fidelity polymerase. So you can see the incremental improvements. If I think about the customer, what does it mean for them? Think about what it costs to press the start button on your sequencer. It doesn't matter which sequencer. Press the start button. It's a well-finished BMW. So utilizing that sequencing real estate matters. Now, if I'm in a commercial setting, independent of which workflow I'm using here, If I get a 5% improvement in the number of samples I can sequence, or a 10% improvement in the number of samples onto the sequencer, the impact to my business, the impact to my research is greatly improved. And that's the bit that matters, right? At scale, these incremental improvements lined up give you absolute success in your lab. And again, we've probably heard that a few times today. We'll meet the customer where they're at. We don't make kings nor queens, but we're going to enable each of the segments to do a good job. So again, PCR-free, and so by definition, no polymerase. Or our library prep, TrueAunt library prep kit, that utilizes the polymerase to kind of give you the best of both the worlds. Lower input, but still getting incredibly good data out of that, or out of that workflow. So that's C1's two for two. So now he's resting on his laurels. And so now what's coming is, what other enzymes do we need to build out into the portfolio? And methylation is an important marker in the oncology space, but there are some constraints on workflows. Bisulfite treatment, you know, it's been around for quite a long time, and it's a very, very useful workflow, but to industrialize that, put that into a facility that's running hundreds of thousands of samples a year, it's a little bit of an art form, and it struggles with low-input quantity of template. It destroys the template, basically. And now that you've got the emergence of enzymatic methods, which are good, but have room for improvement, and, you know, you want to really think hard about supply chain. So now we've had C1 and the team working hard on a cytosine deaminase that makes the difference or detects the difference between methyl C and C. Again, important methylation marker. Same idea that we've gone through before around the utilization of AI for design. We've come up with an enzyme that 42% of the residues are changed from wild type. Now, I do like myself a good dose of evolution in our platforms, right? And there's no better advocate for evolutionary methods than Dr. Arnold that was here earlier. And imagine that, 32 million oligonucleotides every single day, the ability to screen for features of interest because we know our application. And think about being able to play in that space and start to deliver enzymes and reagents with features that are coming from remarkable primary amino acid sequences, that's true enablement, and there's differentiation here and more differentiation to come in the future as we build this portfolio out. So just in case you hadn't noticed, quite excited about the enzyme portfolio and how we can utilize that internally. So a bunch of enzymes is fine, and you've seen how good we are at making DNA. The quality system that stands behind making sure the right nucleic acid, right tube, right customer at the right time. But you also have to have an MPI machine that's sustained by a quality system that scales. And we have a unique challenge. You have the wild west of the gene synthesis community. They need a molecularly pure product. But if you're a researcher, my guess is you're not too worried about ISO 13485. But then on the other side of our business When you go all the way through to a clinical lab Or even as you start to think about nucleic acid therapeutics The quality system has to scale up So you've got sample extraction, library preparation, target capture, sequencing on any platform The quality system to wrap around the products that you'll see And there's some examples up in the back there Has to stand behind that customer base Something that's very understated in the company James gave a beautiful presentation showing the fine details of how we move a piece of DNA around the building that underpins a quality system that supports our diagnostic customers in perpetuity. We audit incredibly well and it's actually become a weapon of offense to help support, sustain, and help our customers grow. If I just pause for a minute and start to build our segue over, Jimmy's presentation was beautiful earlier. It's very hard to follow Jimmy from a presentation standpoint and he also personalized the topic. If you look at the continuum of cancer care and research, and Jimmy is far more articulate than I around that continuum. But obviously you've got your screening, early cancer detection, you have a bad outcome, potentially into surgery, you've got your molecular residual disease test, therapeutic intervention if required, and then continued monitoring. And this is what excites me about Twist Beyond Belief, and we're just at the very beginning of this on the NGS application side. From a screening standpoint, the product portfolio is incredibly strong. as you go through early detection you know this efficient use of your sequencer really efficient and effective target enrichment is enabling whether it's whole genome well exome methylation markers, liquid biopsy or you know comprehensive profiling of a tumor Jimmy talked about MRD and our capacity and the emerging trend of sensitivity and the impact it's having to earlier detection of recurrence of disease, massively impactful and a growing body of clinical evidence saying that's good for the patient, and then through into, obviously, treatment response and monitoring. And just walking along the bottom here, that is an exquisite collection of products. And if you think about the future of precision medicine, this platform's incredibly well positioned today around the sequencing side, the diagnostic side, what we're going to enable in MRD, and then ultimately, you heard how the pieces come together. On the DSPS side in drug discovery, the platform's incredible. Most new therapeutics are biological. And so at the risk of turning it into a seminar, where do biological molecules start? A piece of DNA. It doesn't matter the mode. It starts with a piece of DNA. And if you want true precision medicine, you're going to need a lot of different DNA sequences. So we're excited about the therapeutic pathway. Clonal genes are antibody discovery capability. IgG characterization shortening the time of drug discovery against the most complex of disease. That's a very, very powerful offering. And not to mention what we can do in the mRNA space. So you can see we're well positioned through that continuum of care where the technology and the products enable the community to address some very, very difficult challenges. I'm going to dream just a little bit, just a smidge. Maybe AI won't get that like a mulligan on smidge. I'm going to dream just a little bit. but having spoken to the lads at Onco earlier on they're not really dreaming that much but just imagine the situation where we'll talk about nucleic acid therapeutics it's one of many potential new growth frontiers for Twist so let's pause for a second the infrastructure that's been built out in CROs, CDMOs is built up so you can spend approximately a billion dollars to build out the infrastructure to make a nucleic acid therapeutic at scale. If you need to make kilograms of nucleic acid, that's what it takes to make it happen. But we think there's an opportunity in building out. As precision medicine matures, it's no longer about, or it's not just about, the keg or the swimming pool scale manufacturing. It's about building out across the way different sequences at smaller quantities to help challenge a well-characterized disease. And so that's the need is capacity, scale, and economics to enable needle-to-needle success. Needle for liquid biopsy to start you into the standard of care. You go through the journey that Jimmy had described earlier, and I tried to copy my slide, back to a needle in the arm with your personalized therapeutic to attack your form of cancer. That needs built out. And so just to capture and picture what I was trying to say, we talked about the workflow, we talked about speed, we talked about economics, we talked about really the impact that that's going to have for patients. And so the question is, what's that going to take? And I'm not about to claim that the problem is solved, but it's going to take affordable price. A therapeutic that costs a million dollars for each individual, that's not going to happen. It's got to be high quality. I'd like to have the right sequence shot into my arm at the right level of purity, please. It's got to have scale. So WHO is predicting, I think, I might be off by a couple of million, but it's about 22, 23 million cancer cases by 2029. We've got ever-improving diagnostic tests, disease monitoring. That patient population is going to continue to increase. So you need scale to get hundreds of thousands, millions of doses out to the global community. It can't just be a medication for a few people. You need speed. Now, I'm horribly underqualified to talk about neoantigen escape. And I'll leave that to some of our customers to talk about that. But the point stands. The longer you wait to get a therapeutic into an arm of a patient, the disease is changing. And ultimately, you end up treating something where you're looking in the rearview mirror rather than treating the disease they've got at the time of therapeutic injection. And then complexity, I'm sure you heard we launched a complex product just recently, expanding the sequence space, the number of sequences we can accept and deliver to support an even broader population. And that product portfolio sits quite well. So if you can think of a company that can make millions of genes, doses, per year, that has form for putting the right nucleic acid, the right concentration, in the right tube, shipped to the right customer, on the right day, at economics that are truly enabling, then I would challenge that, I'm not saying that the biological challenge is fixed, but if you think about distribution, if you think about molecular quality, that puts a massive dent into the challenges facing companies in this space today. And so we're writing the future of nucleic acid therapeutics. Dylan, your stuff was brilliant. I cannot tell you how much I enjoyed that from about 100 years ago, learning about antisense oligonucleotides as a PhD student. It's brilliant to see what you're doing. It's absolutely incredible. If you remember the key criteria to deliver on this promise, it has to. It has to go through here, because there's no one else with distribution or scale to make it happen. My next favorite topic, commercial execution. Emily said it very eloquently this morning. Twist, our due north, we are in the business of delighting our customers. And Paul is going to talk about culture in a little while. And it's one of these things to me that really matters. When we started the commercial side of the business together, I want every single twister to care about a customer's scientific success. That is absolutely it. No compromise. If you don't care, you don't belong here. I don't have to explain the economics of a business with high customer retention versus not. I think Bain Capital describes that better than I ever could. Every twister cares. Our tone at the top, and we mean this. We're going to play in markets where there's going to be a number one and there's to be no number two. Old school methods have made their contribution, they've been super, but now with speed economics throughput, they're holding the community back. It's time for this platform to be worldwide. From a sales standpoint, our philosophy, our commercial execution, there's one way to go And that's up and to the right. We've had 13 sequential quarters of letting it rip. And that matters, not just for the business result, but from a culture standpoint. Do people care day by day, hour by hour, when they execute? And the answer is yes. That's why you're here. You're drawn to this company, because we know the importance of our technology to the global community. Philosophically, we have an OEM strategy. We sell product to people we see in the field. We'll sell to our competition. Now, there's two things that go into that. First of all, the economics matter. But secondly, it's a twist sales team. I expect the twist sales team to win every deal. So if the other team has a chance to use our platform, that's great. We'll sell the product. And I expect the twist sales team to outperform anybody. In fact, at the risk of being complimentary, I would say I would put that team up against anybody else, and I would expect a very favorable result. As a salesperson, you are where your numbers say you are. You are what your numbers say you are. If you're behind, we're going to muscle in to help you. If you're ahead, guess what? We're going to muscle in to help you, and we are going over the top together. As a commercial leader, if you don't address your commercial problem, you become the problem, and I will manage that. But on the other side of that, we reward performance, and we reward performance well. But if you're not performing, it's much better that you work for the competition. And when we look at our channel strategy, it's kind of omni-channel. There's direct sales, there's key account management. As our accounts have grown in size, the needs change. It's no longer about selling a product. It's truly the product, product quality, scaling, supply chain, procurement, quality, regulatory audit. It's a much, much bigger challenge that takes a certain skill set to drive key account management. And then we've got a strong team out in academic and government sales taking the word of twist out all across the community. We have scaled growth channels. There are certain areas we can't afford to send to sales rep. So you've got inside sales, an exquisitely well-educated team, great farm for potential future account and key account managers, and we have digital channels that continue to improve e-commerce, punch-out, API-based ordering to make it easy to interact with Twist. Due north, we are in the business of delighting customers. And then we have our OEM channel, which expands our reach out into areas where we may not be strong, which is very complementary to what we do. But the multi-channel model allows Twist to serve everyone, global community, as soon as reasonably possible. And this is where we're excited about the business.

Emily Marine Leproust, CEO

It's diversified.

Speaker 14

It's robust. If you look at some of the simple things here, customer types, that is a broad range of customers. Large pharma, biotech, big tech, academia, government, AI native. I don't need to read the slide. diagnostics, and we see what's happening in the cancer space when you see the flow of cash into the area. Our demand drivers. It's an all-you-can-eat DNA buffet. The applications that continue to develop when you partner with the community, the next great ideas that our customer base is using on the platform. It's incredible. It's fast and it needs our scale. And obviously our funding pools. It's not too difficult to see the flow of cash from AI-driven companies into our space. So we like the way the funding environment is developing. And just as a reminder, the word resilience, we've survived and thrived in the hardest biotech funding environment for quite a long time. So I hope it's clear and you can see we're making good progress as a business across many, many, many applications, many markets. And because of that, the resilience in our business is incredibly strong. So I think with that, that's probably enough from me. And it's back over to you, Angela.

Angela Bitting, Head of Investor Relations

And so for the webcast, we will see you in 20 minutes. We are going to cut the webcast for two customer presentations. You will then join us for the last three. The last three customer talks are amazing. So, and they are with us in the room today. It's my pleasure to introduce John Overton, Chief Sequencing Officer of Regeneron Genetic Center, where he oversees large-scale genomic sequencing initiatives that advance precision medicine and human genetics research. I saw John present at a sales meeting for TWIST, and when we were talking about who to bring in for customer stories, I requested John. He graciously accepted. Over to you, John.

John Overton, Other

thank you very much yeah and thank you for having me here today I'm going to take a slightly different angle than the first couple of talks and I'm going to give you a story and I'm going to talk to you about why in this world of rapidly decreasing whole genome sequencing costs I don't think that's the best approach that we use for our large scale drug discovery research especially at a place like Regeneron So the RGC has been around since 2013, and it was founded with the individual goal of we were going to use the power of the human genome to find the individual differences between two people that make you either susceptible to or resistant to developing disease. Now, I wasn't in the talks through most of the day, so I'm going to take one step back here and just talk about DNA for a second. All right, DNA, four pieces of information, AT, G, and C. It's those patterns over and over again make each one of us unique, but the genome is big. It's 3 billion pieces of information. It's actually 6 billion because you have one copy from your mom, and you have one copy from your dad, so we're looking for these individual variants that make each person different, and now at Regeneron, what we do is we sequence people. We sequence their DNA. We have access to their medical records. We compare them. It drives our drug discovery process. To date, we've done that for over 3.5 million people. It's one of the world's largest database to do this research. It drives our drug discovery and clinical trials. We do this through two different types of technology. One is genotyping. The other one is whole exome sequencing, which Twist is an expert at. So what are these approaches? Genotyping, it's been around for a very long time, a couple of decades. It's an array-based approach. You have a probe. Each one of those probes detects a variant in the genome, tells you what variant is that position. it's kind of low throughput. They're spread throughout the genome. It's not high resolution. They're tools that we use. They have a couple hundred thousand probes in them. You look at common variation, and what you do is you create something called an imputed genome. And the reason we can impute your genome or guess at it is because when we inherited our DNA from our ancestors, you didn't get one ATGC at a time. You got a whole bunch of them. You got a million. You got five million. You got 10 million. You inherited them in big chunks. And so I can look at your genome, and if you have a variant here, here, and here, and the reference has the same variants, here, here, and here, everything else in the middle, it's probably the same. We can assume that. So we do that genotyping. We combine it with something called whole exome sequencing. We do this with twist, where we get very fine resolution of the coding portions of the genome, that 1% of the genome that makes the proteins. We can do that very quickly. It's about 20,000 genes in the genome. We take these two things, we combine them together, and we get a very, very good imputed genome. But the question I get asked all the time, probably at least once a week, why not just sequence the whole genome? Costs are going down, just sequence the whole genome. If I sequence the whole genome, I get a lot more variants. It's a lot more than if I do an imputed genome. It's about 3.6 million variants on average in a person, but this comes at a significant cost in a significant amount of time. Most of the variants you detect in a genome They're incredibly rare. They're what make each one of us unique, but they're not seen very many times. They don't have a great impact on the research that we like to do. It's also a lot more expensive. And so I'm going to take you through some data that hopefully convince you of that. So if you haven't heard of the UK Biobank before, this is an incredibly interesting cohort. It's half a million people in the UK. They've agreed to be part of a research cohort. They signed up a couple of decades ago. We've sequenced their DNA, and they've followed them longitudinally so we can use them for research. These half a million people, incredibly unique because they've been genotyped, they've been exome sequenced, and they've been whole genome sequenced. This is unprecedented. This is never going to happen again. This is not a cost-efficient thing to do. But it's moved as the progression of the technology has moved. But it allows us to compare how each one of these technologies performs in the drug discovery and in the research world. And so first, just going to take you through the number of variants that are detected. It's kind of small here, but in the top box, whole exome sequencing, that was done at Regeneron. You detect about 17 million variants when you do that. The genotyping and the imputation, it's about 110, 111 million variants. Combine those together, get about 126 million variants in 150,000 people that we studied. Whole genome sequencing, nearly 600 million variants, five times more than that imputed genome. It's a lot more data. But what does that really mean? If you look in here and you compare the number of variants in the coding sequences, the coding sequences, the 1% of the genome, this is the exome, these make your proteins. These are the things that we can make drugs against. Whether you do an imputed genome or you do a real genome, you get the exact same numbers. It's about 6.7 total million variants that you get out of there, and the overlap between those, it's 97%. So whether you do either one of these assays, you're getting about the same exact number of variants. You look at the individual level in the chart that I've just highlighted. Per person, it's about 20,000 variants in the coding region of the genome, regardless of which assay you look at. But the really important part is right here. In the right-hand side of this panel, what I'm showing you is the number of observations in an imputed genome or in a sequenced genome of each variant that we find. So there's 475 million additional variants in a whole genome. Over 300 million of those, they're only seen in one person.