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

Hey, everyone. Welcome. I'm Tejas Savant. I'm the life science tools and diagnostics analyst here at Morgan Stanley. It's my pleasure to host Seer this morning, and from the company, we have Omid Farokhzad and David Horn, CEO and CFO. So thanks, gents, for joining us. Before I kick it off with the Q&A, research disclosures, please see the Morgan Stanley website at morganstanley.com/researchdisclosures. And if you have any questions, do reach out to your sales rep.

Omid, thanks for doing this again. And maybe just to set the stage, clearly an exciting year for you guys following the commercial launch of the Proteograph. Can you just highlight what you view as your key accomplishments year-to-date? And what are you most excited about heading into '23?

Yes. I would break it up, Tejas, in buckets. So one is, there was a significant strengthening of the management team. We recruited Scott Thomas to be our Chief Commercial Officer. Scott came with a long tenure in the tool space, most notable 11 years Illumina, where he ran certain geographies and just a really great guy, a good culture guy, good team builder, good knows for the way the business in the tool business runs. It's been great having Scott's partnership. We got Kenny Ross also kind of Illumina ABI to come lead our operation -- Kenny as SVP of ops, runs manufacturing operations, incredible talent. And so that's on the kind of the leadership side of it, and that was great. On the commercial front, we launched into Broad Release and continue to expand in our customer base, obviously against a very tough macro picture. We should not lose sight of that. But despite of this macro picture, we managed to expand our geographic footprint. We now have a Proteograph in Asia, in Europe and U.S. Just to highlight, for example, access in China to do that install was no easy task. I think there was maybe some 20 somewhat number of people managed to get a permit -- special permit to go there. One of them was our sales rep to actually go there and do the install. So a tough environment nonetheless. So commercially, I'm quite happy in what we've been able to achieve. And then the third is our R&D effort and our road map. Number one, the instrument, the current iteration of the product, which is the product or product suite, it's got the consumable, the instrument, the software that is performing exceptionally well in the hands of customer and has got a long leg on its own. But we have previously disclosed that with a cadence of about a product a year will be coming up with new iteration of various elements of our products for improvement, really kind of following customers ask and a ton of innovation has gone into what would become the next product, which we expect some customers will get their hands on the latter part of this year. And so that's been also terrific. And as you've seen now, we're beginning to see customer data come at conferences, and it's tremendously gratifying to see what the customers are doing with the product. In some cases, they're pushing it in areas that was not our own focus, but it's nice to see them using it in ways that they see fit and presenting those data in the entire community is great.

Got it. That's actually a great overview. I want to start with getting your take on the broader landscape here. How are you thinking about sort of the number of labs and researchers utilizing unbiased proteomics today? And among the sort of the broader mass spec user base, how many of these labs are focused on pursuing unbiased proteomics?

Yes. Let's kind of quantify some of that to put it in context so that what I say has some solid footing. Number one, our product sits upstream to a detector Workflow, we detect for agnostic. That detected today for us is a mass spec. But if in some other time, another detector becomes one that the customers like, then the Proteograph will send upstream to that detectors. That's an important point to make. And the reason is what we uniquely solve is what no detector company that I have visibility to either the ones that are currently commercial like the mass spec or the ones that are under development actually solve, which is that the proteins have a very, very wide dynamic range and highly complex -- much, much more complex than the genome and the Proteograph withstanding the buyside unit for any detector. And today, we do the mass spec, okay? That's the first thing. Next point I would make is that today, there's roughly about 50,000 mass specs growing at a 6% CAGR globally. Now about 1/3 of that is involved in the proteomics space, and about 1/3 of that does deep proteomic work, but using conventional Workflows. And just to understand the convention of Workflows, given the complex to the proteomics conventional Workflow typically involves in the depletion of abundant proteins and then fractionation of the remaining proteins to put those bits of information in buy-side units for the detector. So that's what a typical customer would do. What the Proteograph does is it replaces their upstream Workflow and allows them to go from a sample to a Proteograph to the detector much, much more efficiently. Put that in context before Seer right before IPO what we had put in the S-1 was that the largest and deep plasma proteomic study ever published was 48 samples. And that's because you do a lot of fractionation that takes time. So the user, the scientists always have to compromise in terms of do I go deep or do I expand the size of my study. What the Proteograph does is, it takes that compromise off the table, and now we have customers like Oregon Health that complete a 1,000 sample study collaboration with the Brigham doing 1,500 samples. I think Prognomics has not done a couple of thousand samples, and we spun off the Proteogenomic consortium that's going to ramp up to 100,000 samples. So the magnitude of improvement was dramatic, okay? Now -- so that's the second point of it. The third one I would make is that very early phase of this movie, so I want to be very careful about answering in terms of question in that we've just been a broad commercial company less than a year. And I'm learning, and I'm communicating what I'm learning, but there are surrogates of this movie, and we've seen it going back 20 years, the gene-chip market was robust and there was a couple of million of revenue in that as you had. And that revenue was a solid revenue when access to genome became more prevalent with NGS technologies, and we're not seeing an expansion of those. It didn't erode that market, per se, but it created an entirely new end markets and a lot of access to content. I would say, if you look at the genomic market today in total and a lot of logos contributed to that genomic market, many of those logos exist because of a single company allowing for content. So what I see happen is that we're going to expand not only capture some, but even expand the existing proteomic market because of content. We're now seeing customers present the kind of data and the kind of content that was previously not possible to access. So it tells you, I think here's the way I would answer it is, if we execute and that's a big if, if we execute well and somebody will, if we don't somebody else with EDAR launch, if we execute well and somebody will, then I think the proteomic TAM would actually be larger than the genomic channel. Why? Because it's functionally more informative than genomic information is, right?

Helpful. I think you sort of touched upon this earlier in your answer to me. I just want to talk about, in your view, how do you think the mass spec of the future looks like, particularly in terms of throughput Workflows, ease of use? And I know you're currently sort of sitting upstream to only a mass spec, but are you starting to do internal sort of R&D efforts to look at other potential downstream detector options as well?

Yes. Let me answer the second one first, and then I'll answer the first one second. If you look at the Proteograph Workflow, it involves sampling the proteom across the entire dynamic range using our technology core to which is our proprietary-engineered data particles. Once we do the protein capture today, the Workflow of the Proteograph that involves the naturing and breaking those proteins into peptides, getting those peptides ready for the mass spec. So let's say that in the future that you want to do another detector, I don't know, let's say that QSI or Nautilus or others or Encodia, those detectors -- or [indiscernible] those detectors become commercially available, number 1 and number 2, they get commercial traction. Because for me to alter the Workflow of my Proteograph, I need to have a [ needful ] market. Today, my market is 15,000 mass specs that does proteomic work. So if I sense that there is some displacement among some customers of the mass spec with that new detector, the moment we smell that then the terminal Workflow of the Proteograph changes to become suitable for that detector. Now interestingly, in almost every case, it's a simpler Workflow for us because if you actually look at the totality of the time of when Proteograph run, which is about 7 hours, roughly about half of it is to get this thing ready for the mass spec. So if I need to do that, even the Proteograph Workflow shrinks and probably the instrumentation footprint that needs to run, also shrinks. So the instrument cost goes down, the time flow goes down, you can even run more samples. So we are well prepared for that to happen. But to me, the high bar is seeing those detectors get developed and commercialized, right? Okay. That's the first part. I'm going to answer your first question now, and which is, where do I think the mass specs are going. Number one, I think we're seeing constant innovation with the mass specs. We have always said, we're Switzerland, and we have an extremely good relationship with our mass spec providers. In our own labs, we have SCIEX instruments, Bruker instruments, Thermal instruments, and we love all of them. There's some that we use more for different reasons, but nonetheless, they each have their unique capabilities. What we're seeing come from the mass specs is kind of -- and they're after the same thing that we are is reducing sample volume. For example, requirement. For example, for readout. For example, if you just look at the Bruker instrument at the [ STP team stuff that came ] versus the Pro, there's a reduction in input in the amount of peptide needed, but you get the same amount of content. So we're seeing innovation come on the mass spec side, and those innovations are very important because as you try to do large-scale proteomics, you would, at some point, need to tap into places like proteomic biobanks and et cetera. And in those cases, sample is very precious. And so the less -- the lower amount of sample you have to use, the more easy it is for adoption. But where Seer add value is, for example, even if you look at the timsTOF SAP is, you put a Proteograph upstream to a tims Pro or an SCP, suddenly, your content goes significantly up. So you always get significantly more content delivered because you're essentially going deeper into the proteom than any mass spec can today, but we are seeing innovation happen. Now the reason that's important to consider is because these newer technologies that are being developed, I named a bunch of companies that are after them, they don't need to be better than the mass specs of today. They need to be better than the mass specs of the future when those products are actually hitting the market. And the thing is, these technologies are constantly innovating and constantly getting better. So my sense of it is that there's a lot of room and a lot of leg on the existing mass spec technologies that are constantly improving. Of course, for the sake of science and medicine and patients, I hope the other ones also become successful, though we'll have to see how that movie plays out.

Got it. It's been several quarters now since you launched the Proteograph, Omid. Can you just talk to us about what does the sales funnel look like? And do you figure out you need distinct approaches across your customer types as you go from lead generation to sale?

Yes. So we've now -- we started off -- if you look at our commercial strategy, it was a collaboration phase. We had 4 customers in that. These are kind of their thought leaders who tend to publish and talk. In fact, some of our collaborators and customers have been very vocal about -- speaking about our technology. Then we had Limited Release phase, and those customers, we said, we would be hunting single-digit Limited Release customers. They tend to be kind of very early adopters. In some cases, they had more than one instrument that they would purchase right out of the gate. In some cases, they would come in to a minimum number of sample sizes. And then we had the Broad Release, which we went into at the beginning of this year. Now the -- if I look at the funnel of that customer, what it looks like is, it's roughly split 50-50 academic commercial. But it's interesting because if you then look at how fast could they move and that's also changing in the macro environment is that while the funnel is about half-half and there's a lot of interest in both buckets, the commercial entities could move faster than in academic entity, which, of course, makes sense because a typical economic entity would need to have grant funding, et cetera, and that grant cycle kind of delays their ability to move versus a commercial entity. And so if I then look at it in the context of, okay, so what does the actual close looks like? It's more like a 60-40, 2/3, 1/3 kind of split in terms of commercial versus academic. Now I think that's going to kind of flip back probably closer to maybe 60-40, half-half as academic entities begin to get grant funding. And we're seeing, for example, our early folks are now beginning to present data, the kind of data that makes its way into RO1 and other grant funding mechanisms. And of course, the process that you take to sell these things they just -- is evolving. In the early days -- first of all, we are in early days. Let me say, in the infancy days. You almost need it always to have third-party data, and in the absence of that, they would say, hey, would you run my own sample for me in a proof-of-principle study. So we would do these POP studies, except we said, look, we don't do POP studies for free. Any POP studies we do, we consider a service. So we would do these POPs and then they would buy it. We're now seeing width of an emergence of data and width, I mean, for example, the first half of the year, roughly about 30 year or so presentations. And that's -- and that was all the presentations we had all of last year, for example. So the number of presentation is increasing. There's now a handful of customers that now are sitting on manuscripts that are going to be begin to get published. So we're going to begin to see third-party data, not our data get presented and published and we're seeing that. So I think the need for those POP studies will decrease. So our sales process will evolve some. We've always said, we're not interested in the service business. We want to have a distributed business, place the instrument, sell the consumable. Of course, there are some customers that just that are not interested just by virtue of the way they operate to have those CapEx and infrastructures. We created COEs to kind of give access to the technologies to those kind of customers. So we are seeing, Tejas, a little bit of a change in terms of the dynamic of that. Now that is clouded by the macro picture because it's used -- I mean the CapEx is in a high-price ticket item. It's a couple hundred K. It used to be that in the pharma environment, a Director-level person would have approval authority to buy that instrument. Now we're seeing kind of like that CapEx kind of get bumped up a notch in terms of a decision process. So we're seeing a little bit of a change. I think that's because of the macro part of it. Anyway, that's the dynamic that's visible.

Got it. Quickly on new products, Omid. I mean you recently launched the sample-specific variant peptide Workflow. Any early user feedback there? The PAS 2.0, that's the one I'm referring to. And then beyond that, I mean you've talked about sort of throughput being the loudest ask from your customers. But how are you thinking about the cadence of future iterations, focusing on things like perhaps sample input or increasing content, et cetera?

Yes. So there is a product coming. Just to understand, the customers ask largely -- I mean they ask ton of stuff, but largely has been in 3 buckets: hey, I like to put lesser of my sample amount because it's precious to me; and two, I want to get more content than you're giving me; and three, I wanted to go faster. Now when we started this process, this game, the key bottleneck, the absolute biggest bottleneck was the upstream Workflow. The Proteograph eliminated that. Suddenly, the upstream Workflow went totally away. We went from largest study being 48 samples to now having the DLS for their consortium of building capacity to 100,000 samples a year. So then the biggest Workflow became -- once this data comes, like analysis was tough. And analysis -- and there's a lot of great software innovation that was happening, but this was in the hands of experts. Remember, proteomic was an expert-driven field, total experts. This was not a user-friendly platform where any lab could adopt a technology and began to run proteomics. It's literally sat in very concentrated hands, and a lot of that was because of also the data analysis. The idea for the past was to facilitate the way you go from an output of a mass spec to a readout with biological insight. And the past really did that. It's literally like going from, let's say, DOS to Windows in terms of usability. What we heard was now that proteomic, if you would scale content increasing and potentially impedance matching with genomic, and I say, potentially cautiously, but I think we're going to absolutely be there. And in fact, theoretically, we're there today. Could the platform integrate these 2 classes of omic information? I said the past did that. Early -- very early feedback from the proteogenomic folks like a DLS is that why this is going to be amazing. Already, they're asking more from us because now there is a genomic information. They're not saying, well, could you add transcriptomic content in the two from a proteogenomic? So what we're iterating, Tejas, I think you will see another iteration of that platform also come. So I think it's amazing. And by the way, I think it's just a matter of time for us to integrate other omics in there because the future is multiomics and you've got to be looking at the data in totality of all the omics. So that's that. Now in terms of the way we're thinking about our own -- we did the software -- now where is the limitation, now the limitation, now the upper bottleneck is the mass spec bottleneck because the upstream was sold. The downstream, I think, has now become a lot easier. The mass spec remains the significant throughput barrier, and we're innovating to decrease the -- if you would, the LC time with the mass spec in order to kind of reduce the mass spec work for a problem. Now the mass spec guys themselves like SCIEX with their platform and Thermal with theirs, they're doing that on their own, but we're also contributing and facilitating to minimizing the number of injections in terms of increasing the throughput. I think the upcoming product, Tejas, we may have disclosed at the last earnings call is actually focused on the throughput, right? And the next thing that's important to me that I think will make a big difference from a user perspective is content.

Got it. Over sample input.

I think so, Tejas, because I think we can -- almost if you would, in a software update, it's kind of a thing like -- I'm just using the Tesla improvement model, for example. I think we can improve the sample volume a lot easier than the content. So our focus as a team is heavy on the next iteration of the product being more about content side.

Got it. Makes sense. Quickly on the proteogenomics consortium. Omid, I mean, I think you were in the midst of the installation and validation process. Is that now sort of behind you? And do you still expect to begin running initial samples in the fourth quarter?

Let me have David answer that because David has been most closely tied with the DLS folks and lead that effort. David, do you want to take that?

Yes. Sure. So Tejas, we are working closely with the folks at DLS and SCIEX to stand that up. They are still working through their installation-validation process. They've got a facility up in Boston that they are building out. So the Proteograph is there. Their SCIEX, ZenoTOF are there, and they're just going through their kind of standup process on that. As they've said, they're going to run a small number of samples and then move into their -- what they've announced is a 500 sample kind of validation/marketing study. And then they do anticipate running customer -- their first customer projects before the end of the year, so in the fourth quarter. So we're still on track for that, and we're super excited about it. They've got a new Head of EVP Proteomics, who's kind of leading their charge up there. And they've been great partners, and they really wanted to get this thing moving.

Got it. And on that sort of the revenue contribution statement, how are you thinking about Prognomic? I mean they've contributed about 1 million per quarter. How do you see that sort of scaling over time? And do you view that is a proof point of a fully scaled customer in terms of what some of your higher tier sort of power users could get to?

Yes. Yes. So let me break that down, and I'll come to the answer around Prognomic. But really, what we're seeing from overall customers is that there's going to be a wide distribution, right? You're going to have a really large number of customers who are running -- I don't know, it's scale for them means hundreds of samples rather than thousands. And then we're going to have folks probably in the middle to the larger end who are going to be running thousands of samples, and then you're going to have folks at the most extreme like the proteogenomics consortium that's running tens of thousands of samples, right? But the distribution is not an even distribution as you would get. So I think today, I think we expect Prognomic to be fairly consistent at that level. It certainly has the potential to go more, but again, they're also project-driven, right? So as you do projects and you come off one project and you go to another, you could get some variations in that quarter-to-quarter, right? As they may pause to like we're going to analyze the data, but then they also could ramp it up and start looking at larger cohorts. So -- but for what we can see, it's going to be pretty consistent. It's certainly in the near term, and obviously, we want to expand our non-Prognomic customer base such that Prognomic as a percentage of our overall revenue continues to decline.

Got it. I'm going to take you in a world and tour in terms of thoughts on each geo economically. I mean you talked about this like elongating sales-cycle dynamic. Is that sort of focused on any specific geography? Or is it more broad-based? And then are you starting to see any specific pockets of weakness in Europe, particularly given the energy crisis, some shift in budgets and government priorities towards subsidizing energy versus academic funding, et cetera? And the same question on China. I mean different context, with the COVID lockdowns, they are continuing, what are your conversations with Enlight Medical suggesting? What are they seeing on the ground?

Do you want me to take that?

Please.

Okay. A lot there, Tejas, to unpack. But look, I'd say, broadly, the macro headwinds are global, right? I mean everyone is facing the inflation, supply chain issues, and so we're seeing a slowdown globally. I will say, the most extreme is in China, right, because of the lockdowns. And it certainly had an impact in terms of just elongating time frames of which we're -- us and in lighter looking at the sales cycle there, just given that people aren't in the labs and they're locked in their houses. So that certainly had an impact. I'd say, the second worst, if you will, is Europe, right? They've just had a little bit more difficulty. So again, things are elongating there. And then even here, as Omid had mentioned in the U.S., people are just being more thoughtful, right? No one knows what's coming, and so everyone is being a little more cautious. So whether that means now I can't approve that CapEx budget, I got to go get my bosses approval for that CapEx. That just slows things down, right, as they look at things. So again, we're seeing it globally, but it's something where we're working hard to try and mitigate that as we can, things that are in our control in terms of being able to, again, kind of push forward with those prospects through the funnel.

Got it. And the last 30 seconds here, David. You still feel pretty good about $14 million to $16 million this year and sort of the cadence that the Street has you doing in the back half?

Yes. So we -- as we said on our last earnings call, we reiterated our guidance of $14 million to $16 million. We were $6.9 million through the first half. So we feel good about where that is, and we have said that we expect it to be back-end loaded. So again, we feel good about where we are, and obviously, the macro is the unknown. So we'll have to navigate that, but we're comfortable with where we are now.

Got it. Well, that's a great place to leave it at. So thanks so much.

Thank you, Tejas. Appreciated.

Thanks for having us.