Seer, Inc. (SEER) Earnings Call Transcript & Summary

Hey, everyone. Thanks for joining us today on day 1 of our Health Care Conference. I'm Tejas Savant, and I cover the Life Science Tools and Diagnostics Space here at Morgan Stanley. I'm delighted to have Seer joined us this morning. And representing the company are Omid Farokhzad, CEO; David Horn, CFO; and Omead Ostadan, President and COO. Before we get into the discussion, just a quick safe harbor here. For important disclosures, please see Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. And if you have any questions, please reach out to your sales rep. So with that, Omid, maybe we'll just start with a quick overview of Seer for those who are new to the story, and then we'll dig into the details.

Okay. Thanks, Tejas. Good morning, everyone. Thanks so much to Morgan Stanley for the opportunity for Seer to present. So at Seer, we're developing transformative products that's going to open up a new gateway to the Proteo. Over the last 15 years, our understanding of biology has grossly expanded, many thanks to my colleagues in the genomic space. And as profound as that impact has been in terms of health disease and management of disease, I think that broad scale access to proteomic is going to have even a deeper and more profound impact in, frankly, making the world a better place. Over the last 15 years, with cost of genomic dropping technology is getting better, we've now sequenced over 1 million genomes over 10 million exomes. The cumulative data of which has identified over 700 million genetic variants, yet we know literally the tip of the iceberg of what this means in terms of its function. To get to the function, you really need to understand proteins, which are the functional use of life. Now at Seer, we've developed a Proteograph Product Suite. It's a solution that leverages our proprietary engineered nanoparticles. The product has -- our consumable, which includes the particles, an instrument, which is in Hamilton OEM instrument designed for assay and a software suite that helps the customer go from biological data to biological insight. The business model is a razor-razorblade model. We install the instrument, then we have consumable pull-through. Now before Seer, it was possible to access the proteome in an unbiased way. Those methods involve depletion fractionation. It was costly. It was time consuming. You just really couldn't do any studies at scale. In fact, prior to Seer, the largest deep, unbiased study was a study of about 48 samples. Now there were scalable approaches to proteomics. Those are targeted approaches, and they have a different set of limitations. And their limitation really is that they don't get to the depth or breadth the complex of the proteome. So Seer is the only commercially available solution that gives you deep unbiased proteomic at scale. We began commercialization of our Proteograph Product Suite in late 2020, decided to take a calculated 3-step commercial approach, a collaboration phase, which we completed, followed by limited release, which is what we're in now. And then culminating into a broad release, the approach has been used to introduce other first of its kind highly disruptive platforms in the past, frankly, many of it in genomics. And we think if you do it this way, are you able to work with lighthouse customers and KOLs that can really demonstrate the Proteograph Product Suite for key applications, really develop strong reference sites, create blueprints for others to follow as we shifted a broad release. And I think, ultimately, if you do this well, it's really going to accelerate adoption at the broad release. So you're starting off kind of very measured, but then you really accelerate adoption as we shift into broad release. So anyway, I hope that was helpful, Tejas. Really excited. And let me turn it back to you.

Sure. So Omid, that's a great jumping off point, actually. I mean the proteomic space has existed forever, like you said. With all these new innovative technologies coming to market, including you guys, do you sense this groundswell of interest that's starting to rise here? In other words, are we at an inflection point? And can you point to some sort of anecdotal evidence or your own sort of voice of customer work that convinces you that now is the time, now that you guys and some of your peers which we'll get to in a minute are starting to sort of make inroads into the market?

So yes, I think we're at an absolute inflection, Tejas. As I mentioned in the intro, Seer is the only technology that lets you kind of get to the depth, breadth complexity at scale. And before, you just couldn't do that. And the thing is, the market wants to get to function it always has. There has been a pent-up demand in that. Genomic provided an enormous amount of content. We just couldn't get to the function. And in the absence of access to proteomic, the market try to do that, for example, looking at methylation or looking at the transcriptomics. But every one of those brings you closer to function, but doesn't get you to function. And so the ability to get to function has been a long-waited goal for the scientific community. And I think with the enormous body of genomic that's coming in, and frankly, really just a tremendous amount of innovation that's not gone into the space, we're beginning to see that content becoming available for scientific community and the adoption of it actually speaks volume to that, Tejas. I think some of the data that's going to be emerging in the upcoming conferences is going to speak volume to the kinds of information that is now becoming possible for the customers.

Got it. And then as you think about the competitive landscape, I mean, you mentioned the Olink and SomaLogic of the world who are doing the more targeted work. But then you also have new entrants like Encodia and Quantum-Si and Nautilus and so on. How do you see the market evolving with the entrants of these new players? And where do you see Seer sort of position in that ecosystem?

Yes. So look, I see the market -- I break it down in my simple mind into 3 buckets. You've got those that says the existing detectors, which are the mass specs are not good. So I'm going to detect and I'm going to develop a new one. That's QSI, Nautilus, Encodia and then you've got those, and I'll comment on that. But then if you got those that says, "I'm going to approach proteomic in a targeted way." You've got Quanterix, you've got Olink, you've got SomaLogic in that bucket. And then you've got folks that says I'm going to approach proteomic in a deep unbiased way. And let me be very clear, there's exactly one company in that bucket, and that is Seer. Now our detector of choice today is a mass spec, but we are detector agnostic. So among the mass specs, we've used all of them. In fact, we have SCIEX, Bruker, Thermo instruments in our own labs, but we're more broadly detect diagnostic. The reason we want to mass spec initially is that there is 15,000 of them installed globally that look at proteins. I mean a total number of mass spec installed is about 50,000. And so that's a large market opportunity for us to go after. But in the world that QSI and Nautilus, Encodia develop instruments, those instruments get commercial traction and the customers want to use them then we would tweak the end of the workflow, the proteom -- the Proteograph to then sit upstream to those instrumentations. We just have to see that come. And the thing is some of these companies have guided that their instruments would come in at the end of 2023, possibly 2024. And so in our view, the limitation is in the mass spec, the limitation is in deciphering the complexity of proteome, which is what Seer uniquely does, that technology will apply to every one of those other detectors, QSI, Nautilus, Encodia as well. And those instruments need to be better, not compared to the mass specs of today, but compared to the mass specs of 3 years from now, or 2 years from now or 4 years from now. And of course, there's a lot of innovation that's happening in that space as well. And so we just have to see how it comes. From my perspective, with Switzerland, just like today, we work with every mass spec. If in the future world, those other factors become prevalent, we would absolutely work with those companies to make the Proteograph work with that. Now if I look at the targeted companies, like the Quanterix, the Olink and SomaLogic, by the way, each one of those 3 targeted approaches has its own unique kind of niche. Quanterix was super deep in terms of sensitivity. Olink is on NGS platform. SomaLogic does the aptamers. But all unique. But the thing is, again, comparing Seer to those companies, frankly, it's like comparing apples to umbrellas. I mean we just fundamentally do totally different things. We fundamentally answer totally different questions. If you look at the those targeted approaches, they vary from having panels of maybe 30 year or so proteins upward of several thousand. But if we lived in the world that you went from 20,000 genes to 20,000 proteins, then a targeted approach would be okay from a discovery perspective. But we go from 20,000 genes, and we end up with 1 million or 2 million or 3 million or 4 million, depending on whose papers you're reading, protein variants. And so you cannot discover the complex of the proteome using those targeted approaches. And if you make an analogy to the genomic space, when you looked at the targeted approaches like the SNPs, they had limited value ultimately both for discovery and for clinical utility and technologies like NGS that allowed you to look at nucleic acid at amino acid -- sorry, nucleotide resolution, not only were they useful for discovery, but ultimately, the clinical utility was also based on the same platform. I think a very similar movie is going to play again in the proteomic space. In that, not only are we going to look at the [indiscernible] landscape at the amino acid level, look at all these protein variants, but actually given the complex biology, many of that application, not only will go from discovery, but also would all spend all the way down to translation and also clinical utility.

Got it. We keep getting the question on nanoparticle panels and when you might sort of find enough of, I guess, customer feedback to justify launching a different or perhaps expanded nanoparticle panel. Can you comment on that? And as a second part of that question, beyond just nanoparticles, you are working on increasing protein coverage and throughput and driving down sample volume, et cetera, which of those sort of dimensions beyond just nanoparticles do you think will be the most impactful from a customer standpoint?

Look, again, Tejas, we're at the very beginning, right? So we literally -- the instrument has been in the hands of customers for, in some cases, just 6 or 7 months and then in some cases, weeks, and we won't even hit broad release until the end of the year. So too early to judge when the other panels will come. Suffice it to say that there is -- that the first product has an enormous runway in terms of the value it adds. And you're going to see that beginning at, for example, the ASMS conference late October, with data being presented from these collaboration and limited release customers in what is possible. So I think the existing product has a very long runway in terms of value-add on what becomes positive. Now that said, we have now built an arsenal that -- of nanoparticles that can do different things and combining them in different ways, create different panels. So we can go from product 1 to product 2 to product 3 very rapidly. And my expectation, frankly, is without giving any guidance that you will see other products come despite the fact that the first product is awesome to begin with because the first product is designed to apply a very -- a big part of the market segment. But if you're, for example, working with model animals in some cases, small animals, where the plasma volume may be small of what you can use, then being able to use the Proteograph with smaller sample volume may be useful for those preclinical studies. Or if you're interested in some post-transitional modification, having panels specifically look at some of those preclinically useful. So we will see other particles panel come possibly as early as 2022. I will give more guidance, Tejas, by the end of the year in terms of what that would look like. I think if I look at what the customers want, look, the customers will always want more cheaper and with using less of their samples. And -- but that said, we're moving in the direction. The existing product, frankly, fits so much of their need, but we have a long runway, Tejas, in making improvement across any number of those dimensions as the market requires it.

Got it. How does the throughput and turnaround for the Proteograph compared with, say, an Olink or a SomaLogic? I mean, they now -- you mentioned that they have sort of targeted libraries in the multiple thousands, but they can't sort of at least for now handle variants. So in terms of throughput and turnaround, can you help sort of benchmark the Proteograph? And then what does the volume per sample requirement look like? And does that sort of shut you out from biobank sample work, for example?

Yes, excellent. Tejas, so first of all, 1 Proteograph instrument and 2 mass spec can do 48 samples in 2 days. Just from a context perspective, that is exactly the throughput of 1 NovaSeq right? That gives you basically 6,000 samples that you can process in a year with that setup. And by the way, just like the genomic guys, if you want to do a larger study, then you will just have one of those instruments and you can run it. And by the way, improvements in the mass specs are coming. So that 4-year sample in 2 days may shrink further as the mass specs even get better, right? So that's the first one I would make. If you look at sample volume, I think we're making an excellent point in terms of the very, very precious sample that, Tejas, in some biobanks. I do think, Tejas, that if you look at clinical utility today, like, for example, what is the amount of blood that's required to do a whole genome, we're actually in that same kind of range into do a proteomic study. But that said, today, we use just a little over 200-microliter of sample to do a study. And my expectation is that plasma volume will decrease. Now remember, we use that because we believe that delivers the optimal content. A customer may actually choose to use significantly less amount of plasma even with the existing panel and get 80% or 90% of the amount of data, they may be satisfied with that. So if you have truly a sample limitation, even the current protocol, allows you to sacrifice some content. And by some, I don't mean 50%. By some I mean like 10% or 15% or 20%, and use really a much, much smaller fraction. Now that said, with the subsequent panels coming, my expectation is that not only do you not have to sacrifice content, you might even be able to get to more content with lesser volume plasma. But again, Tejas, we have to build the market and kind of approach it in a very thoughtful way, exactly the way the genomic guys, by which one of their pioneers, frankly, is on this call, did that in the genomics space in the last 15 years.

I assume you're referring to David, but we'll...

I was referring to Omead.

When do you expect to run sort of tissue samples on the Proteograph, Omid? And are there any sort of technical challenges that you need to overcome before you can sort of get there?

Tejas, we've already exemplified the use of the Proteograph for a number of different sample types. Obviously, plasma, but in synovial fluid, CSF, tissue lysate, cell lysates. So those are all now exemplified. And even model animals, those are all exemplified. And I think as they get presented in conference and the customers see that they begin to come. So there is not a miracle or even an invention that needs to happen. We just kind of need to put the protocols up for the customers to get used up.

Got it. Fair enough. And then between basic research and discovery and translational applications, where do you see the greatest adoption in the near term for the Proteograph? And how do you expect that mix to evolve over time as users gain more experience?

Omead, do you want to tackle that?

Sure. So I think, Tejas, what we're seeing right now and for limited releases that we have a broad range of interest from a cross segment of customers with a heavier mix towards commercial. And this is somewhat intentional and selected by us. And in commercial, I would say probably the 2 most prominent customer types are pharmaceutical companies and I would say, translational -- the genomics translational company. Those are probably two. I expect the mix to continue to be more heavily tilted toward commercial entities, probably for the next 18 to 24 months. It's largely driven by the competitive dynamics within commercial. They're more readily accessibility of samples and their ability to, quite frankly, get the scale faster. Over the course of time, I do expect that mix to then shift a little bit more towards academia, but really for the foreseeable user, and it's going to be the inverse of what you saw in genomics, where the business started in academia and that transitioned into commercial. I really think the momentum that's already existing for multi-omics discovery and work -- in commercial is a strong tailwind that's going to allow us to establish a strong presence within those market segments and then complement it with an expanding academic base.

Got it. And while I have you, Omead, what are you -- what were the key learnings for you from the Phase 1 of your launch? And as you think about sort of early feedback from the collaboration sites even, what do they like about the Proteograph that really resonates with them and the work they're doing? And what areas do they feel as potential areas where you could get better?

Sure. Obviously, you're fairly well connected to a number of those users so much of what I'm going to say you already know, but others on the call may not know. I think, overall, I've personally been extremely pleased with the ease with which the system has gone in. The robustness of the instrument in terms of its performance. And the extent to which even literally the first runs of data coming off of customer hands are mimicking what we've seen in-house. As somebody who comes from the genomics side of the world and has been involved with introducing a fair number of sequencers, this is unusual because sequencers, usually, there's a break-in period, if you will, both in terms of the installation and coming up to speed, and we're talking about extraordinarily successful platforms. What I have seen in terms of the performance that are Proteograph into hands of customers has exceeded anything I have seen previously in terms of any instrument introduction. And the customer feedback reflects that, that customers are really quite pleased. And those are customers who come from a proteomics background as well as those from a genomics background with literally no familiarity with proteomics that they've been able to get up to speed and get going very quickly. A lot of it comes down to the fact that it's a highly automated, well-considered assay that has -- that requires very little human intervention. And the reproducibility of the platform overall has been exceptionally high. And what we're hearing from customers is, and again, Omid alluded to this, but I think I don't want to let the cats out of the bag here, but you'll start to see some of those customer experiences at ASMS and HUPO coming up, that customers have been extremely pleased with the depth and breadth of the data that they're seeing that they're seeing things that fundamentally they have not or could not see before and the relative ease with which they're accessing that data is also extremely surprising, and it sort of fits what we expected. So overall, I couldn't be any more pleased with the ease with which the system has gone in. In terms of learnings, I think really some of these have been very COVID specific, to be quite frank. I think COVID has created complexities for all of us, but the team has responded exceptionally well. So for example, how we help get the installation, get the sites right ahead of time because you can't really visit sites as often or as frequently as one might be able to, how we conducted training early on so that the customers are more prepared even before the technology arrives and how we do things like remote learning and whatnot, have been some of the things that we sort of piloted in the early phases of the launch, and that's proven to be extremely beneficial. And it's one of those things that I think fortuitously is going to help us as we move forward, hopefully, in the post-COVID world, that even those capabilities are the sorts of things that you're going to be able to deploy to have an even more efficient installation and support of the system moving forward.

Got it. And then on the decision to enter China a little bit ahead of plans, Omid, can you walk us through what sort of drove you to pull that trigger there? And in the medium term, is local manufacturing on the cards here?

Yes. Tejas, look, China has been a substantial market for us. We, of course, see that. And we had initially planned to kind of get there, let's say, second half of 2022. Really, 3 things drove the decision to get there sooner. The first was we saw the way the product was performing in the hands of customer. It was just really exceptional. Just across any metric, robustness, run success, time, uptime, quality of data, the product just did really, really well, and we felt that it actually was ready to go in the hand of an international customer. Second one was really the caliber of the established commercial academic institutions that are demanding the Proteograph. And we were very encouraged by seeing that demand from these customers, and we wanted to frankly satisfy it. And then the third -- and I would actually say probably not an insignificant one was that we had identified an awesome distributor who is obviously someone with a process of expertise, capabilities in regulatory compliance manufacturing and also someone who was known to Omead. So the CEO, Ruilin Zhao, had worked with Omead closely as a GM of Illumina China business in the past. And so the combination of those 3 factors, Tejas, encouraged us to get there sooner. As of today, there is no plan to do local manufacturing in China. But I'm not going to rule it out, Tejas, we're in the early days right now, obviously.

Got it. Fair enough. And how do you see the path to penetrating the clinical market over time? I mean, obviously, you're working with at least one customer. But as you think about sort of the cancer screening opportunity and potentially other disease areas over time, how far out do you think that opportunity is from becoming meaningful for Seer?

Look, I think our technology has applicability across research discovery, obviously, transitional ultimately clinical. We agree with that. And we plan to submit our system for clearance to enable these customers to leverage the platform for diagnostic applications. The challenge, Tejas, is given where we are right now, I'm just not prepared to give you a time line. We're literally just coming out of the gate, it's early. But of course, we're committed to support these customers, and that includes getting effective clearance of the Proteograph with time. I think if you just give me a little bit more runway for us to see how this thing does, I'll be able to be much, much more clear as far as the time line on those clinical opportunity patient's questions. Yes.

Yes. Got it. And in the last few minutes here, I do want to pull in, David. David, I mean, one of the questions we've gotten after the quarter is just with initial product revenue of about $800,000 that you just recorded there. Should we assume that there was a decent amount of discounting that went into that early phase of the launch? And could you sort of provide a framework for how to gauge consumables usage in the initial part of that revenue stream?

Sure, Tejas. And I do just want to make one thing clear upfront, which is you noted the $837,000 in product revenue. We additionally had another $300,000 plus of revenue from PragnomIQ. PragnomIQ is a customer. They were obviously our spinout, but they are -- we have arm-length arrangements with them. We do report it separately on our financials just because we still own 19%. But I would consider them a customer, and I would consider their revenue -- product revenue. So just for future reference, you should look at the 2 together, I think it's important for people to understand. It's just more of a disclosure item that we have to report it that way. In terms of discounting, I would just say no. We don't -- we're not into heavy discounting. In fact, these limited release customers are making significant financial commitments to us over a period of months and in some case, years. And that's really the nature of this rollout that we're doing is that we really wanted to find as we say a high single-digit number of customers that really are willing to make that commitment to us and us to them. And so that doesn't come with a lot of heavy discounting. It just -- it's signaling a commitment that they want to adopt the technology at scale and work with us to exemplify it in various other instances. So I think that's something that I think will -- is important for people to understand as well. And then finally, on just how do we think about consumable pull-through again, too early and too few data points to kind of give guidance on that. And I think over time, we will look to be more transparent on that. But again, I think you can assume that the majority of our revenue in the early stages will be instrument-based revenue that will then ramp to consumables. And look, the project sizes and scope that people are going to do in terms of sample sizes is a good proxy for what kind of amount of consumables that they need in terms of pull-through and the capacity of the Proteograph. So again, more to come on that. But again, we're very encouraged at the current trajectory.

Got it. And I just got one for you over e-mail here. Just given sort of some of the COVID resurgence we've seen in July and August and into September here, are you anticipating any supply chain issues or inflation in the back half of the year? And then in terms of just Street modeling here in the mid-teens range for '22, do you still feel pretty comfortable with that range given sort of the COVID dynamic here?

Tejas, so on supply chain, I mean, obviously, we've had our guards up from the beginning. And frankly, we never put our guards down. So the Delta resurgence actually was a validation of why we should not put our guards down. So I see no supply chain issues today, and we're comfortable with the numbers that you have for 2022.

Brilliant. And then one final one for you, Omead. How are you thinking about the size of the average project evolving over time in terms of just the pull-through Proteograph installed site? And how do you see that ramp?

Omead, can you tackle that?

Sure. It was a little bit unclear which Omid, but hence I paused. Yes, sure. I think, again, this is the first of its kind product. So talk about a fuzzy crystal ball, but drawing from my experience in genomics. And I think that's actually a really good proxy because ultimately, people are trying to get to this. They're trying to understand what is happening at the molecular level within biology. And so it's a reasonable proxy for project sizes. And so our sense is that in general, project sizes in commercial are going to be larger than those in academia that -- and they could be larger by potentially even factors, close factors of 2, if not more, on average. There's going to be a broad distribution. This is one of the things that I really dislike about saying averages because we know from experience, you're going to have a very varied approach to experimental design across customer sets. So with those 2 caveats out there, I do expect that you're going to see an expansion of project size as we saw with genomics, right? It's a little hard to put numbers on it, but we typically have seen in the case of genomics is that the first 3 months after an instrument goes in, and this is the ones where you've done a little bit release. When the first 3 months really sites are trying to get their arms around the system, use some example -- samples on it, the following 3 months is where they do typically their initial pilot. And so typically, it's about a 6-month time before they begin to start the very first sort of what they would consider to be large-scale studies. And those large-scale studies in academia are going to be number in the couple of hundred to perhaps about 500. In the commercial sector, you're going to see them probably starting in the 300 to 400 range and perhaps going up into the low thousands, right? And then you'll see a moderate ramp in those things. So it's the first 6 months, I think, people are just going to be sort of dipping their toes and scaling up and then starting their initial projects probably 6 months in and with that scaling up. And that's how we've just generally modeled our expected, if you will, performance of our customers. And as we get more data, we will refine that model.

Fantastic. We have to leave it there, unfortunately. But thanks so much for joining us this afternoon, guys. We appreciate it, and hope you have a productive conference.

Thank you, Tejas. Really appreciate it. You did well.

Thanks.