Nautilus Biotechnology, Inc. (NAUT) Earnings Call Transcript & Summary

Hello. Hi. My name is Yuko Oku, and I work on the Life Science Tools and Diagnostics team at Morgan Stanley. Before we begin, I'd like to remind our listeners important disclosure information that can be found at morganstanley.com/researchdisclosures. It's my pleasure to host Nautilus, and speaking on behalf of the company, Founder and CEO, Sujal Patel. Thank you for joining us today. Maybe to start for those that are not as familiar with the story. Where do you see Nautilus fitting within the evolving proteomics landscape? How is Nautilus approach to proteomics different from other proteomics platform that exists today, including proteomic sequencing companies like QSI and Encodia, targeted proteomics platform like Olink or SomaLogic, nanopore protein sequencing and protein fingerprinting?

All right. Well, that's a lot to dive into right there. First, thank you, Yuko, for the invite to the conference. It's great to be here again. So Nautilus is pioneering a new method to analyze proteins. And proteins are the key part of your cells that do all of the work in your body. And unlike DNA, which the analysis of DNA is a commodity, costs a few hundred dollars, you can get 100% of your genome. It's accurate, it's reproducible, it's reliable, easy. Proteins are very different. Protein analysis is incredibly hard. It's complicated. There's many different dimensions. And in the end, the very best we can do is get a small fraction of the answer that we're looking for out of a sample. That's a big problem. It's a big problem because 95% of our FDA-approved drugs target proteins, most molecular diagnostics target proteins, proteins are incredibly important for therapeutic development, for diagnostics, for precision medicine, and the world does not have a good way to measure them. The way that we measure them today, the gold standard, if you will, is a complex workflow that's built around mass spectrometry. Billions of dollars of mass spectrometers are sold into protein discovery environments, every single year. But for example, if I take a drop of blood and I'm looking for the proteins that are in a drop of blood, the high confidence proteins that are identified and quantified out of that can be as little as 10% of what's actually in the sample. And when you have such a small percentage of the sample being accurately identified and quantified, you don't have the ability to do complete analysis. You don't have the ability to figure out, if a drug candidate can be cross-reactive with some other part of the body. You don't have enough data for this next generation of AI technologies to really analyze it and figure out where the next therapeutics are coming from. Nautilus is building a new instrumentation platform that is a complete platform end-to-end, that's focused on delivering comprehensively the entire proteome out of any sample from any organism. That is a -- you mentioned a bunch of other companies in this -- in your opening. That is a very different proposition than what many of these other companies, not many, all of these other companies are after. All of the companies you mentioned and things that are separate from the mass spectrometer are focused on some small niche of a solution. I have some percentage of the proteome, but not sensitively. I have some applications in sequencing, but only very, very short fragments of a protein. I have some ability to measure 10 or 15 proteins, but not the whole thing. We're building a platform that can identify the entire proteome and dig deep into single molecules of interest, which is becoming increasingly important in the customers that we're talking to. So that's kind of us in a nutshell. We are a development stage company. We are planning a full launch of our full proteome solution at the end of next year. And this year, we've begun early analyses with customers for some of these deeper proteoform applications. I'm sure we're going to get into.

Yes. I want to dig right into the science here now. You recently made manuscript available on BioArchive, which introduces iterated mapping of proteoform, IMAP, a method that enables interrogation of proteoforms in a massively parallel manner. Now one of the things that jumped out at me, was the methods dependent on availability of highly specific and sensitive antibodies for the particular targeting question. In light of that, in what format would you make this application available to customers? Do you anticipate having kitted solutions that have already been validated at Nautilus? Or would you also enable custom solution by allowing the customers to bring their own antibodies?

Yes. So the preprint that you're talking about in this capability is what we call proteoform analysis. And just to kind of differentiate that from what we expect to launch at the end of next year, when -- one of the most basic questions that biologists ask is, here's a sample, tell me all of the gene encoded proteins in it. And that is an unsolved problem and one that we expect to address at the end of next year, and we think that product is an absolute killer product in pharma, in DX, academic nonprofit research. The other half of our platform, same platform, different application is focused on digging into 1 or 2 or 3, a small number of proteins and mapping all of the different chemical modifications and forms of that protein. Why is that important? It's important because there may be 20,000 canonical gene encoded proteins inside of a human, but that is a very small fraction of the complexity of a human. All of these proteins are degraded in different ways, are functional in different ways, and that diversity is reflected in the chemical modifications that are done by kinases and enzymes on these protein molecules, and it's encoded in that form of the protein. There really is no good at-scale way to understand what are the forms of proteins that are out there. So why is that important? Well, take, for example, this preprint that we put out. The preprint is all about tau and looking at hundreds of different forms of tau. And in fact, our assay is capable of up to 2,000 forms of tau. In Alzheimer's disease, the key pathology is that there are many phosphorylation events that occur on tau. Phosphorylation is one type of modification, there's too many of them, and it causes an accumulation in the brain, which ultimately causes neural damage and leads to the awful symptoms and awful lifelong effects of AD. This preprint was the first time, that it has ever been -- the entire form of the tau protein as you progress to AD has been analyzed. We, for example, were able to find a quadruple phosphorylated, meaning 1 molecule that has 4 modifications version of tau, that showed a distinct pattern that got to that point, meaning a pathway that led backward in time. If we can follow those types of pathways, you can intercept AD earlier and our customers and future customers in therapeutic development space are going to use that information to build better therapeutics that are earlier, that are more precise. That's really the crux of that preprint. Tau is the first protein that we're going to work on. We'll likely do 1 or 2 more in neuro-degeneration, but this is a generalized problem across cardiology, across inflammatory disorders, autoimmune, cancer. And so there's a great deal of white space for us to expand into. But this is a market where each of those applications, we have to build a new assay. We have to go and evangelize it. We've got to go put it out into the world. It's a slower build than the product we intend to launch at the end of next year.

Okay. That makes sense. And then one of the other interesting aspects of the paper, is that you demonstrated the ability to measure over 130 different proteoforms of tau, some of which had many of 6 co-occurring phosphorylation events. Tell me, how common is it to see that many different protein alterations on the same gene? And how big of a problem is the lack of ability to interrogate these forms in determining biological function?

Well, I'm going to zero in on an interesting part of your question, which was how common are 6 phosphorylations on a single molecule. The answer to your question is, we have no clue. This was the first time that anyone has ever gathered this level of information on any molecule. And this is on the tau molecule with our assay. We don't know how common it is. We know through some methods like top-down mass spectrometry that modifications are extremely prevalent. We also just intuitively know that as well, right? We have 20,000 genes in a human. It's less than a banana has. The complexity of the human is not encoded in our genes, it's encoded in all of these different chemical modifications and iso-forms of these and splice forms of these proteins. And you asked a question, well, do they have biological relevance? That's a question that is answered. Every one of those modifications, creates a different pattern within the protein. It creates some -- sort of messaging change. It creates a degradation. It causes protein to migrate from the nucleus to the cell-surface. So they all have functional changes. And so understanding this is going to be critical if you're trying to build better therapeutics, you're trying to build better diagnostics for the future.

Another highlight from the paper to me was extreme reproducibility of the platform, which other proteomics vendors have also highlighted as a key differentiator. Could you explain to me why this is important?

Yes. Let me try describe -- I'm going to answer the question two ways. One, I'm going to answer the first part, why is it important? And then I'm going to talk about how we're able to deliver this reproducibility and how it compares to others' claims? So reproducibility is absolutely critical, because reproducibility means that you have data that you can rely on. If you're trying to do an analysis and understand what is the marker that is a therapeutic target that I'm going after. If you're trying to look at a diagnostic for AD and you want to understand, is this a biomarker that always precedes -- change always precedes disease or not? Those are things you have to be able to rely on. If there's differences and you look at the same sample and there's variability that's significant, you're never going to make sense out of what you're looking at. And particularly when we move to this world of the future, where AI and other data science technologies are analyzing these data sets, having data that is reliable and always correct is going to be 100% critical. So that's the reason why it's important. The way that we get reproducibility is fundamentally different than what anyone else has attempted before. The way that others get reproducibility is that they tighten every aspect of their analysis and their assay. They make sure that every single reagent that goes in their system is in these tiny, tiny narrow bands of specification. They go and make sure that assays always run in the exact same way, the same time, the same temperature, at the same instruments. This is really the level of specificity. Our approach is the only approach which takes a single intact protein molecule and probes it over and over again with all sorts of different reagents, gathering information that increases our confidence about this molecule. Because of that, we're not just taking one data point. We're taking hundreds of data points and putting them together to come up with a comprehensive exact identification of what the molecule is. By doing that, we are able to be so confident about the molecule's identity and proteoforms that our CVs co-efficient to variation, a measure of [ reproducibility, incredibly ] tight. We showed CVs between 1% and 5% in the first experiments that came off of our platform that made it into that preprint. 1% to 5% for a platform that just, is publishing first data is crazy. Most of the platforms that have been around for 2 decades have 20% CVs. And even then, you can look at the asterisks on their disclosures. It's really, really -- it's really, really hard to figure out, is that really a true CV? How did they look at it? Because we're able to gather all these data points, it's a fundamentally new method of gathering reliable data. And we think that's going to be critical for us over the course of the next few years of productizing both those proteoform capabilities and our broad scale capabilities.

Great. While this study used cell lysate, do you imagine that over time, the technology could be used in fixed tissue slices? Are there any technological limitations that will restrict you from doing that?

Yes. So when you -- your question more broadly is about sample type. So when you think about each application, the sample type is going to be different. When we're talking about tau, tissue is fine, but that means that your patient is diseased. So there is a need for -- from our customers to move to CSF as a potential sample input and then ultimately to blood serum, which, of course, is the easiest and most prevalent out there. And so those are two capabilities that we intend to pursue over the course of time as well. And the issue, of course, is that CSF has at least a few orders of magnitude less tau in it than brain tissue, another few orders of magnitude to get the blood and some of the forms might begin to degrade. And so we'll work through that over time. Things like frozen tissue and so forth, the only limitation for us is that, in order to do these proteoform analyses, we're going to want whatever has occurred with the tissue to make sure the proteins are still intact. They can be denatured. They could be -- they could lose their structure, but they have to stay as a whole.

You also announced that you entered into an agreement with Allen Institute to evaluate connection between tau protein and neurodegenerative disease. Tell us a little bit more about the goals of the partnership and when we might begin to see data from the collaboration?

Yes. So the Allen Institute for Brain Sciences agreement is really a pilot to start taking some of the samples that they have of brain tissue and looking at how our data relates to the data sets that they already have on it. Their goal is to demonstrate with their own samples that the depth that we showed in that preprint is what they should expect out of their samples. And then the goal there for us, of course, is to go and turn that into a Phase II or Phase III after that and really start to do some analyses that help understand the basic pathology that occurs with tau as it relates to AD and frontal temporal dementia and some of the other tau-opathies that are out there. This deal is probably a lot like what we will see both with other research institutes, but as well pharma, right? It's always going to start with a very, very small pilot. Some of those may be paid, some of them may not be. But when they are paid, they're small dollars. Really for us, revenue is not important in any of these engagements. Our goal is to show the world that if you don't measure tau proteoforms at this depth, you are not going to be able to build better therapeutics, you're not going to be able to build better diagnostics. And so our goal is to make it known, that this is a necessity, not just for tau, that's a proof point of 1,000 other biomarkers beyond it. But the goal is really to show the world what the power of this technology is.

And just based on customer -- potential customer conversation you had so far based on these protocol analysis capabilities, could you provide color around the mix of customers that have expressed interest in potential collaboration partnerships?

Yes. I mean mix of customers is probably too early to generalize off of that. These capabilities are about 3, 4 months old. And so it's very, very new. The first people that read these preprints are the KOLs in the neuro space, and they all call up, they're like, holly macro, how did you do this? And then we have a conversation. We also just went to the AAIC, which is the big Alzheimer's conference just occurred. So we met with most of the academic and nonprofit research KOLs in this space. And so I'd say that most of our conversations have been with them, but the pharma conversations are starting to pick up as well as we start to show more of the data that we're -- that we've been generating with some of our early, early collaborators.

Great. And then just moving on to now broad-scale discovery capabilities on the platform. In conjunction with targeted proteoform analysis capabilities on the platform, are you also advancing broad-scale discovery proteomics? You're also advancing broad-scale discovery proteomics capabilities. Are there key takeaways or learnings from the IMAP publication that can be read through to the efforts of broad-scale discovery efforts?

Yes, that's a great question. Just to level set, right? So we have one platform that has 2 use cases. One use case is these proteoform analysis and one use case is what we call broad scale analysis, which is a discovery application. I have a sample tell me everything that's in it. The platform elements are the same. The only thing that broadscale requires is, it requires a few hundred proprietary reagents, which are the probes that we've been talking about on earnings calls that we've been building for many, many years. The rest of the platform is actually pretty much identical. The go-to-market for each of these opportunities is also quite different. While both, I firmly believe now both are really big opportunities, multibillion-dollar opportunities each. There's a ton of market development that has to go into the proteoform work because no one has even conceived that this type of data was producible until that preprint came out. The second thing is for every biomarker, I have to make a new assay. And that's -- it's a significant amount of work for us to build an assay to go through verification validation, to ship it and so forth. So panels are going to come out more slowly and the revenue wheel is going to spin more slowly on that side. Not to say that it's less important, but that's just a fact. On the broad scale side, customers are buying million-dollar instruments every single year to try to do discovery proteomics. The sales cycle is going to be shorter. There's an existing CapEx budget that we can fit into, and customers readily understand our differences relative to the incumbents that are out there. And so we expect that when that product is done, it's one product that will sell at a much more rapid pace. With that, we're a company with very limited resources. We have -- while we have $180 million-ish of cash, we are preserving the vast, vast majority of that for our completion of development on the broadscale side, the commercialization, building the commercial team, commercialization, early revenue. So if you ask me what the split is, it's probably 5% of our resources are being spent on proteoforms right now. It's way lower than I'd like, but it's kind of a fact of where we are, like 95% of our energy is still focused on broad scale because we believe that the differentiation relative to the competition is enormous, and we think the revenue opportunities are more immediate.

Would you ever consider just launching a product with proteoform analysis capabilities alone? Just to get researchers more comfortable with the data that comes off it and build a brand awareness?

For sure. Yes, absolutely. And it's not a question of if, it's just a question of when, right? We will launch those capabilities. And today, it's just collaborations. It's a one-off process each time, but certainly, we're going to launch those capabilities. You asked earlier, how are we going to let people access those capabilities? Initially, we're going to let customers access those capabilities through a service, send us the samples, we'll analyze them, prep them, get you the results. And that is a good model for proteoforms, because there's not a lot of appetite out there to buy $1 million instrument just to do proteoform analysis on one marker. But over time, once the instrument is being placed because of our broad scale capabilities and that we don't have panel, we have 10 or 20 panels, we expect that we'll be using kits to drive those proteoform capabilities on the customer's own instrument as well.

One of the reasons for delaying launch was to reconfigure the broadscale assay, which is expected to reduce technical risk and yield, greater number of affinity reagents. First of all, how do you know that changing broadscale assay configuration will yield more affinity reagents to meet your specifications? And then second, I understand that it's a significant undertaking that's likely to take multiple quarters, but could you provide any updates on those efforts?

Yes, that's a great question. So our broad-scale assay depends on us building hundreds of antibodies. We call probes, affinity reagents, whatever you want to call them, hundreds of probes that go and will bind to pieces of a molecule. And those pieces are very short, usually about 3 or 4 amino acids. And that level of information by itself is nothing. We can't tell you what a molecule is. But when you stack up 100 or 200 or 300 pieces of information from these probes, you can come up with a shockingly precise idea of what the molecule is, and then we can do that for billions of molecules across our system at once. One of the key problems that we talked about on our earnings call at the very beginning of the year was that, too few of the probes that we're building are functioning well on our platform. And by functioning well on our platform, what I mean is that every antibody on the planet has a natural state. How long does it take to bind? How long does it stays put before it unbinds? The unbinding was too fast for us. And we had to deal with that either by building thousands and thousands of more probe candidates with a very small yield, which is not an approach that you want to take for a number of reasons or biting the bullet and changing our assay configuration, so that it's much more tolerant of these shorter rates that antibodies come off of a protein. And that is -- that was what led to 1 year delay in our schedule from this most recent round. And how do we know that it's going to make a difference. We've done extensive proof-of-concepts before we started to do the work to change the assay configuration. Changing the assay configuration for us requires a chip and flow cell change. So we've gone through the process of doing that. We have external partners that help with that. And I'd say that we're through the development work on that and are at a phase now where we're proving. And so what I'd say is over the course of the next 1 to 2 quarters, you'll hear a lot more specificity from us on where those efforts are. And those efforts are expected to allow our probe library that we have today to be much more usable, which means that we will be able to get to our goals by the end of next year.

Great. Looking forward to it. And then beyond affinity reagents, what to do are remaining prior to commercial launch? What types of internal performance metrics, milestones are you watching to gauge, whether you're on track for a launch in late 2026?

Yes. So what I would say is that from a development perspective, many parts of our platform have had a lot more time to mature because it's taken us a lot longer to get these affinity reagents done, to get the probes that we need to get a full platform out. And with our work that we're doing on proteoforms, we're showing that the platform is capable of very tight CVs. It is capable of cycling affinity reagents in one at a time. The chips and flow cells are fully functioning. And so a lot of those capabilities are being hardened by the work that we're doing in proteoforms. The go-to-market work that's required as we head to broad scale, is really related to getting a preprint out that looks like the preprint we just did on tau, but for broad scale. It's launching an early access program to let early customers have access to these capabilities via a service, which gets them comfortable with the results and it gives us the data that we need to continue marketing to other new customers out there. And we've really just -- those are the big steps that are coming up that get us to that final release at the end of next year.

And how should we think about time lines to when you might kick off the early access phase?

You should think about the early access phase is about 6 months before the launch. And so that's the goal for us. And so we can work backward from whenever you think, end of next year.

Okay. Sounds good. All right. And then moving on to manufacturing. What have you done to ensure that you're ready for launch in late 2026 from a manufacturing perspective, both on reagents as well as on the instrument side?

Yes. So on the instrument side, like I said, the instruments had good time to bake. We have multiple facilities ourselves. So we know what it's like to ship an instrument. We know what it's like to get an instrument up and running on another site. We've worked through some of those early things that you would expect. We have enough scale on the instrument side to go and get done a reasonable number of builds for a revenue ramp, and as well, we've worked with our supply chain to reduce any of the single parts that are difficult to have long lead times. We don't have any of those sort of things on the instrument side. On the reagent side, we have a robust reagent manufacturing capability within our walls. And then we have a number of partners on the antibody side, in particular, where we have [ burst ] capacity, and we have capacity to bring in product that is the same specifications that we make internally. And so we've got a robust pipeline there of partners and CROs that are backing us up, and we feel confident that we're good to go from that perspective as well.

Should we anticipate the launch in late 2026 will be fully scaled? Meaning you'll be able to ship as many instruments that are in demand? Or do you anticipate that to be steady as you gauge demand?

If we were fortunate enough to have so much demand that we had to worry about that, we would [ stage the mouth ] a little bit. I mean, this is a brand-new launch of what I think is the most ambitious proteomics platform that's ever been built. And we will take it a little by little, to make sure that our customer satisfaction is very, very high, right? Those things are super important when you're introducing a new capability like this.

Makes sense. Okay. And then while focus has been predominantly on progress of developing affinity reagents, could you share where you are with respect to software side of things? Is it possible to work on software to back-end analytics in parallel? Or do you essentially need to wait until you have a complete set of affinity reagents for the launch?

So there's a lot of different aspects to software. The instrument requires a tremendous amount of software to run. We send all of the data that comes off the instrument to the cloud because the type of analysis that's required to identify molecules is pretty compute-intensive. And so all the technology for that data pipeline, for the cloud software, for the customer portal, all of that technology is being built and is in good shape for launching at the end of next year. There are a whole set of analytical capabilities after that as well. And we think those capabilities will be critical differentiated capabilities for us that perhaps we will even be able to charge more for. But no, we haven't substantively started building a lot of that technology yet because, yes, we could parallelize some of it, but it's better if we see the data first and understand what we're looking at and how our customers want to use that, right? There's going to be -- there's never been complete information in proteomics before. There's never been this level of mapping of proteoforms before. And so the process of understanding what our customers want to do with that information is -- it's the learning process that we're in the middle of with customers on the proteoform side, and we think that, that same learning process will happen on the broad scale side.

Got it. And then as you touched on earlier, you're talking about a $1 million price tag for a complete bundled solution for your platform. So first of all, is that how you're still thinking about it? And second, if the environment continues to be challenging at a time in a commercial launch, could you provide your thoughts on the ways you could help to facilitate accessibility to the platform?

Yes. So we'll talk about it in more depth on the next earnings call, but we are just about done with a new pricing study. And I can tell you our price point is absolutely solid for the value that we're going to deliver. And so a roughly $1 million deal is what I expect that we'll be launching at. That deal is instruments, software support, some reagents like prep install to get you going. And given the value that the platform is capable of putting out, what the data that we're seeing from customers is that, that price point is solid and that a few thousand dollars a sample, which is what we've been talking about is a solid price point on the consumable side as well. Now your other question, what do you do if it's not accessible to users? There are a number of things that you can think about doing there. One, is customer can access these capabilities in a less scaled fashion using our service offering. We expect that customers on the academic and nonprofit research side will need to apply for grants. And when they do apply for grants, we'll provide capabilities for them via the service until their grants come through. There's the possibility of things like reagent rental models and so forth, but we'll be careful with that, and we'll have to feel our way out as we get closer, right? There are some platforms that don't quite have the value that we do, in the tool space today that have started to use those models because customers don't want to pay for the instrument. Our instrument is fitting into a bucket where customers are buying mass spectrometers, and we think that the budget is there, and we don't want to compete against ourselves here. So we'll have some flexibility, but we're going to be careful.

You've been extremely prudent in your spending and extended your cash runway into '27. Could you remind me of your cash position and provide examples of how you're able to manage costs so well?

Okay. So I did -- our cash position is about $180 million. And when we took the company public 4-plus years ago at this point, we raised $345 million. So we have done an excellent, excellent job with cash management. How do you manage cash? Like that is a daily job. My CFO is in the audience here, Anna Mowry. She is like the best at it. Anna and I worked together at my last company, which was a publicly traded tech company. We had the company at a positive 20% operating margin before we sold it. That was on a scale of roughly of closing $100 million a quarter for the last quarter before we sold that business. Making a business as efficient as we think we are, is like thinking like a start-up. Everything in our organization is like a startup. There's not one headcount where I'm like, that's a headcount that isn't necessary. Every single dollar that we spend, we're spending in a prudent way. And we think that's important for all of our shareholders. We also think it's important for us, right? Parag Mallick, my Co-founder and I own 1/3 of this company still. So every dollar we spend, with $0.33 out of our pocket. We take it seriously.

Great. And then in the last couple of minutes here, I just want to close up with a bigger picture question. So how would you anticipate proteomics to evolve in the future with new emerging technologies, improving scalability of proteomics as well as increasing number of targets that can be identified on the platform? And in your view, what are the key differentiating factors for those that take a majority of the market versus those that are limited to niche applications?

I mean that's a great question. And I think that we can take some hints from the genomics era on this, right? There have been no less than 50 different platform attempts in the genomic space. And out of those, one, Illumina emerged as by far the leader. They had a solution that was reliable. It was packaged really well. It gave you a largely complete answer, and they did a great job of executing. I think proteomics will have lots of winners, but we think we're the one that will emerge, assuming we get done what we say we're going to get done, which I'm hopeful of. I think we're the one that will have the opportunity to really democratize access to the proteome, and that's why I get up every morning.

Okay. Well, thank you so much, Sujal.

Thank you, Yuko.