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

All right. Good morning, everyone. This is Julia Qin from the life science team at JPMorgan. It's my great pleasure to introduce you to our next presentation by Seer. As a reminder, if you have a question, you can submit it through the website. And with that, let me turn it over to Omid.

Good morning, everyone, and thanks so much, Julia, Tycho and the JPMorgan team for the opportunity to present today. I'm Omid Farokhzad, CEO and Founder of Seer, a company which I founded back in 2017 and went public in December of 2020. I'm going to walk you through some of the foundational aspects of our company, the progress we've made in our first year as a public company and share with you why we believe Seer is leading the proteomics revolution. My presentation today will be followed by a Q&A with myself and David Horn, our Chief Financial Officer. Now please note our safe harbor disclosure, which indicates that this presentation may include forward-looking statements. At Seer, we have a bold vision of imagining and pioneering new ways to decode the secrets at a podium to improve human health. We're focused on developing transformative products. In the field of proteomics which have the potential to deliver a vast source of molecular information, but this enormously valuable content has remained largely undiscovered today. And we believe we're in excellent position to address this unmet need. Proteomics is the next frontier in biology. And every day, we're more excited by the opportunity in front of us. Now that said, doing so requires focus, clarity of approach and the right strategy. Coming into 2021, we were focused on a four-pronged approach for market expansion that included first establishing Seer as the premier provider in proteomics by setting the pace for innovation in this space. Second, enabling our customers and KOLs to build our technology with us as we develop customer intimacy, enable customer innovation throughout our phased commercial approach. Third, building an ecosystem around our technology really would select partnerships to enable access to deep unbiased proteomics by more labs expand geographically and positioned to open up the market. And fourth and most importantly, continue to build a world-class team with unparalleled expertise, experience to develop and bring first-of-its-kind innovative technologies to the market. This slide puts into context the magnitude of the impedance mismatch between our access to genomic versus proteomics information, and this was our key impetus for starting Seer. Access to the detailed and complex information in the proteome is key to understanding biology. Virtually, every function within a living organism occurs by the action of a protein or a group of proteins interacting with each other and working together in concert. With large-scale access to deep and a biased genomic information, we've now sequenced over 1 million genomes over 10 million exomes. And across these efforts, the research community has identified over 1 billion genetic variants. Yet, we still don't have functional context for reliance of this genomic information at the protein level. Now part of the reason for this gap is that biology is a dynamic, complex and purposeful matrix of interaction among the vast universe of biomolecules. This complexity exists because it's necessary for life as we know it. And science is showing us that complexity goes even deeper than we had previously imagined. What this has been known for decades, there is a growing literature that describes this complexity in more and more detail from papers showing that different genes expressed differently in different tissues, to those showing that multiple protein variants can be made from a single gene to those describing post-transitional modification of proteins with drastically different function, the evidence around this complexity continues to increase. As we move from the approximately 20,000 gene in the genome towards the proteome is estimated that we will encounter more than 1 million different protein variants. Now prior to Seer, the fundamental challenge with the available proteomic technologies was that we were not able to reach the depth, the breadth and scale to unravel this biological complexity, and Seer change that. Seer's approach allows for this complexity to be explored more efficiently than any other method that exists today and the corresponding market opportunity is large and growing. Our approach can be used to accelerate our understanding of biology and human health across both the proteomics market and the genomics market driving demand and expansion across both. We also believe that our technology will enable novel insights and methods that will lead to new companies and end markets that don't exist today similar to what we've seen happened over the last 1.5 decades in genomics. Before Seer, in a highly complex samples, such as plasma, it was not practical to consider doing unbiased proteomics studies with greater than tens of samples. Conventional approaches for deep unbiased proteomics involve cumbersome workflows that take a lot of equipment, manual labor and time and are not accessible to most labs and more scientists, they're fundamentally limited in their scalability. Seer's approach mitigated the need for this complexity and enabled access to the proteomics content at scale speeds stepped previously not possible. Our technology uses proprietary engineered nanoparticles, bringing together key attributes and unbiased approach, deep proteomics interrogation and rapid automated protocol, enabling researchers to take on studies of hundreds, thousands or even tens of thousands of samples. The approach is inherently scalable, and this is key. We are at a watershed moment in proteomics, the likes of which we saw in the mid-2000s in genomics, where access to novel content progressively became possible a larger and larger scale and new markets were created and expanded. Now let's unpack this a bit more. There are 2 approaches to proteomics today. With targeted approaches, a scientist typically uses an analyte specific reagent or ligand to screen for specific protein for which the ligand was designed. The maximum number of proteins that can be interrogated is inherently fixed and the information you're able to get is largely limited to the fluctuation in the quantity of that set of proteins for that panel of ligands that may be present in a given sample. With an unbiased approach, you're not limited by what you know today. And the more samples you study the more new content you discover. As the number of samples in your study increases, so does the number of protein variants you discover, rare protein variants, post-traditional modification, protein-protein interactions can all be interrogated. You do not need to have set up your assay specifically to capture them. As unbiased proteomics scales over the coming years, we're going to get closer and closer to understanding the vast complexity of the proteome. In this way, a huge amount of biological content can be discovered. Now how does an unbiased approach discover more content? Well, peptide level resolution is critical to identify protein variants, something only an unbiased approach can provide targeted approaches missed the vast majority of protein variants. We're progressively seeing publications that show this. For example, the recent work by [indiscernible] in Nature Communication, reported a few critical findings. Let me highlight them for you. First, protein variants induced by genetic variants are largely missed by targeted approaches and different targeted approaches can find different protein variants. That should not be surprising. It's because an average human protein is 472 amino acid long whereas an average epitope of a ligand use and targeted approach is 5 to 8 amino acid long. And as long as the changes in amino acids are within this short epitope, it can be seen by the ligand. The second point they make in the paper is that there is inconsistency in the data generated by the different targeted approaches using the same biological sample, really just a poor correlation. These differences can be explained by variation in the protein structure that are identified by one targeted method, but not the other. And third, the effects of genetic variants resulting in protein variants need to be explored by direct measurement of peptide sequence at the amino acid level. Now we've shown that this is the case in our own data. At right, we show a chromatogram with each of the peaks represent a peptide in the protein sequence. From peptide-based data, we can get to amino acid resolution. In this example, we're showing a peptide harboring a heterozygous variant for which the alternative is a rare proteomic variant. Let's take a look at this even one step further. Let's say we wanted to understand proteomic differences between healthy subjects and non-small cell lung cancer patients. Using data from our major communication paper that we have published in 2020, we identified several biologically important novel cancer biomarkers at the peptide level that would have been missed if we had only focused on overall protein expression. Let's take a look at bone morphogenic protein 1 or BMP1 as one example. BMP1 has been reported to have a dual rolling cancer, but the biology is not well understood. You can see on the top right figure that if you look at BMP1 at the protein level, ignoring the variance of BMP1, there is no apparent difference in the overall expression pattern of BMP1 between cancer and healthy subjects. This is confirmed by the volcano plot on the top left where BMP1 is highlighted and is close to the center line, meaning statistically unchanged. Based on this data alone, one would have concluded that there was nothing interesting and move on. Now if you look at the peptide level information, a more complex picture emerges. The peptide level volcano plot on the bottom left highlights individual BMP1 peptides, where you can see individual variants of BMP1 at play. In this way, we determined an opposite pattern of differential expression in a short versus long variance of BMP1 in cancer versus normal. It is possible that our findings may actually explain the dual role of BMP1 in cancer. BMP1 is just one of multiple examples that we found in this relatively small study of only 300 subjects that we focus really on just 1 disease. Now this discovery is enabled because unbiased approaches look at peptide sequences at the amino acid level and consequently, in the study of this size you actually may be looking at more than 20,000 different peptides that can shine light on which protein variants are present. In summary, our unbiased approach lets you see what is actually present in a sample. This is important since different protein groups may be present in different sample types due to changes in either cell or tissue-specific expression or simply biological variations among [indiscernible] samples. The key is to reproducibly sample what is present, and the Proteograph does that exceptionally well. We have used the Proteograph took a proteomic content across numerous sample type. Today, we have identified over 10,000 different protein groups in human and nearly 3,000 in mouse at a false discovery rate of 1% and rigorous criteria that defines how accurately you're identifying proteins that are present in a sample. Now as we look across sample types, spanning cell lysates to plasma to CSF different species, there is a change in the overall proteomic content. But importantly, the Proteograph samples this proteomic content in an unbiased way, so is able to detect what is actually present. As cohort size increases, increased sampling allows for identification of more protein groups and the discovery of more protein variants across a given population. The Proteograph is not constrained in the protein variants it can detect, so it's able to pick up variations across very low or very high allele frequency as well as identifying different isoforms, PTM and other variants. Now switching gears. Let's talk about our first year as a publicly traded company in 2021. We finished the year strong demonstrating tangible progress across all areas of our strategic plan. During the year, we completed the first 2 phases of our 3-phase commercialization approach demonstrating the power of our technology with over 25 abstracts at different conferences and adding a high single-digit number of Limited Release customers. We established industry-leading partnership with Thermo Fisher, Bruker and SCIEX. We demonstrated our commercial team and infrastructure and expanded the size of this group. We began efforts to expand our supply chain partnerships to reinforce our ability to support our growing customer base globally even during the unprecedented challenges of the COVID pandemic. And none of this would have been possible without the critical talent to drive progress across these important fronts, and we have more than doubled on organization in 2021. We're very proud of our progress and we're extremely well positioned to scale our business as we head into 2022. Now as a result of this progress, I am thrilled to announce that we have moved into broad commercial release, and we have already booked multiple orders. As we enter Broad Release, we have built a strong pipeline of demand for our Proteograph Product Suite, which is highly differentiated and easy to use. The Proteograph Product Suite includes the proprietary engineered nanoparticles, consumables and automation instrument and an analysis package called the Proteograph Analysis Suite. The entire solution readily fits within a researchers budget, both from a CapEx and a consumable perspective. And in a similar range as other large-scale omics study costs. In addition, the entire workflow sits upstream to a large installed base of detectors over 16,000 mass specs globally that are focused on the proteomic space. This makes a solution broadly accessible to most labs interested in adding unbiased proteomic data to their studies. We believe we are well positioned to become the definitive tools leader in the proteomics. We envision a future in which entire ecosystems and end markets could be created or expanded with customers using our Proteograph Product Suite to access unbiased, deep and rapid scalable proteomics across a myriad of applications. Over the next few slides, I'd like to show you how we're positioned to lead by first, demonstrating the power of our technology. Second, enabling our customers to drive unique insights and applications. Third, empowering studies at a scale previously not possible. And fourth, making unbiased the proteomic at scale accessible to more labs. First, demonstrating that our technology is uniquely enabling. We have 3 seminal publications in preprint or in press. In a paper accepted in PNAS, we studied nanoparticle protein interaction using machine learning approaches to characterize and model protein corona formation for more informed protein sampling. Next in the manuscript under review and available on bioarchive as a preprint, we further demonstrated how protein corona could be optimized to enable deeper and deeper protein access. And finally, in a preprint available on bioarchive, we explored the utility of combined unbiased proteomics with transcriptomic analysis and identified over 400 subject-specific protein variants, demonstrating the utility of the Proteograph Product Suite to support proteogenomics studies. The power of our technology was further exemplified by more than 25 abstracts at scientific conferences in 2021, many of which came from our early customers. These abstracts included studies from the Knight Cancer Institute, the Broad Institute, Sanford Burnham Prebys Medical Discovery Institute and Protein Metrics. They describe novel applications such as Tandem Mass Tag, also known as TMT, on the Proteograph identifying PTM such as glycoproteins and investigating noble biomarkers in cancer and complex disease. It is very gratifying for me to see the power of our technology in customers' hands. Let me highlight the abstract from Broad Institute as one example. They looked at the plasma proteome of patients undergoing therapeutic ablation and monitored the result in myocardial tissue damage and the time course as a model for myocardial infection or MI. The Proteograph dramatically reduced time and labor required for deep unbiased plasma proteomic versus the state-of-the-art depletion and fractionation method described in the group's 2017 Nature Protocol paper. The Proteograph allowed the researchers to detect an average of 2,200 proteins in each of 3 different plasma sampled across 5 time points with a total of approximately 4,000 proteins identified across all measurements. With the Proteograph, they were able to detect a validated biomarker for MI troponin and an earlier time point that they had identified multiple new biomarkers as well that needs to be investigated. We feel these early results are just the tip of the iceberg in demonstrating the unique capabilities of a Proteograph Product Suite and its ability to unlock new biology, new applications and new insights. Today, proteomics and genomics are largely distinct fields and rarely overlap in multi-omics size at scale. The missing link to truly enable proteogenomic has been large-scale access to proteomic content at the amino acid and peptide resolution at a similar scale to match the current large-scale access to genomic content at the nucleotide resolution. The Proteograph is well positioned to bridge the gap between proteomics and genomics to accelerate the impact of proteogenomics and to better connect genotype to phenotype. As we start 2022, we have multiple large-scale studies using our technology in early phases. Because these studies are empowering groundbreak in research, we expect the findings will enable important publications, respecting our customers' privacy we have agreed to share the nature of these studies without disclosing the names of the institutions or the investigators pursuing them. There is a prostate cancer study underway with approximately 1,000 serum samples to look at biomarkers that define clinical endpoints in prostate cancer. There is a multi-omics study underway across multiple centers to look at disease biomarkers across a cohort of over 2,000 samples. Aging is an important emerging field of multi-omics, and we have a recently signed project with over 1,500 samples gearing up. We look forward to keeping you abreast of the important progress and any resulting publications or presentations on these first-of-their-kind studies. Now we've also come a long way in making it easier for our customers to analyze deep unbiased proteomics data, and go from data to insight using the Proteograph Analysis Suite. When we developed the Proteograph Product Suite, we removed a key bottleneck in unbiased proteomics by enabling investigators to process sample in just 7 hours with an automated workflow in just about 30 minutes of hands-on time. Analysis still remain cumbersome. And historically, much of proteomics data analysis was very fragmented and pulling together advanced data analysis would take multiple experts sometimes weeks of time. Today, a single user who doesn't even need to be a bioinformatician can perform similar analysis in a matter of minutes with just a handful of clicks using the Proteograph Analysis Suite. The analogy I would make is basically comparing DOS to iOS in terms of user experience. This system is highly scalable, and we will continuously add additional capabilities to it, for example, including enabling proteogenomics and multi-omic data analysis. Now we're also making it easier for customers around the globe to access the power of the Proteograph Product Suite through unique partnership, such as the one we signed within Enlight Medical last year to expand into China. Today, we announced the launch of our Center of Excellence program, a program that partners Seer with premier select multi-omic service providers to offer unbiased deep proteomic services to customers around the world. We have strategically selected each COE based on geography and expertise, including Evotec in Europe, Soulbrain in Korea. And in North America, we have Biodesix, Discovery Life Science and Sanford Burnham Prebys Medical Discovery Institute. We're working with these partners to take the program live this quarter. I'm also pleased to announce today a key collaboration we've established to expand access to deep unbiased proteomics for genomics customers in order to enable proteogenomics and multi-omic studies at previously unimaginable scale. Together with one of our COEs, Discovery Life Sciences, a leading genomic services provider and SCIEX, one of our current commercial partners and a leading provider of mass spec platforms, we have created the first Proteogenomic Consortium. This is a multiyear collaboration and incorporates a 3-phased approach for discovery to set up, expand and offer deep unbiased proteomics capabilities to their existing genomic customers with an annual capacity of over 100,000 samples per year using the Proteograph Product Suite and the SCIEX XenoSoft-7600 platform. We expect this service to be launching in the second half of 2022 and look forward to updating you on the progress of this consortium. Looking ahead, as we look ahead, our vision goes beyond becoming the definitive leader in proteomics. Biology is a dynamic and complex, and we have a long way to go to unlock the meaning of this complexity and its impact on human health. We have an eye on the future and what might be needed long term. Connecting genotype to phenotype will require an unconstrained view of biology. Our long-term road map will focus on allowing this to happen at scale allowing us to become a definitive enabler of our understanding comprehensive molecular phenotypes. Our proprietary engineered nanoparticles are well positioned to expand this unconstrained interrogation of biology across different sample types, different species, different study types and even different biomolecules, each of the areas representing a significant market expansion opportunity. We have already exemplified many of these areas in our labs and are continuously exploring the breadth of the application that our engineered nanoparticles can enable. In this way, we're paving the way for a portfolio of products that will expand well beyond our first product in proteomics. We look forward to expanding the impact of our technology in 2022 and beyond. Now as we move forward in pursuit of this vision in 2022, we will continue to drive execution against our core strategies. We're very much at the onset of this journey. And while much work remains and we're excited and inspired by the opportunity that lays in front of us, I'm incredibly proud of our team for the progress we've made in such a short time. I'm excited and humbled to lead such an amazing team, and I'm inspired by the passion, the hard work, the dedication that has allowed us to commercialize such a transformative product. In summary, I believe we have the technology, the team and the strategy to bring the next phase in omics to labs around the globe. Thank you so much for your attention.

Great. Thank you, Omid. That was a great overview. And congrats on commencing broad commercial launch today. So can you tell us about how the order book is shaping up? Your press release noted multiple orders? So should we be thinking low single digit, mid-single digit or high single digit? And how confident do you feel about 35 to 40 instrument placements this year? And also talk about the centers of excellence program. Like how are these partners selected? And are they expected to complement your direct go-to-market? Or is that going to be the main driver of the order funnel?

David, can you take the question for me on the Broad Release pipeline?

Sure. Yes, Julia, as we did say, we do have multiple orders. We're not in a position to give guidance today on this point. But I will say we do see a really strong funnel and we've had great interest across the board from different markets and customer types. So we will be providing guidance on our year-end call in late February. So -- but we do, as I say, see a lot of strong interest and are having a lot of great discussions.

Got it. And then you announced the proteogenomics consortium with SCIEX and Discovery Life Sciences to deliver that 100,000 sample annual capacity. Can you tell us about the time lines? How many instruments are involved? And is that 100,000 annual capacity to be achieved in the near term? And how should we think about the demand pipeline for such population scale, proteomics -- proteogenomic studies?

Yes. Julia, let me take a bit to that. I might ask David to add, but let me take a bit to that first. So number one, the community, the genomic community has always been very interested in annotating the genomic content in terms of their functional relevance. Now the missing link has always been the access to large-scale proteomic content in an unbiased way to be able to annotate genomic content. And so what they've done is they've looked at surrogates of it, transcriptomes, et cetera, methylation to kind of get there. So there is definitely a ton of pent-up demand in terms of being able to tie proteomic information at the amino acid level to genomic information at the nucleotide level. And of course, the ultimate demonstration of that is engagement of a site like Discovery Life Sciences, a real preeminent site in the genomics space with enormously large number of customers being able to add access to proteomics for their customers in an unbiased way. The way the partnership works is that a combination of SCIEX and Seer provide the requisite instrumentation for DLS to build infrastructure, which will happen in the second half of 2022 and then to gradually ramp up to a capacity of 100,000 samples of unbiased deep proteomics per year, which, frankly, when you look at where we started from before Seer is just totally remarkable. Before Seer, the largest deep plasma proteomic studies that were published were in the tens of samples. In fact, to be exact the largest that I'm aware of, and there may be others, but the largest that I'm aware of, was 48 samples. So we've already gone from studies that were in the neighborhood of 48 samples, I shared with you studies that are already starting in 1,000 to 2,000 plus samples. And so add these customers up, and I think the 100,000 sample a year capacity, it may just be tip of the iceberg of a starting point with where the community needs to go to really kind of empower what the need is in terms of the information. And this is particularly relevant because the proteomic content is also dynamic, which means that you may actually be testing similar subjects at the proteomic level repeatedly, which is really not something that you do in a genomic level. David, do you want to add anything to what I just said?

Yes. The only thing I would add, Julia, just to get to your question on, is this the primary driver or is this an addition to, look, I think in general, what we've seen is we're trying to increase accessibility, right? So there -- this is a new disruptive technology platform, and there are certain customers who would prefer to access the technology through a third party rather than bringing it in-house. And I think not only this proteogenomics consortium, but also the centers of excellence is a way to simply expand the ecosystem to allow experts who know how to do multi-omics who provide excellent customer service and can also quite frankly, help us with product and workflow feedback to expand the ecosystem. So we see this as kind of additive to our own direct selling efforts.

Got you. And you currently have high single-digit number of early access customers. Can you maybe break out the mix by academic versus pharma versus clinical? And is that mix in line with your original expectations? And how do you see that mix changing going forward with the broader commercial launch now? Is one group of customers more likely to take up than others?

Yes. Let me start off and maybe I'll hand it to David. So if I look at the mix, Julia, is exactly as we had predicted. We have biopharma. We have liquid biopsy diagnostic companies. We've life tech drug developers. You have academics. You also have CROs. So that mix is exactly the representation that we had expected heading into the Limited Release phase. Now as we shift from Limited Release into the broad commercial launch, I think that a lion's share of what you can imagine, just because of the length of the sales cycle is going to be on the commercial side. Academic customers tend to have a longer sell cycle because they have to -- they're constrained by grant funding, et cetera. So I think as we shift from the Limited Release to the Broad Release, the mix will remain very much similar. And I think you can assume that a lion's share of our customers coming out of the gate initially will be on the commercial front. And we'll continue to expand, build our relationship with the academic community as they also ramp up with their funding going forward. David, add anything to that?

No, I think you covered it very well.

Great. Can you maybe talk about the China order book specifically? I mean your partnership with Enlight Medical suggests strong interest from that region. So how much visibility do you have into the China order funnel? And how much share of revenue do you expect to come from China in the near term?

David, please?

So we are seeing good demand in China. And so that's been a very strong interest from that market. The issue with China right now, Julia, is actually trying to get folks into China. With COVID, it's been extremely hard to get individuals on the ground there for training and installation, that type of thing. We do have Enlight there. But again, just to highlight that, that has been a challenge for us. But again, we do see that, that market growing very strongly being a very good market for us. I would say in 2022, we do have a fairly modest amount of revenue from China. So we're not -- again, we're trying to take it slowly and do it correctly, kind of like we did here in the U.S. is kind of a phased approach, but we are seeing very good interest there. And so we do have high hopes for that market. It's just we're going to be phased and methodical in our approach to how we roll out there.

Got it. I know you talked about the significance of the recent OHSU and Broad data releases at ASMS. What other publications can we look forward to in 2022?

Well, Julia, I hinted that we got a paper accepted in PNAS. That was a nice Christmas gift that came to us right after Christmas. I expect that paper will come out. Most of the journals, by the way, have a backlog because they're all focused on publishing COVID-related content appropriately so in the midst of pandemic. But I expect that PNAS paper, which is not accepted to come out. In addition, there's another paper of Seer that's under review though we have made that available on bioarchive. And then the proteogenomic paper, we have made available on the bioarchive. So those 3 papers, I think you can expect to -- well, it's 2 of them you can read today and one as soon as it's out the PNAS, one as soon as over the coming weeks. I also expect that you're going to see the same level of activity, both from us and also from our customers progressively at conferences. So abstracts coming from us and from our customers in the coming months throughout the year. And then finally, my expectation is that you're going to also begin to see publications from customers, Julia. So to that, I have a lower degree of visibility. But of course, the visibility I have is to the level of the depth of the studies that they're doing and the types of data that is being generated and the likelihood for you to be seeing those in conferences. So I'm hoping for a robust year of demonstration of the -- really the power of the platform, both in terms of papers and abstracts over the coming months.

Got it. Sounds great. And lastly, can you talk about your near-term R&D priorities? What kind of new features or enhancements are customers demanding? And we know the Proteograph can currently do about 2,000 plex while some of your competitors are at higher plex. So what are your plans to increase the MarTech plexing? And is there a road map or time line you can share?

Sure. So let's break it up into 2 parts, if you would. So the first is what does our product road map look like if you would? So I think our first product has a massive amount of value add and a ton of runway and a lot of interest among the customers in what it delivers to them and the value it adds. We have been very active. And in fact, one of the key purposes of the Limited Release was for us to be very engaged with our customers. And the marketing team having done a voice of customer analysis, we're hearing where the market needs are. So for example, the market is asking for, can we have lower amount, if you would, or lower number of mass spec injection to increase the throughput of the mass spec, for example. Or can we have lower sample volume to start off so that proteomic studies can be done not just in human but in model animals that typically their blood volume is lower than human. And so you need a lower sample volume to be able to do, for example, a mouse study or rat study. And we're also being asked that, can you integrate or making it easier to integrate proteomic content into genomic and multi-omic content because as the level of information grows, integration is going to be key. So I think if you can imagine, we are -- we have an R&D plan that essentially mirrors those asks from customers, addressing those areas that were being asked to be a priority to them in the next iteration of the product or subsequent products that add to our existing product. So that's the first half of your question. The second half was with regards to 2,000 plex versus 7,000 plex or 3,000 plex, let me challenge that approach to the question, if you don't mind. Number one, when I think of plex, I don't think Seer is a 2,000-plex platform because if you actually scientifically look at it, we're technically 5-plex because our product is a 5 nanoparticle product, not 2,000 ligands. It interrogates 2,000 protein in plasma, but it does a different number of proteins, some overlapping and some unique in CSF and synovial fluid in other tissue. In fact, the same panel of 5 particles is what has now identified over 12,500 protein groups and growing, and it does that at a 1% false discovery rate and it does that in a highly reproducible way. Now that said, we can constantly increase the depth of protein coverage more and more to get to a larger and larger number of proteins. When you also think about a 7,000-plex proteomic platform, that 7,000 protein plex platform does not identify 7,000 proteins in plasma. It's 7,000 ligands across all of the biological content that it can identify. It identifies a subset of them in plasma, a different subset, for example, in intracellular compartment, a different subsegment of it in extracellular compartment, some in tissue, et cetera. And so in any one sample type, the number of protein IDs is different. And of course, we don't have a sense of what the false discovery rate is to understand what is actually accurately being covered. So if I think of Seer, I look at it as it identifies what is present, and it does that in a highly reproducible way with a very defined FDR and a very tight coefficient of variant. Now if you look at variant across the studies, those variances can come as a biological differences. So if I have cancer, I might look different than to the next person that has cancer. My cancer may be a bit more advanced. There's going to be a bit more advanced, might be slightly different. Those are biological variances. And again, the Proteograph identifies exactly what is there, and it does that in a highly reproducible way.

Great. We're a little over time. So let's leave it at that. Thank you very much, Omid and David, and good luck with the rest of the conference.

Thank you so much, Julia, and thank you, Tycho.

Yes. Thank you very much.