Mikson, Aquifer Life Sciences and Diagnostics for Canaccord. Pleased to present this fireside chat with Seer here with us today. Seer is a company participating in the exciting proteomics space with the Proteograph Product Suite. From the company, we have Omid Farokhzad, CEO, and David Horn, CFO. Thanks, guys, for joining us today. I appreciate it. So, yeah, so I guess, Omid, maybe just walk through the second quarter results that you announced recently, maybe from a you know update perspective, because, you know, not too much on the financial side, but I guess, you know, what's going on with the business as well, like kind of trends, things like that.
If I understood you correctly, you wanna talk about the most recent announcement that we did on Q2, and where we are in the business. So let me. I'll split that up into two. I'll talk about part of it, and I'll have David comment on the other part, which is, I would say Q2 was probably the quarter that we saw external validation of our platform be demonstrated more than ever before from our customers. When we launched the Proteograph back in the very, very beginning of 2021, there were really no customer publications. In fact, it took about almost two years for the first customer publications to come. The velocity of those picked up.
The first half of this year has been a time where, we're now seeing a significant momentum in terms of customer papers come out, and the validation of what is possible using the Proteograph is becoming increasingly clear, in terms of what, what can happen when you use a Proteograph. Now, that tailwind, which is helping us, is really offset by a stronger headwind, which is the broader macro picture difficulty for, companies like us, selling a relatively new instrument. So the totality of all of that translated into, a quarter that was flat to the previous one, and a year-over-year, reduction, in terms of revenue.
That said, my hope is that the trough is behind us, that with the increasingly higher number of customer publications coming, that the tailwind of that will actually be stronger than the headwind of the broader macro picture, and hopefully, we can come out of this stronger for the second half of the year. David, you want to add additional color?
Yeah, no, I, I think you summed it up well, Omid. I, I think just the things to note, you know, we are this year have a headwind of PrognomiQ, which is our related party, doing less as they've had great success with their developing their early cancer lung detection test, and now they're moving into LDT. So they're doing less kind of basic research, biomarker discovery that they did last year, and so we're seeing less revenue from them and, but we're seeing continue to see, you know, growth outside of PrognomiQ.
Okay. Maybe, David, just, on the kind of guidance, I think it was adjusted a little bit. So, you know, you guys initiated this STAC program about a year ago, or, like, close to a year ago, and that was kinda due to, you know, just it almost felt like kind of instruments and capital sort of related as well. And now you have, like, you know, further CapEx constraints and things like that going on. I mean, definitely different issues year- to- year, but, like, it's not, hasn't gotten much better. So how was it sort of, creating guidance early this year? So, like in, I guess, February or whenever you had your-
Yeah
... earnings, I guess, early in 2024, and then also, like now, with this update, how has the environment changed from your perspective?
Yeah, you know, I think we were more optimistic at the beginning of the year that we'd see some loosening in the CapEx funding environment. But I think we've really haven't seen much movement or improvement in that, certainly over for the first half of the year. So I think really what was driving that is we've got some really nice opportunities in the pipeline. The question is: How quickly are they gonna move through the pipeline? And they're just, it's just kind of, you know, grinding in terms of elongating the sales cycle and putting a, you know, a quarter or two delay on that, as we've said on our call. And so we're just being, you know, abundance of caution in terms of the back half of the year.
Now, that said, we are starting to see some more positive conversations, where people are willing to engage with some of these publications that Omid mentioned coming out. We do have people that are, you know, reaching out to us and do wanna have these conversations, but it's just unclear how, how quickly we can move those. I think. You brought up the STAC. I do think that that is a positive 'cause it's a lot easier for people to consume it as a service, and so we are seeing good demand from the STAC. But ultimately, our objective is to not be in the service business, but to, you know, sell instruments and consumables.
So it's a good tailwind for us, and we'll just see how long the overall CapEx cycle takes to loosen up.
Okay, and then, Omid, you've referenced some like publications coming out or I guess like you know in the works or maybe in the past. So what's maybe we'll just double-click on that. What to expect there? Really just like the impact that those could have and why those are impactful like what kind of data is gonna be in there and what types of customers are really like generating this data in the first place and how are they using the Proteograph Product Suite?
Look, scientists, I'm one of them, are inherently skeptical, meaning that you need evidence in order to agree with a point of view. And when we launched the Proteograph, I, as a scientist, felt, you know, extremely bullish about what we were doing. You know, and having been a professor for 20 years of my life, that carries some credibility. But at the end of the day, the customer needs to see what the Proteograph does in their hands and what are the biological insight that uniquely becomes enabled with the Proteograph that the customer wouldn't otherwise be able to do with an alternative platform.
That evidence then builds credibility, and if you do these in the right accounts, meaning the lighthouse account that others follow, then that can actually give you significant momentum in terms of adoption. So with that background laid out, we are now seeing some seminal papers come out that clearly differentiates the scientific data that comes from the Proteograph versus the alternatives. Before Seer, in order to do proteomics at scale, you were limited to targeted approaches in proteomics. And we had seen some very large targeted proteomic studies done, like the UKBB done with Olink or the deCODE study done with SomaLogic. But mass spec-based proteomics before the Proteograph was limited to studies that were typically in the tens of sample number, whereas those other ones were tens of thousands of sample number.
And so when Seer Proteograph came, for the first time, you can do deep proteomics at a very substantial scale. We've now had multiple customers complete studies over multiples of thousands. PrognomiQ is starting studies of 15,000 samples, and we're now in conversations with some customers that are talking about 5,000 or 10,000 sample studies. These were just fundamentally not possible before, and we're beginning to see some of those at-scale studies get published. So, for example, very recently, about two months ago, a study from Karsten Suhre in Cornell looking at pQTL, so a proteogenomic study, tying up genomic variants to changes in protein expression, get done with the Seer Proteograph, looking at studies of about 1,600 samples, going to a depth of 1,000s of proteins.
Karsten's study was extremely insightful in that it showed, number one, if you look at a mass spec that looks at a protein, and because you're looking at multiple peptides per protein, you can see along the length of that protein, that you could see pQTLs, meaning genomic changes that translates into protein up and down regulation, that the targeted approaches were missing. So that was interesting, but the more interesting part of it, right, this equally interesting part of it, was that previously published pQTLs for both Olink and Soma, in about a 1/3 of the time, they were wrong pQTLs, meaning changes in protein expression were incorrectly attributed to a pQTL, whereas very likely those were changes in epitopes of proteins or epitope effect that happens because of a genomic variant changing a protein conformation.
Keep in mind that an average human protein is 470 amino acids, and average binding site of a ligand is to about a five-eight amino acid long entity. So any changes anywhere else in the protein is missed, and those changes can change the epitope, and so the binding site changes, and you get misguided.
When that paper got published, there was incoming emails from some of the leaders who were actually involved in the UKBB study, saying, "This actually fundamentally changes the landscape, and that we really have to begin to do these similar types of studies at scale, at the scale of ten of thousands of samples, to really begin to understand it." So if I look at studies like the Karsten study or other ones that are going on right now, or the Chris Mason one that it was just published in Nature and Nature Communications about two months ago, again, 2.5 months ago, they allowed the customers to visualize what is possible and help with adoption.
We are now at a place where about every other week or so, a paper is getting published, and we actually don't even know about some of these papers that are coming, and this is a fantastic place to be because before that, we knew what was coming. We were waiting for it. We would actually reach out to the PI, wondering: How was your review going, and when do we expect this paper to come out? But now they're actually coming out, and our marketing team are identifying them because they're popping up on the PubMed, and I think that velocity of customer publication is only gonna take off, and with it would then come an ease for adoption.
Okay, that was helpful. I mean, the publication aspect is kind of like more... It's qualitative. It's hard to sort of like quantum- I was, like, I'm tempted to ask, like, how many publications do you need? How long do they have to be out there? But I guess we'll just see. Hopefully, it is like a tailwind.
Kyle, I would say the following, which is, publications come in two flavors, and they're not mutually exclusive. One is quality. So in other words, one Nature paper can really move the needle in a very different way than, let's say, you know, five or six or 10 papers that are much, much lower impact. But what's interesting is, if I look at the papers that have come to date, they've all been at a very high level, so Nature Comp paper, two of them, Nature paper, two of them, Advanced Materials, impact factor 30, PNAS. So the publication quality is high, and I think the reason for it is that the content that's coming is highly novel, and it's ending up in high places.
I think if we have the velocity of publications that are coming and the quality of publications coming continue, I really think that the headwind that we're gonna be facing in the macro, us and every one of our peers facing, is probably going to be weaker than the tailwind that is publication pushing us forward in terms of adoption.
Got it. Okay, and then also, something interesting that you just mentioned was, like, the population proteomics potential of all this, and, like, maybe just talk about if it's possible, if, you know, Seer can be included in, like, some of these, like, larger population-scale projects, like a UK Biobank, for example, so.
By the way, not only do I think Seer can be included, I actually think Seer should be included. I think the most valuable commodity of these biobanks is their samples, and to utilize a sample, and to not be the beneficiary of the content that comes with a Seer is almost not good science. But the thing is, again, if you're sitting on a biobank sample, which is enormously valuable, before you commit that to a resource like Seer, you need to make sure the platform is extremely validated. So as these papers are coming, they're also lowering the barrier for these biobanks to then say, "We will actually do a study with Seer," because they're not compromising their sample using a platform that is not useful to them.
So I think as the validation comes, in fact, my prediction is that we will see population-scale proteomics get done. I always thought it would be done this year. The year is coming to an end, and I don't think it's gonna be done this year, so I will then wait and say maybe it's next year. But I do expect the population-scale deep unbiased proteomic to be available to us in terms of the data, probably hopefully by the end of next year.
Okay, then, and how, like, dependent on, on that is some of these like, almost like next gen, like, really exciting mass spec instruments, like the Orbitrap Astral and, like, the timsTOF, the next generation timsTOF, things like that? Like, I mean, I feel like, those need to be, like, you know, you need to, like, be able to improve upon what they can do. Maybe like, maybe just debate that, I guess.
Yeah. Look, mass specs are getting better and better. Just in the last seven years that we started Seer, roughly three years that the Proteograph been in the hand of customers, we've seen substantial improvement, and probably the most pronounced improvement in mass spec is the Orbitrap Astral from Thermo Fisher. That really moved the needle in terms of performance. As amazing as the Astral is, when you put a Proteograph upstream to it, the performance of the Astral in plasma goes up by seven-fold. So there's a meaningful improvement. So as the instruments are getting better, the fold improvement of Seer upstream to that mass spec continues to be the same, and the reason is, Seer fundamentally solves a different problem than a mass spec does. We solve the dynamic range problem, and they solve sensitivity of detection problem, and they're very complementary.
So I don't think their improvement needs to be there for us to do a population scale, because at the current improvement of an Astral and a Proteograph, we are able to do a population-scale proteomic study with very nominal CapEx investment. So the existing platforms today, the best mass specs and the best assay from Seer, which is a Proteograph XT, is necessary and sufficient to do a population-scale study. The bar will only go down over the course of the next 12-24 months as we launch our next product, as Thermo launches their next product, it'll get easier and easier, but it already is possible.
Okay, I guess, like, that was helpful. Also, the part of the question is, like, is the mass spec install base, especially these newer instruments, kind of broad enough at this point to enable, like, a population proteomics project that kind of can enable, like, an add-on Proteograph Product Suite as well, kind of?
Remember, usually when you look at these population scale studies, they tend to be centralized in the way they get done. When you have a large install base, it helps in the distributed model in terms of adoption, not necessarily for those population scale studies. So I think if, for example, if we were to do a FinnGen or UKBB or like, or similar types of a study, it will likely be that we will dedicate X number of resources in terms of a Proteograph and mass spec and do that in partnership with them. But that you don't necessarily need a very broad install base to get that done for those types of studies.
But to your point, though, in terms of adoption of our technology going forward, absolutely, there is a massive install base of mass spec today, you know, that does proteomic work. Roughly of the 55,000+ installed base of mass spec, about a 1/3 of them do proteomic work. So there is a massive install base, and if you look at that, you know, I don't know, roughly 13,000 mass specs-14,000 mass specs that does proteomic, and look at the install base of Seer, we're at the very, very beginning of this journey. So I don't think that's gonna be our bottleneck. I really think the key bottleneck to adoption is gonna be basically the publication and the adoption from the, from the customers when they see the evidence and the value add to them.
Okay, that was great. And then just kind of, before the acquisition of Olink by Thermo Fisher, like, you know, closed recently, we would get a lot of questions about the overlap with those two technologies. I mean, there was some like, you know, global regulatory entities that were like, kind of, looking into that, basically. The key question was, like, what, what happens long term? Like, like does, does this protein measurement, does this targeted protein measurement from Olink, you know, overlap with maybe mass spec from Thermo, things like that over time? Now, how do you, where do you position, like, Seer, maybe, like, long, long term? We're talking, like, five-10 years going forward.
Like, you know, could this be, end up being some kind of like a, not necessarily like a holy grail of proteomics, it's not like that, but just some sort of like a end all, almost like an end all be all, kind of like a really, you know, comprehensive approach to, like, solving some of these, solving some of these proteomics problems that we kind of see today?
David, what do you think?
... So, look, Kyle, I think it's a great question. I think, you know, Thermo even said this in their press release when they acquired Olink, which is the two technologies are very complementary. And they, they view them as such, right? They view that targeted. You need both, right? I mean, you need the unbiased mass spec-based technologies, and you also need the targeted. The targeted are good when you know what you're looking for, right? When you have a defined set that you know what you're looking for. I mean, you know, personally, if I was Thermo, I would kind of refashion Olink to be more kind of clinical targeted focus rather than, you know, quote, "discovery focused." And because to understand and discover novel content, you definitely need an unbiased mass spec-based approach.
It's the only way you're gonna see new content, 'cause by definition, an array has a limited set of content, right? You're kind of fishing with the fishing pole rather than the net, right? So you're not really ever gonna discover new, pure new content, new biomarkers for disease without an unbiased approach. So, but you need both, right? I think they're both very complementary. The good news, it's a, you know, rising tide lifts all ships. I think it, you're gonna continue to see expansion in, in both markets. And I think if you talk to Thermo, they'd tell you the same thing, that they see them as very complementary, and you, you kinda need both to, to solve the problem.
Okay. Yeah, I mean, I guess just from the Seer perspective, maybe omit, like, where does unbiased - there's not, like, a lot of unbiased methods today for proteomic analysis. I guess, like, the mass spec type stuff. There's, like, there's ways to sequence a protein using mass spec as well. And then there's, like, those next generation, possibly de novo protein sequencing approaches that, you know, kind of exist or could, like, maybe get better and better over time. But again, like, you know, this is a pretty, like, this is like, this could be very exciting, Seer, right? You talked about, you talked, you talked about it for years. I right. Like, I mean, where could this go towards in terms of-
Yeah
... being able to, like, help the world?
Look, I think the question is, is there other detectors that are coming? And, of course, there are folks working on it, Nautilus, QSi on the public side, Erisyon and Encodia on the private side. And I've always said, for the last four years, that as these companies are developing their platforms, they have to be better, not compared to the mass specs of today, but to the mass specs that are available at the time this product got commercialized, so the mass specs of three, four, five years down the road. And of course, mass specs are getting better and better. Now, that said, my expectation is that new detectors will come.
But as detectors come, every detector, at least the ones that are being developed today that I have visibility to, has fundamentally the same problem, which is that proteins are not only different in terms of their structure, which is how a detector can separate them and compare them to each other, but proteins are also different in terms of their quantities. So unlike genomic sequence or where the genes are lined up and you sequence them, in the proteome, the most abundant protein to the least abundant protein can be 10^10 order of magnitude or 10 order of magnitude difference in concentration. So if you begin to sequence that, you're gonna sequence 10^10 times of the most abundant protein before you can sequence one of the least abundant proteins. That is where Seer Proteograph solves the problem.
We compress the dynamic range quantitatively so that that detector can sequence the abundant protein, but also sequence the least abundant protein. That solution that we offer applies not only to the mass spec, but to every one of these other detectors, like the Nautilus, QSi, Erisyon, and Encodia. They have the same exact problem. So if they are to commercialize a product that does get market traction, then we will adjust the Proteograph to sit upstream to their market. But that all said, it would be very exciting for me to see that. I'm a little bit skeptical, as a scientist, that the traction would be such that would be relevant to us in terms of a market share to go after.
Okay, that was interesting. Thanks for that. I guess, like, we're basically out of time. Now, looking to 2025, we had the STAC the past year or so, that, that program. Hopefully, the macro headwinds are, like, you know, alleviating or something by, by next year. Then you have these, this traction from publications. It could be, like, an interesting year for the company, maybe 2026, but either way, 2025 looks, looks, promising. Maybe just, like, a walkthrough, like, your reasons to be excited for next year. But if you guys wanna tag team, then that's. It'd be fine.
All yours.
No, look, I think, I think, you know, we've, we've kinda said, you know, really what, what stage we're at, is trying to get that biological insight. I think with these publications, by the end of the year, we should have. You know, I think we've got 11 publications now. You know, we should, we'll obviously have north of that by the end of the year. I think that does give us good momentum going into 2025. And hopefully, if the macro, you know, headwinds alleviate a little bit, budgets free up a little bit, you know, I think we will, you know, we're excited about 2025. The other kind of interesting stat as the finance guy is that, you know, our commercial revenue is kinda 60/40, commercial versus academic.
But our pipeline, if you look at the opportunities in our pipeline, it's exactly flipped. So we've got 60% academic, 40% commercial, which to me says there's been a lot more interest from academia. They're doing a lot to get the funding they need in place for next year. And so we'd certainly hope to see, you know, more interest from kind of the basic research labs and having the funding to come and be part of it.
Okay, yeah, that sounds great, promising. Thanks, guys, for joining us. This was great. Appreciate it.
Thank you, Kyle.
Thank you, Kyle. Really appreciate it, Kyle.