TD Cowen Healthcare Conference. I'm here with Omid Farokhzad, CEO of Seer, and David Horn, CFO of Seer. I'm Kyle Boucher, an associate in the Life Science and Diagnostic Tools team here at TD Cowen. So with that, maybe we can just jump right into Q&A. So just to start, 2023 was a difficult year. The macro environment continued to be challenged for capital purchases and nascent technologies. That said, Seer continued to make progress at the end of the year with $16.7 million in total revenue, a bit better than the guidance that you had set at Q3, where you're pointing towards the lower end of the guide. So from a high level, can you discuss how 2023 played out relative to your expectations? What were some of the key puts and takes for the year?
Kyle, thanks so much for having us. So as you said, we finished the year with 8% growth, 16.7%, which from my expectation, I would have liked to do a lot more. I think there is really a disconnect between how well the technology is doing and kind of what is happening in terms of commercialization. I'll get into why that is in a second. We did make a number of advancements for the year. One, we launched the Proteograph XT at the ASMS in June. We launched the Proteograph Analysis Suite, the PAS 2.0, that lets you analyze data, really make it very usable, not for the mass spec experts, but really for the biologists or folks that are today in the genomic space, to be able to digest the proteomic content that's coming off the Proteograph for themselves.
We launched the protein catalog library, which is the proteins that you can see with the Proteograph in plasma. That's a list of 10,000 proteins that can let you, for example, if you're looking at your data, what proteins, what pathways are they connected to, just helping you connect the dots. We launched the STAC, which was a Seer Technology Access Center, letting a customer go from a sample to data in a service model. We don't want to be in a service business, but we thought that was important to lower the barrier for a customer to access the proteomic content that comes off the Proteograph. And by the way, the STAC has been a real hit. I mean, from the beginning, we said we're not going to build a lot of capacity, and we're not going to expand the capacity.
We launched this STAC in collaboration with Thermo Fisher Scientific. The STAC has done great, 100% quarter-over-quarter revenue growth. We now have a backlog. All of Q2 is locked, and now we're splitting into Q3 in terms of STAC. But that says we're not going to add capacity because if we do that, we're going to push ourselves toward becoming a service business. This is not what we want to do. We launched the SIP. That's the Strategic Instrument Placement Program that lets a customer essentially take in a Proteograph as a loaner though the intent, obviously, is to sell it. But then it's tied to a consumable purchase upfront, again, lower the barrier for a customer to access data. And then on the commercial side, we expanded our geographic kind of reach, signed four distributors to help us penetrate more broadly.
As you know, we have boots on the ground in Europe. We have boots on the ground in the U.S. and Canada, but the other geographies we work with distributors do that. So overall, and we expanded our Centers of Excellence Program with the addition of Panome Bio and the Vingroup partners. So a lot to still do, Kyle. We're making progress. I think the key for a really kind of jump-starting adoption is going to be the evidence of biological insight that's going to trigger, that's going to start turning that flywheel. And we also saw a number of papers get published on preprint and bioRxiv, including one on medRxiv that showed, and we can get into that later.
But then the first actual customer paper just got published about 2 or 3 weeks ago in Nature Communications, two more coming later this month, one in Nature, another one in Nature, and another one in Nature Communications. So customer papers are now beginning to come, which should not be surprising because it takes a couple of years to finish a study. It takes a few months to go through a review process. And we're a young commercial company, so this is just the time frame where the customer studies are getting done to a point where they're beginning to be published, and we're seeing that start. And so the momentum for publication, I expect, will continue to rise going forward as the customers that have had access to the platform.
Got it. So on that, what are the key performance criteria in these publications you're hoping to see? And how will the broader proteomics community view the performance, and how should we consider the impact on catalyzing demand for Proteograph?
I think probably the most important thing to keep in mind is what does the Proteograph give you that you cannot get using another platform? And so what is the differentiated, unique biological insight that comes from the Proteograph Product Suite? So in that context, let me refer to maybe three customer publications, two on bioRxiv and one that is now published in Nature Communications. Maybe I'll even hint at the two that are coming in Nature and Nature Communications to make those points. One is that there was a study published by PrognomiQ in medRxiv that is now being submitted for peer review into a first-tier journal. That paper was a 2,500-sample study looking at patients with lung cancer or comorbid controls. These are smokers or healthy controls, and looking at multiomic classifiers for detection of early-stage lung cancer.
What you found in that study is that the multiomic signal, largely driven by the unbiased proteomics that you get with the Seer platform, gives you an AUC of 0.96 for classification of lung cancer versus healthy. That translates into a specificity, sorry, sensitivity of 81% for stage I lung in an 89% specificity. So these would be patients that would be high-risk patients that would normally be recommended to get a low-dose CT, but compliance rate is very low for that. They would get a PrognomiQ, sorry, they would get the PrognomiQ test, and then based on that sensitivity, they would be recommended to go for the LDCT in that case. Okay. So if you look at what drives classification, it's multiomic. So you have RNA, you have metabolome, but you also have unbiased proteomics that are in the top proteins in the classifier.
That AUC and the sensitivity, the specificity that you see with the PrognomiQ test is best in class. It's better than Delfi, and it's better than the Guardant in terms of the signal that you see for early cancer versus comorbids. All right. Then there was a Mass General study looking at Alzheimer's disease, and MGH had longitudinal samples on 1,800 samples looking at a 10-year span to see is there patients that are cognitively intact, but they'll end up getting a decline and eventually getting late-stage AD, Alzheimer's disease. And so that 1,800 sample was done in collaboration with Seer. 138 proteins were identified that over the 10-year span either go up or down with cognitive decline. Interestingly, 55% of that 138 protein is not in the high-plex 5,000 panel that is commercially available.
Of the 138 proteins, only 44 of them are known to be associated with AD, so another 94 are putative new biomarkers of AD that you would never know to go look for. Then MGH folks did a very interesting follow-on study on their own, unrelated to us, which was: Is there a subset of those 138 proteins that if you had them could dictate that you would be a rapidly declining patient? So they identified 8 proteins that if you have those 8, you would be a rapidly declining AD subject. Of course, then you would be prone to getting novel therapeutics that may be available to you. Again, 6 of those 8 proteins are not available in a commercially available high-plex panel. The final paper I would highlight is the Karsten Suhre paper published in Nature Communications looking at pQTLs in diabetes.
You identify new pQTLs using the Seer platform that you could not identify using targeted approaches, and also some ability to be able to detect an accurate or correct pQTL because if you're looking at pQTL, meaning that you're looking at genomic variants and then their correlation with up or down regulation of proteins, if that protein itself is biologically changed, either through a mutation or through a post-surgical modification, that no longer binds the targeted agent, then you would miss the up or down regulation. But because with the Proteograph, you're looking at amino acid and peptide level resolution, you actually see true pQTLs, and you're able to identify false pQTLs. That paper was published by Karsten Suhre from Cornell in Nature Communications. Those biological insights are not possible to do using the alternative solutions.
If you try to do them using conventional mass spec approaches, it doesn't work because it doesn't scale. If you look at the PrognomiQ study of 2,500 proteins, they go to a depth of about 7,000-8,000 proteins in that study. To go that deep, the number of fractions that you need to create using immunodepletion and fractionation in the conventional approach, I mean, that study could have taken multiples of years to complete, just totally impractical.
Got it. Got it. So just want to move over to the technology a little bit here, some specifics. So the Proteograph XT Assay Kit you came out with last year, you said the product enables analysis of up to 10,000 samples per year with one XT in two mass specs. Can you talk a little bit about what this new kit has enabled customers to do above and beyond the original kits, and how has customer adoption been of these new XT kits?
We've now converted 75% of the existing platforms that are installed to an XT. The feedback has been fantastic. You can now do large-scale proteomics. I mean, today it's now possible to do population-scale deep unbiased proteomic. I mean, if someone said such a thing to you before the Proteograph was launched, you would think that they were just completely clueless about how these things get done. But this is now totally possible to do thousands or tens of thousands of studies with relatively modest CapEx investments.
Got it. So on that, if we look out, sorry, 12-24 months, what does the headroom look like for the technology? Is there anything we should expect in terms of new products, improvements, etc., over the next 12-24 months?
Yeah. So the XT, again, was 2.5-fold improvement in throughput just in terms of number of samples. We went from 16 to 40. In terms of what's the pipeline going forward, there's basically three areas that people kind of have asked us to look at. One is throughput. One is content, so get deeper in the proteome. And the third is sample volume. And I think XT addressed a lot of those. It certainly was similar, if not a little bit better, on content, certainly much better on throughput, still the same sample volume. So there are areas where we can continue to improve on throughput, content, and also sample volume. And so you can imagine a roadmap, Kyle, that over the next, I mean, again, we just launched XT last June, so it's still got some legs on it.
But over the course of the next, I don't know, 12, 24 months, that there are new assets you can come out with. Certainly, we also have our Proteograph Analysis Suite, which is the software, which you'll continue to see improvements. It's a cloud-based solution, so it's pretty easy to launch and roll out updates for that. People have asked us, last year or 2 years ago, launching the proteogenomics component of PAS. And so people have asked us to look at other analytes to maybe marry with the proteomic data in terms of analysis. So we're working on that. We're also working on our biologist customers. Right now, PAS is programmed to really speak to the mass spec user. Some of our biologist users have asked, "Just tell me the insight I need to see." And so we can see us rolling out versions of that.
Across hardware, assays, software, I think it'll continue. We've got a long roadmap of product innovation we can do.
Got it. Got it. And maybe going back to you talked about the STAC and the SIP opportunities earlier. Can you just talk a little bit about how many have adopted the technology under the SIP program? And maybe what does that look like as those relationships sort of progress? Is there a point in time where they have to purchase the instrument, or how does that ultimately play out over the next year or two?
Yeah. So what we've said on the SIP is we shipped 23 instruments last year. About a third of those were SIP folks. And so those are, again, strategic instrument placements, i.e., we expect that these folks will ultimately buy the technology. And it's really a question of funding cycle. So they say, "We're super interested. We don't have the funds right now, but we're going to apply to get the funds," or, "We're going to budget for the funds." So we generally give them a loaner for, call it, 9-12 months. But they do have operating budgets, so they purchase upfront kits. And so we do get revenue from it.
And then what we're trying to do is get them started and let them generate data so that they can then make the case of again, it's all the customers who have data and have seen data love the data. So we want to get people exposed to that as quickly as we can. And then I think on the STAC, we said at the conference earlier in January, said we had about 48 projects that we did over the first six months of last year, and quite a number of large pharmas were part of that. So great traction on that, as Omid said, and we've got a backlog there, which is great. And so again, in terms of trying to bring down these barriers to access, it's working, right?
Got it. Got it. So moving over to guidance for 2024, you got a 2024 revenue, 14% growth year-over-year at the midpoint. Can you sort of discuss the key drivers in terms of instruments, consumables, funnel conversion, etc., and what can you point to for revenue growth acceleration beyond this year?
Appreciate that.
So in terms of the guidance, really, it's based on a few things. One, STAC will become a bigger part of our revenue, right? So while I would think about it in terms of maybe a quarter-ish of our revenue would be STAC this year, and then the balance will be consumables and instruments, kind of the majority of that being consumables, as you would expect from a revenue perspective. And so really, what are the key drivers of that? I think one is continuing to see people wanting to do these larger projects, which we are seeing. We're having quite a number of conversations with people that want to do two, three, four, 5,000 sample studies, both either themselves or in conjunction with us, either running the samples themselves and sending us the work for the mass spec or having us do kind of the full range of sample-to-answer.
In terms of just driving adoption of the instrument, the STAC and SIP obviously help that. They feed that. And we are seeing some conversion there from that. But I think, as Omid alluded to on the papers, I mean, we've seen whenever a person posts a preprint or gets the paper published, we get inbounds. People see that. They notice that, and they give us a call. Obviously, we're out there with our salesforce making calls, but it certainly helps when you see some of these prominent KOLs in a prominent publication. That certainly helps with the inbound efforts too. So continue to drive the revenue growth with those publications.
Kyle, I just want to add. I mentioned that we don't want to be in the service business, so we're not going to add a ton more capacity. In fact, I like to add no more capacity in the U.S. despite the demand for the STAC. If there's more interest, we're going to push them toward the centers of excellence and let them do it. In other words, I will sell the kit to our COE, and the COE will run the sample. We'd rather do that instead of doing the services ourselves. Now, that said, the STAC does something fantastic, which is it gave the customer the ability to touch and feel the proteomic data off the Proteograph, and then it's easy to become a converted. With that said, because of GDPR rules, we can't take services or samples from overseas.
As you can imagine, there's obviously a lot of demand in the U.S., but there's also demand ex-U.S. You can assume that we are thinking about geographic expansion of the STAC to be able to capitalize on some of the customers that are outside of the U.S., giving them the ability to firsthand experience the STAC services before we hand them off. As we know more about that, we'll make an announcement about that. Directionally, we're thinking about that.
Got it. What does your sales funnel look like? How has this changed over the past year? How has this been impacted by the macro environment, bio-pharma budgets, funding, NIH budgets, etc.?
I would say that if I look at the funnel, it's probably 60/40 academic tilt. But if I look at, it's about 50/50 academic-commercial. But if I look at sales, it's probably 60/40, 65/35 tilted toward commercial. And by the way, if I look at large pharma customers, they're tough not to crack because in this macro environment picture, they don't want to take on new platforms that are not very validated. But once you break into those places, they're actually excellent customers. It's a run-rate business, recurrent revenue, quarter after quarter orders from those. So my expectation is that STAC will continue to grow to about 25% of our revenue for the year. And then the mix of academic and commercial will continue to look the way it does now, which is 60/40, 65/35 tilted toward commercial.
Got it. Okay. So moving on a little bit to the competitive dynamic side, there's several companies out there vying for traction in the proteomic space from discovery tools like Olink and SomaLogic, which is acquired by Standard BioTools, Alamar Biosciences. There's protein sequencers like Nautilus Biotechnology and Quantum-Si. Now, there's newer multi-omic spatial players in the spatial side like Akoya Biosciences, Standard BioTools, 10x Genomics, Element Biosciences, not to mention ongoing performance improvements from mass spec vendors like Thermo Fisher Scientific with the Astral. Where does Seer for today, and is your competitive profile differentiated enough to really stand out amongst this broad field of competition?
Yeah. Look, let me give you my simplistic view of the space, which is there are three buckets. Those who said that mass spec is not a good detector, I'm going to develop a better one, Nautilus, QSI , two public ones, but Encodia/Erisyon on the private side. And then those who approach it in a targeted way, Olink, Soma, I would say Quanterix is another one that would be in the same bucket. And then the ones that say, "We're going to approach this in an unbiased way at scale." And at scale, unbiased, there's only one player in that space. That's been the case when we went public. It is the case today. If I look in the rearview mirror, I don't see anyone catching up on us on that front.
Now, how I look at the competitive landscape is while today the Proteograph sits upstream to a mass spec, we're detector-agnostic, we're mass spec-agnostic, meaning we've placed our instruments upstream to the first-generation timsTOF, the next generation, and then most recently, HT. The timsTOF has gotten better and better. But if you look at how much improvement the Proteograph brings to that instrument, that multiple improvement has been consistent. If you look at the Thermo, same thing. We were ahead of their early instruments and on the later and now most recently, the Astral. Again, if you look at the improvement that the Proteograph provides, that's been consistent. By the way, no better flattery or no more than the fact that when Thermo released the Astral, the launch at ASMS, that launch presentation actually included the performance of the Astral with the Proteograph.
The Astral on its own in plasma does about 760 proteins. The Astral with the Proteograph does about 6,000 proteins. If you look at HT, HT alone in plasma, about 600. HT together with the Proteograph, about 5,000. So the full improvement on top of the mass specs has been consistent, and we continue to add value. And we consider all of these guys our partners, and we work with them. Now, in the event that Nautilus eventually launches their product or QSI eventually getting traction, then that's actually good for us because the Proteograph will then go and sit upstream to their instrument. In other words, the problem we solve, none of the detectors have solved. That problem is a dynamic range problem.
If you take plasma and you put it on the Nautilus platform, you're going to read 10^10 albumin before you get to the first IL-2. What Seer does is it compresses the dynamic range, so you read less of the albumin and you read more of the IL-2. That is how we add value to the mass specs. It's the same value add that would go to QSI. It's the same value add that would go to Nautilus. By the way, the reason I don't think of it today in terms of putting a Proteograph upstream to their instrument is because I have no visibility to those instruments, A, being launched, or 2, getting market traction. But if they do, it's an expansion for us in terms of our opportunities. So I continue to be very bullish about what it is that we're offering.
The thing is, we weren't the 3rd targeted platform to come like Alamar, which is a very easy pattern recognition. Okay, Olink did that, so I'm coming in and you can follow it. What Seer did was fundamentally disruptive. There was no targeted sorry, there was no unbiased mass spec at scale possible before the Proteograph. Before the Proteograph, the largest unbiased proteomic study ever published in plasma was 48 samples. The deepest study published from Broad was 4,500 proteins. Today, we have multiple customers that are at thousands of sample level, 7,000, 8,000, 9,000 protein depth. The question that the customer asks is, "Well, what does this even mean?" So once the biological insight becomes clear, I think the traction will follow in terms of market adoption. That's the beauty of being first in the market. You're building the plane. You're flying it at the same time.
We're building the market as we're commercializing it into that market, and we're learning and we're adjusting. I think the STAC and the SIP are two very good examples of us looking at the market and reacting to it to better serve our customers.
Got it. So I have two sort of financial questions and then a conceptual one after that. But to start, Seer is sort of in a great position. You've got $331 million in cash and equivalents in the balance sheet at the end of the year. What's your burn look like in 2024 and over the next few years? And do you think you can get to free cash flow positivity with the existing cash on hand?
373, by the way. But David.
He's right on cash and market. We add investments long. You've got some long-term returns there, and it jumps it up to 373.
Even better.
Yeah, even better.
Even more, Kyle.
So we burned about $53.5. Actual cash usage last year was about $53.5. The free cash flow was around $66, but the actual cash use was around $53. So if you just if nothing changes, that's seven years of cash. So we've got plenty of cash. Look, do we think we can get to cash flow break-even? Again, it's kind of too soon to tell. We think it certainly is possible. But again, it just depends on investments we want to make and opportunities that we see, right, whether there's an inorganic opportunity or something we may look at. So again, I think we feel really good about our position. We feel really good about our burn. We continue to monitor that. We expect it to be maybe similar, maybe slightly down this year.
Again, it's something where we feel like we're in a really good position to try and drive the growth.
Got it. What should your margins look like in 2024? What should your margins look like in 2024? How should we think about that over the next 3-4 years?
Yeah. So I think we still continue to believe, as we said at the IPO, that this is a 70%-75% gross margin business long term. The margins this quarter were around 45% overall for the year. They were a little over 50%. We had a lot of instrument sales in the fourth quarter, which was good, but they obviously come at a much lower margin. And so I think as you think about it, consumable kits, really high margin, instruments, pretty low margin, the STAC kind of somewhere in between. And it's really going to vary quarter to quarter with the growth of the STAC this year. That's the reason we guided a little low on the gross margins, which is going to have a little bit of impact on our gross margins as we continue to grow that business.
So again, I still think fundamentally outlook has proven pretty good in terms of the gross margin structure. It's just a matter of driving the volume to get the margins in the right place.
Got it. Got it. So conceptually speaking, we've seen some M&A in the space, in part given large market TAMs, differentiated technology, but also some strong balance sheets on targets. So how do you guys create the most value for shareholders? Is it on a go-alone basis? Can you continue with your tech development, commercial plans? Or is it maybe seek an acquirer to buy Seer, generate more meaningful operating leverage that ensues with a bigger acquirer? What's your philosophy around that?
Look, I think from the very beginning, we thought that the value proposition that we have is on par to those who've made the biggest impact in the genomic space. And I'm alluding to Illumina here. And we said that Seer is going to span everywhere from content discovery all the way to clinical. If I look today, just three years after the launch of the instrument the first time as an early access, our largest customer, which is the PrognomiQ that we spun out, they're entering into an LDT later this year using the Proteograph for their test. So the notion of Seer going all the way from discovery to clinical use is actually happening. And so the opportunity is huge. Now, that said, scale is key. We do need to scale.
A lot of that is going to come in terms of getting market traction for our own platform, continuing to innovate, to be right on the edge of that innovation and build revenue and be very judicious about the way we spend money. We're extremely frugal. David said we've got 7 years of cash, assuming no growth and basically where we are. Let me tell you that is not the way we run the business. If we had no growth, I would cut our burn so much that it would not be 7 years of cash. It's 7 years of cash. If you just do the simple math of my revenue today was my burn, my expectation actually is that as we continue to invest in innovation and the revenue growth that I expect that we're going to have, that we will be a cash flow positive company.
Now, that said, we got calls all the time in terms of opportunities, whether we would be considering acquiring X, Y, and Z, some of them have passed that have been acquired by others. Now, that said, the reason for that disciplined approach is I feel very bullish about what we have. If I thought we had a failed technology with my balance sheet, I would go and merge with someone. If I feel like I have something extremely differentiated, then I sit put where I am. I invest in my own platform, and I will bring something in-house that is suitable and complementary to us. That has not happened yet, but we're always looking.
Got it. Well, I think we're out of time here. Omid, David, thank you so much.
Kyle, thanks a million.
It's great to have you.
Thank you, Kyle.