Good afternoon, everybody. Hopefully, you've all had an opportunity to watch the technology update from Clive at London Calling event last week. You were all sent links to that. Joining me today is Rosemary Sinclair Dokos. Rosemary Dokos, as we say. Sinclair. I'm reading this script. SVP Product and Programme Management, who will give a brief recap of the key tech updates, and then we're gonna dive straight into Q&A, if that's okay with you guys.
Thank you, Gordon. Hi, everyone. Good afternoon. Last week we had our technology update at London Calling. It's something we do every year. It's a really great opportunity for us to engage with our community of users and provide a roadmap as to where Oxford Nanopore is going. We've had a lot of interest over the past few months around and many years around our accuracy in Simplex and Duplex. We kicked the show off with Duplex. We've updated our users on how our Duplex rates have continued to improve both internally, but also in the hands of a very select number of users who had developer versions of the product.
We showed internal improvements such that you can get 90 gigabases of human data off a single PromethION flow cell that is composed of 50 gigs of Duplex and 40 gigs of unpaired Simplex. That is over 90 gigs of data which is Q20 and above, with a lot of it at Q30. Those are long native human reads. We showed our timeline and our roadmap to get all of this implemented into MinKNOW, it's really easy to run. We also showed what the technology is actually capable of when it's fed ultra-clean PCR DNA, which is over 100 gigabases of Duplex data from a single flow cell, with a further 40 gigs of unpaired Simplex.
We're making this available to users now. Users today can log in and register interest on our website. From there, they are spoken to by members of our technical team, just to make sure that they have everything in place that they need in their labs. Then they will be given access to this technology. We reminded all of our users about our Simplex capabilities. It's really important to remember that Duplex has been phenomenal for really difficult problems like telomere-to-telomere assemblies. However, when it comes to whole genome sequencing, detecting variants, detecting methylation, Simplex continues to be a market-leading data type. Our outputs on Simplex are routinely over 100 gigabases in our customers' hands.
Our SNP and our SV and our methylation detection are market leading, and we've made great strengths and great improvements in our indel performance, particularly in coding regions, where we are getting over 99% indel performance in coding regions, across the whole genome. We have shown a roadmap on how we're gonna get Simplex from one genome per flow cell to two, three, and then four, as we continue to have immense headroom in our platform. We've also spoken about our ability to now provide basecalling in keep up mode on the A100. These are improvements that are going to go out shortly, such that customers, when they start a PromethION run, their analysis, not just basecalling, but also methylation detection and alignment can all happen on device.
On modifications, we've spoken about our ability to do the industry standard 5mC, our market-leading ability to do 5mC detection. Going into research now is the ability into research, but this is actually research with our customer base, is the ability to detect novel modifications in all contexts. RNA was a really big star of the show, not just in Clive's talk, but we also had an RNA talk just before Clive and many other during the conference.
Our direct RNA chemistry is getting a significant output boost as well as an accuracy boost, and we are making available tools, early tools for people to start interrogating the modifications in RNA. Around our platform, we've spoken about the availability, the early access of our P2 Solo device and our developer release now of the P2 Integrated with Compute, and how all of this continues to drive utilization and drive adoption of our platform. We have a lot of new platforms in the pipeline, so we are, the MinION is due its first facelift in 10 years.
That will be coming later this year, early next year, and its incorporation with iPad brings together the most distributed sequencer on the planet with the easiest to use user interface from Apple. We also have a roadmap of new ASICs, including the SmidgION ASIC, which will lead to a whole host of new and exciting devices, which are incredibly low power, and potentially, low cost consumable there too. That summarizes all of the key points that we, that we made at London Calling at the tech talk. Now we'd be happy to take some questions.
Ladies and gentlemen, if you would like to ask a question or make a contribution on today's call, please press star one now on your telephone keypad. To withdraw your question, please press star two. The first question come from the line of James Gordon calling from JP Morgan. Please go ahead.
Hello, James Gordon, JP Morgan. Thanks for taking the questions. Our first question was about things you can sequence. There was quite a lot of talk about Duplex sequencing and RNA sequencing in the presentation, but how widely do you think each will be needed? Do you think in future you're gonna get a very large proportion of people using Nanopore using Duplex? Similar sort of question for RNA sequencing, is that gonna be very big or is that quite niche? I think there was a comment before that maybe 5% of the sequencing market will be RNA. Could that be conservative and actually it could be a much bigger commercial opportunity? A second question was just accuracy, even further improvements, but are we now there on accuracy or are further improvements actually needed, and are you gonna see a lot more further improvement?
Thank you. On the Duplex RNA Simplex, I think Duplex, it's been of immense interest. The telomere-to-telomere work that happened sort of last year and bleeding into this year has really sparked an immense amount of imagination. Duplex for telomere-to-telomere work is generating a lot of attention because you can apply it to humans, but also, there's a lot of efforts now in doing perfect assemblies of plants and animals. Whilst it will drive a lot of scientific discovery, when it comes to how much sequencing happens on the planet today, which is de novo assembly, which is what Duplex is really useful for, versus how much sequencing happens on the planet today for just run-of-the-mill variant detection.
The vast majority is run-of-the-mill variant detection, which is why Simplex is such an important technology to Oxford Nanopore. Also we wanna feed not just run-of-the-mill capabilities, we also wanna provide a tool that can also lead science in new directions. Both are very, very important. RNA is a very, again, a very interesting market because RNA changes, it's a dynamic molecule, whereas DNA, of course, modifications change slightly, but it's a more static genomic fingerprint. RNA is something that changes all the time.
Whilst again it is today perhaps not as big a market as DNA sequencing, in the future, in years to come, RNA can be, can become much, much bigger than what it is today, especially given, interesting tools and interesting capabilities that we'll be working with our community to develop over the coming years. The other question was accuracy, further improvements, are we there yet? I think it's really important as a company that we never say we're there.
There's always more to come from an innovation standpoint, that's what will keep us ahead of the competition. Where we are today is incredibly competitive performance versus the rest of the market and the ability to answer customers' questions. We are there in terms of what our customers need today, but we always need to be ready for what they want tomorrow, and that's what the innovation team work hard to deliver.
Thank you. The next question comes from the line of Odysseas Manesiotis calling from Berenberg. Please go ahead.
Hi Gordon. Hi Rosemary. Thanks for taking my questions. First of all, I understand you didn't talk about RNA, the sensing and protein sequencing. I'd say all the other avenues were quite clear, but could you give us a bit more color on your progress and in particular on the RT side? Then, I'll ask another one. Thank you.
Sure. I think the trouble with London Calling is that we could probably fill three days just of Oxford Nanopore progress, but that wouldn't be really very interesting for all of our customers. We, we do have to be a little bit brutal into what makes it into our technology talk. RT is still in with our research teams. Our accuracies continue to improve, and we continue to get, we continue to develop new capabilities, such as things like size selection that Clive has spoken about before. The team continue to do well. It isn't something that is imminently available to our users, which is why we chose to, we chose to not talk about it. Protein sequencing, again, it's on the same sphere there.
We've made a lot of progress internally, and the teams are really, really excited about where we are. We really felt that this conference we should be focused on what users are gonna be using in the near term as opposed to things which are more sort of medium term in terms of delivery. Again we had to make a very difficult choice in terms of focusing on Duplex and Duplex RNA, Simplex, and we will come back and talk to users about our team protein sequencing as they near product readiness.
That's all clear. Thanks. Could you take us through a bit more detail on the early access program for CyclomicsSeq? What will it look like, and what's the roadmap to getting it cleared for clinical use?
The early access. One thing that is important to note about Oxford Nanopore is that we've got a few stages that our products go through. We have developer releases. Early access for us is already a commercial, a commercialization path in that customers are ordering product and receiving it. The early access Cyclomics, that is a pure play, that is a tool. It is not the, it isn't the kit with a specific clinical use case. What we're doing is making that capability available to our whole community through an early access program so that they can develop applications. Does that? I hope that makes sense. What we're doing is launching a kit so that the research community can develop applications and then come back to us with things that they find are incredibly valuable with that type of chemistry.
That's all very clear. Thank you, Rosemary. I'll jump back into the queue.
Thank you.
The next question comes from the line of, Charles Weston calling from RBC. Please go ahead.
Hello. Thanks for taking the question. The first topic, is on population sequencing programs, please. Appreciate this is not part of a tech update, but there were a couple of presentations from, one from EGP, one from Genomics England. I just wondered if you can give us a bit of color on that. Specifically, I suppose on, for EGP, they said they were looking to sequence 1 million people by 2025, but it was only 83,000 last month. They talked about PromethION and also, a sense of how they're looking to accelerate Nanopore use there. Genomics England, they also talked about sort of 1 test for cancer and piloting with Oxford Nanopore. On that as well would be helpful, please.
Okay. The EGP program is GBP 1 million by 2025, but it's not limited to just long-read sequencing. It's also short-read as well, and it's public domain information that they are working with BGI and Illumina as well. It's a big program. The 83,000 update was Nanopore sequenced genomes. And what the mix looks like and what they end up doing and how all of that plays out, we're not party to that level of information other than we have our contract, which runs until next year, and we have a set number of genomes that we're doing in that contract.
On Genomics England, the so I think test is a bit too forward-looking right now. It's, they're looking with the GLH hubs to roll out and debug what this screening pilot looks like. It'll be on the basis of that there will be a clearer timeline, if you like, and what that looks like. This is really an exploratory phase right now. Exciting, but exploratory.
Thank you. The second topic please on the P2s. I think there was a serial number over 600 that was shipped. Does that mean you shipped more than 600 P2s and it sounded like from Clive, the first ones needed replacing. How many was that? Are these customers new to Nanopore, or are they upgrading existing capacity?
Yeah. On that, we make lots of P2s, some for customers, some for internal, some for production. You can't take the serial number in field for as gospel. We have had an enormous amount of interest in the P2. It's, you know, we're just very, very excited with all of the new applications that come out of that customer group every day. In terms of are they new to Nanopore, it's a really healthy mix. Some are brand new, some have had MinIONs, some have had GridIONs. It's a healthy mix of new users, but also existing users who are going from MinION platform and adding the PromethION platform to that capability.
Great. Thank you.
Thanks.
The next question comes from the line of Kyle Mikson calling from Canaccord. Please go ahead.
Great. Hey, guys. Thanks for taking the questions and thanks for that update last week. Great. Lots of detail there. My first question was gonna be about just kind of like a high-level view on the market here and kind of going forward. You talked about this PromethION really, like, you know, ultra-high throughput, ultra, high output for PromethION. I think 4 genomes for flow cell is achievable, maybe like in the near term. I guess, you know, just curious, who really needs 20,000 long-read genomes per year today?
Like, what's the current demand for something like that and what's a good, like, per genome price point in your view? You know, will this only be like a limited update from you in the near term, or could it expand the market possibly soon and maybe kind of make sense commercially? It's kind of a question about capacity for long-read whole genomes right now and kind of how you kind of aim to serve that market segment, I guess. Thanks.
I think the demand for population scale genomics is growing. The more we sequence at population scale, the more we need to sequence, the less we understand, and the variation is proving to be dramatic from region to region and population to population. The key thing to take away from last week's conference was that native DNA with all its modifications, long reads with structural variation and delivering structural variation and copy number variation are game-changing across human health. That was particularly highlighted in our clinical translational session on Friday afternoon. We believe that there isn't a dichotomy in the market between long and short reads. That was something that was created as an illusion when Illumina tried to buy PacBio.
If you started today with a blank canvas and said, "You can have very short reads where we've bleached out all the interesting things like methylation, and we cannot get copy number, and we cannot get structural variation, or you could have native DNA at any read length you want," the market will naturally migrate. There is already demand, and the EGP program is a flagship for long reads, native Nanopore long reads and population scale programs. We are in multiple conversations across the globe and in the conversations.
The question we don't know how to answer yet is how you see that transition as we have ever-increasing data sets that show native longer DNA is very important. That is really just a function of time and effort, and one driver in population scale genomics is price. The BGI and the Illumina of this world are in that sort of $100 to $300-$400 range. We believe we have to be in that range, we will be offering a superior data set than anybody else on the planet at scale.
I think there's one more point to add there, which is our platforms are really flexible, so the fact that we can do 20,000 if you run a P48 doesn't mean you have to. Certainly the pricing that we've put in place with customers means it's incredibly flexible, and it flexes to their demands.
I think that's a
Great. Thanks so much, guys, for that.
Yeah. That's a really important point. These platforms are pay-as-you-go. You don't have to maximally load these machines either.
Yeah. Sorry, Gordon, for cutting you off. No, that was great detail there. Maybe Rosemary, just going to you for a quick question, for a quick follow-up there. I think that there was a question earlier about RNA sequencing. I think what's important here is like, for this update is the direct RNA sequencing aspect. You know, it sounds like there's a lot of interest there for the developer access, for the flow cells and the kits. I guess, like, how much of a net need is native sequencing of the RNA molecule? Maybe could you just walk through how much of a competitive advantage direct RNA sequencing is in kind of like the overall short read and long read field? Thanks.
I think the direct RNA is, you know, is the next sort of native long read for DNA, if you like. People y ou know, we've had direct RNA capabilities for a few years, but their outputs have been low. It's, you know, price per gig or price per molecule has been high. It's created a whole new market of people who really need that in-depth information. You can only do that with in-depth information. As the outputs increase, then it becomes an ever more competitive. Not only does it do the more specialist experiments and the more high value experiments, but it also starts to do some more run-of-the-mill everyday RNA experiments.
Yeah, it's gonna be a very growing market. In terms of benefits, we offer both. We offer direct RNA, but you can also sequence cDNA. Customers today can do single-cell, bulk, spatial, using our cDNA technologies. Those are long reads already. They capture isoforms, they capture splice variants. Very much in the same way as DNA, we used to just talk about SNPs and indels off Illumina machines. It's really the same thing with cDNA, with RNA-Seq, if you like.
People just talk about expression analysis. That's the only thing they really talk about. With the introduction of long read sequencing into that space, we've started to find novel structures. Direct RNA takes that a step further, and it delivers your isoforms. It delivers all your splicing information, but it also delivers modification. Now that is gonna be a journey because there are many, many RNA modifications, many of them who've never been looked at or studied before. It's certainly a very high value molecule to sequence, and the RNA community is incredibly excited about it.
Perfect. Thanks, Rosemary. Thanks, Gordon.
Next question comes from the line of David Westenberg calling from Piper Sandler. Please go ahead.
All right. Thank you for taking the question. First I just wanna talk a little bit about Duplex and kind of the cost associated with Duplex versus Simplex. I don't know if you know, I know, I think prices kind of vary on a kind of a customer usage kind of basis. I don't know if there's any ranges you want to give in term or maybe like comparative Duplex versus Simplex, or if there's, you know, a nominal amount for how much that costs to, say, an S1 versus, like, an S3 customer.
What the Simplex versus Duplex, you know, comparison cost is. Really, like, I think this is pretty exciting on kind of the accuracy we can get with Duplex. You know, this really does compare on a SNP basis to some of the companies in short reads and, you know, even any kind of long, like, long read competitor. Just trying to really get a sense for, you know, how low you can get that cost and what, you know, what the market can get and what you can deliver to the market in terms of that cost.
Again, one thing really important to note is our SNP detection on Simplex is as good or better than Illumina across a whole genome. It's always really important. I know everyone always gets really excited with the highest of high accuracies. It's really important to remember that Simplex is over 99% accurate and can delve into regions of the genomes that we haven't been able to see up until now. In terms of its, in terms of Simplex's competitiveness, we can get it down to sort of between $3 and $5 per gigabase. Those are the kind of outputs when our customers are getting 150, 180.
We've got all of those tables up on our website in terms of price per gig. It's incredibly competitive. Duplex is a more expensive data type, but again, it's really important to remember that the if you think of a telomere-to-telomere reference genome, I think they typically cost these days somewhere in the region of $20,000 to do a telomere-to-telomere by the time you add all the different methodologies into it. If we're now saying that you could potentially be doing that with two, maybe three PromethION flow cells, then you've taken something that used to cost $20,000 down to under $3. There's immense benefit in that.
Again, I think it's, we need to be careful not to do a sort of a price per gig like for like comparison for Duplex because it's just enabling science that was not possible before. Whereas Simplex is much more the run-of-the-mill price. You know, for the customers who are price sensitive in terms of price per gig, Simplex is incredibly competitive. For the customers who wanna do telomere-to-telomere assemblies, the price just got an enormously amount more affordable for them.
Okay. Thank you. Then, I'm gonna go back to the interest in the terms of the P2. Maybe this is a pretty short, direct kind of question. Any sense for how many of those placements are in a production capacity, i.e., the central lab of maybe the core lab, kind of where you would traditionally use those big machines? I'm just trying to get a sense for, are you also, you know, getting interest from there from a really small platform, and then maybe it has a chance to decentralize, go from centralized to decentralized, kind of different from your normal business model? I mean, anyway, just trying to get a sense for that.
Yeah, again, it's gone to a very broad brush of users. A lot of decentralized, a lot of individual labs are picking up the P2 and doing all sorts of wonderful experiments. But in the core labs, the interesting part about having the P2 platform in cores is that they can run their larger instruments, their P24s and their P48s. They can run those instruments effectively in production mode if they've got large projects going on.
If they want to try a new chemistry or if they wanna try a new improvement, they can deploy the P2s alongside them because it is effectively the same platform. That's why we've hit such a big nerve with the P2 launch in that it serves not just customers who've never sequenced before or never used a PromethION before, but it's also going into PromethION labs who want it to run alongside their production units, so that they can do experimentation on the smaller device and then scale up on a larger one.
Got it. Ian, this is a question outside of this kind of the tech update, so, you know, feel free to not answer it if it's if I'm answering it or asking a question in at an inappropriate time. You know, there's been a lot of noise in terms of the China market just in terms of competitive, not just in, you know, there is a nanopore company there and there is short read companies there. Just any kind of sense for what's going on there overall. Again, I know that's not a tech update question, so maybe I'm asking at the wrong time.
Just, the MinION has been adopted by the Chinese Centers for Disease Control as its chosen post-surveillance pandemic network of choice platform. We continue to make good progress in China. Clearly, the economic landscape has changed over the last couple of years. BGI have been there for a very long time, but there is a very sophisticated market in China, and they understand the benefits of native long reads that Nanopore offers.
Thank you.
The next question comes from the line of Tejas Savant, calling from Morgan Stanley.
Hello, this is Yuko on the call for Tejas. Thank you for taking our questions. Regarding Duplex sequencing, how much of a hurdle is offering this technology for longer reads, such as ultra long reads? Could you talk about the challenges there, and is that something that could potentially occur in the medium term or is it more of a longer-term opportunity?
To do Duplex, you need to put an adapter on both ends of the DNA, and the chemistry that we have today for ultra-long doesn't need that. I think the actual better question is it gonna be necessary to do ultra-long Duplex? I think a lot of the telomere-to-telomere consortia groups have been managing to show incredibly highly contiguous assemblies, which is what the long reads give you with the Duplex as it is today. The Duplex as it stands is generating reads well into the 200 and 300 kilobases, which is more than, you know, it's very long. It's sufficient to sort of build to build the scaffolds that you need for assembly.
I don't think it's gonna create a blocker, if you like, in terms of a capability that doesn't exist today. There will be molecular biology ways in which we can address it. However, for now, customers are getting long reads with Duplex, and they are building the scaffolds and assembling the genomes that they need with those long reads. They're accentuating that with ultra-long if required. Again, remember that the ultra-long reads are 99.3% accurate. Again, they are incredibly high accuracy. They are Q20+ on ultra-long.
Got it. Thank you for that. Then on adaptive sampling, what applications are you seeing the greatest traction? Is this something that could be useful for a needle in a haystack type of applications as well?
adaptive sampling is really helpful. If you, if you wanna look at a genome, if you've got a whole genome, but you only wanna look at a little part of it, adaptive sampling is incredibly helpful 'cause it brings the cost down or the time to result down for that region of the genome. Needle in a haystack is when you're looking for one copy in 100 or 100,000. Yes, it can help with that, but there are other methods that do a great job of that too, such as sort of amplification or Cas9.
Yes, it can be useful for needle in a haystack, applications, but right now most people are using it for targeting regions of a genome and getting to an answer faster, or cheaper by only sequencing the parts that they're interested on and therefore multiplexing multiple genomes onto a flow cell.
Thank you.
The next question comes from the line of Veronika Dubajova, calling from Citi. Please go ahead.
Excellent. Thank you guys, and thank you for the update. Super helpful. Just wanted to follow up on just the P2 interest and I guess this might be a difficult question to answer or you might not want to. Relative to your expectations, here is, Gordon, where you'd put that degree of interest in P2 and maybe anything that surprised you either positively or negatively in the type of customers who are interested in it. I have a follow on after that, but.
When we canvassed pre having P2, we were somewhat surprised that there was broad, strong interest across all, you know, people who were P48 users, P24 GridION users, even MinION users. It's interesting because it's, it kind of cuts one way in that it's a $10,000, the P2 Solo. It's a low-cost entry point. It's not surprising maybe that a MinION user wants it. At the same time, it's quite a powerful beast.
It's not unsurprising that the, you know, having the sort of quick two-channel thing next to your bigger P24, P48 or even next to your GridION platform would be of interest. The P2 Solo can be plugged into a GridION to leverage the compute as well. It's across the board right now. As we get more granularity, we will come back and talk about the use cases and where we're seeing traction, but it's way too early to really kind of figure out where it will stick most.
Helpful. My second is on the sort of Kit 14 chemistry standardization. Again, Gordon, stop me if I'm asking the wrong question in the wrong forum. Just from a practicality perspective, what do you think that means for your operations? Is this something that we should.
You bleeped in and out there, but I think I got Kit 14 is being rolled out, but we've been very careful and controlled. We actually launched it a year ago. The reason we haven't rolled it out widely for everybody is because there are a lot of customers doing sequencing who are very happy with R9.4.1. Moving to R10 series has been a controlled transition, and we will continue that transition, but ultimately we expect everybody eventually to transition over. That may require some validations on some of our customers who are very happy with the data they're getting on the R9 series. I think I would remind you all that accuracy is not binary.
It's a continuum. You can have, you know, a very low accuracy call, but time to result is critical. You wanna know if it's MRSA with a particular mutation. You don't need ridiculously insane accuracy for that. Finding very rare mutations in very rare disease cohorts, yes, you do. There's a continuum of accuracies and we're able to continue the nine series, but ultimately it will be replaced with the ten series.
A practical perspective, Gordon, for you, any implications, cost perspective, ease of supply, logistics, anything like that that we should bear in mind?
No, there's, you know, in terms of everything that goes into a Kit 14 is incredibly similar.
Yeah.
Bar one enzyme to a, to the other kits before. There's no impact on logistics or supply chain.
I mean, yeah. As I was reminding some people at the conference, They're big, they have a big impact on performance, but these are subatomic mutations where the number of changes is very low. they're minor changes to the actual chemistry and how it interacts with the chip and the operating systems. They just have profound leaps in terms of the accuracy because our signal-to-noise increases and our algorithms get easier to make to deconvolute. that will be the game that we'll continue to play. essentially, it's the same Nanopore with small point, single point subatomic resolution mutations that have that dramatic impact because we are operating at the single molecule level and that's what you can gain in benefits.
Very clear. Thank you guys so much.
The next question comes from the line of Shubhangi Gupta calling from HSBC. Please go ahead.
Hi. Thanks for taking my question. I have a couple. First, RNA sequencing has been there for a while. What are the advantages that this technology has over the other existing technology, especially, like RNA sequencing by other long-read sequencing technologies? Second, in its current form, can this be used in like clinical applications? What are the clinical application? Like, can it be used for drug discovery or speeding up clinical trials like, DNA sequencing is being used now? Thank you.
Thanks. There's one thing that's really, really important, which is that there is no other company in the market that does direct RNA sequencing. Everyone else does cDNA. They take an RNA molecule, they convert that molecule to DNA, and then they sequence the DNA. Be it, Illumina or other or other long read technologies, they are all reading cDNA. We are the only company that is reading direct RNA, and therefore the only company that is able to detect the modifications in the RNA as well. Regarding the, regarding the applications, there's an enormous amount of... Again, there's a lot of research to do on this. Direct RNA hasn't been a molecule that's been easy to look at from before.
There may very well be, it's very highly likely that there are gonna be unique and interesting and very important biomarkers in things like clinical applications, but it'll take us a while to get there. An application that we've actually encountered much more quickly is actually in the production of mRNA vaccines. That is a very quick that has been an application that's come to bear very quickly because they need to QC the vaccines that they are producing. mRNA vaccine technology is very exciting. It's, you know, the promise towards personalized medicine. There are gonna be hundreds and hundreds of different types of vaccines, and they all need to be QC'd.
There's only one technology in the world that can read an RNA molecule directly, and that's Oxford Nanopore. Everything else, you would have to read a copy of the RNA, and as soon as you copy the RNA, you lose the modifications. Some of the modifications that go into mRNA vaccines are critical towards its function. I think on the clinical side, there will be lots of very interesting applications that we will hear about potentially next year at London Calling as all of our users deploy this in all of their studies. If you want an example of something that's been very obvious and very quick for us to recognize an opportunity, it's something like direct RNA vaccines.
Follow-up question. This technology reads the complete RNA molecules. In the earlier technologies, there were some, I think, effect on library preparation when reverse transcriptions happened for cDNA. Is it the same in this technology or this has been like overcome over?
Yeah. You don't need to do reverse transcription with Nanopore in order to read the RNA. The reverse transcription, it turns the RNA, the single strand, it turns it into a double strand, and that's what other technologies use. We can do that too. We can do an RT, and we can sequence the result, the cDNA that results from that. We can also read the direct RNA without need for RT. That's what makes it different. I hope that's answered your question.
Yes, it does. Thank you.
You're welcome.
The next question comes from the line of Charles Weston again from RBC. Please go ahead.
Thanks for taking the follow-up. I enjoyed, Rosemary, your slightly more nuanced, potentially lower cost for the ASICs. A bit more nuanced perhaps than Clive on the technology update. When do these ASICs actually get integrated into devices that might actually go into the sort of early users? You talked about cost, but also power consumption and size. What do you envisage this enabling? You know, we talked about the SmidgION, it's called the SmidgION ASIC, but what else could this enable?
Sure. We have prototype units. We think there'll be sort of prototypes around and hopefully with very early customers during next year. Low power, it is something that's actually really important. If you think, you know, for example right now, when we plug a MinION into an iPad, it'll last an hour. If you have a much lower power ASIC, it will last maybe ten. And that is a really big deal as we want to enable more and more decentralized sequencing.
The actual, the sort of most exciting thing with the SmidgION ASIC, we've had Flongle for a number of years as a product, and it has a really strong community that's built around it. The request we often get from Flongle users is they want that type of data, but they want it quickly. A Flongle has 100 channels, and so it takes four times longer than a MinION to get to the answer. A lot of the customers who are at this end of the scale, they want to use the technology for its ability to detect things quickly.
Really what we're most excited about in terms of SmidgION, ASIC is its 400 channels, its time to results, and its ability to go into very low power, very simple to make devices, which should help us really accelerate that decentralization and those tests and that deployment of our technology in things like pathogen sequencing, pathogen detection, in factory settings and low- and middle-income countries, all of those applications that are being built today on the back of MinION, but really need to scale and accelerate.
Thank you. If I can squeeze in another, please. There was a slide shown in the conference about partnerships. There have obviously been a number of commercial partnerships that you or your partner has announced. On that slide, there were a number of academic institutions as well. I just wondered what's the nature of those sorts of partnerships versus them being a customer and, you know, what's the strategy behind those partnerships, please?
The way the slide was set up, the more applied stuff ends up being with commercial partners such as Pufferquest in biomanufacturing QC or 4bases for same day breast cancer and so on. The programs that we are doing with NIH on Alzheimer's, with Exeter on hard to diagnose rare disease children, with GEL at Cancer 2.0, these are foundational, translational from discovery. We know that long reads and native DNA and methylation are important in cancer or Cancer 2.0, which are.
These are now Well, not experiments. They are proof of concept studies providing clinical insights into cancer for the NHS in preparation for potential rollout. They are just as significant. They tend to be more academic-oriented, seeking the clinical insights, versus the ones at the other end of the spectrum with company partners which are more advanced and closer to a go-to-market strategy and timeline.
Perfect. Thank you.
The next question comes from the line of Paul Cuddon calling from Numis. Please go ahead. You might be on mute, Paul.
Sorry. Thank you. I was ye ah, two questions, please. Firstly, on sort of workflow and bioinformatics. I mean, I think the talk mentioned several times about how important bioinformatics will be to get out of the kind of centralized lab, to decentralize of what, is being done there. Secondly, on kind of sourcing the location of tumors using epigenetic profiles, whether there was anything sort of beyond the previous papers, whether you can do more, early-stage cancer diagnosis and sort of localizing of tumors.
On workflow and bioinformatics, we've been investing really heavily into our Epi2me platform. This is a product that sits alongside all of our instrumentation. It has automated workflows inside. If you are sequencing a soil sample, you can click on a button that will tell you what's in your sample. It will do a metagenomic analysis. We've had that actually for many years. What we've really been focused on over the last year or so is the more advanced and the more heavy lifting types of workflows. Single-cell workflows are all in there with a simple click button. Human variation, a lot of the
Again, in the same way that Gordon's just said that we spent a lot of time with people like Stanford developing that workflow for them to be able to sequence a human genome and come out with an answer in about 8 hours. We've taken a lot of that development and put it into a smooth pipeline, so any customer can now press a button in Epi2me Labs, and they can upload their human data, and they can get an answer, a single report that has structural variation, SNP, indel, methylation, all in one report, as opposed to running a lot of command line. That's been really critical in order to accelerate adoption, is making sure that the bioinformatics is really really simple sorry, your line broke up during the, during the second question. I don't know if you could.
Yeah. Thanks, Rose. It was just on the sort of localization of tumors using epigenetic profiles from, I think it was blood samples. Were there sort of more kind of papers that come out on that topic and potentially a pathway to sort of broader clinical use?
Yeah. There were a lot of talks, a lot of cancer genomic talks at London Calling. I'm still catching up on some of them post-event. It is now, it's quite clear that the epigenetic profiling is highly valuable in being able to determine methylation, but it also helps you determine tissue of origin. If you're doing cell-free DNA, for example, and you are, you're picking up DNA fragments, you can tell which organ they've been shedded from. Therefore, if you're finding cancer markers, be they sort of standard like SNP markers, for example, or you're finding methylation markers, you can also do tissue profiling based on that to see what tissue has this come from. There's a lot of foundational research work going into that, and there were a few presentations about it at London Calling.
Brilliant. Thank you. I had submitted a few questions, through text beforehand, but probably because they're offline. Thank you.
Thanks.
The next question comes from the line of James Orsborne, calling from Stifel. Please go ahead.
Yeah. Hi there. Thanks for taking my questions. Just quick one on homopolymers. Seems to be an area of continued research for you guys. Saw the slide around you developing the modified bases to improve polymer reads. Just wondered how far along this research was, you know, and if we can expect an update in the near term or just kind of a long-term project for you guys?
Yeah. It's highly active as Clive said in his talk. Yeah. I think I would expect us to be constantly and continuously updating the community on how we're doing in terms of our homopolymer performance, which has drastically improved over the last year or so with the introduction of R10. The R10 Nanopore has taken us to a point where we could determine homopolymers of maybe five or six bases into something that is more like 12 to 15 bases. There is still a lot of machine learning development as well as we start to improve our visibility beyond that. Yeah. Lots of things happening in that space. It is definitely one that as soon as there's something which is customer-facing, it will be spoken about very, very quickly.
Great. Thank you.
We have another question calling, coming from Odysseas Manesiotis, calling from Berenberg. Please go ahead.
Hi there. Thanks for taking my follow-ups. I had one on the Duplex cost per genome, if you could give us an answer to that, Rosemary? Am I right to say that given the figures you provided earlier, you get about 55% Duplex rates for human samples, and would getting that closer to your overall Duplex rates of around 80% be realistic within the next year?
What we showed was actually duplex rates of about 80% now. That's what we showed at London Calling. I'd just go back to this. A simplex genome is what you need for the vast majority of the research out there. There if you're doing one genome per flow cell, it's, you know, $690. If you're doing two genomes per flow cell, it's $345. Please remember that, you know, simplex has all of the answers that the vast majority of our customers want. When you're doing telomere-to-telomere genomes, you need Duplex data, you need ultra-long data in the same way that today you need three or four different data types today for telomere-to-telomere. You need Duplex and ultra-long, and that's why we've got. It's not easy to put a price on it, because there's so many different elements that go into it.
Thanks. I misunderstood the 50 gigs from the 90 gigs part you touched on earlier.
Yeah.
on the Simplex one as well. I mean.
you get 50 GB
Yeah. Just the last one for me.
You want one flow cell. Yeah. In a flow cell that costs you $600 on a PromethION, you can get 50 gigs of Duplex and 40 gigs of Simplex. That's 90 gigs total of human data that comes off a flow cell that's cost you $600.
All right. No, that's so clear. Last one, just something Clive mentioned in the presentation. On the P2 Solo hardware bag, I think Clive said around 20% of them were replaced. Has that been a material unexpected cost on your side, and were the replacements easily done?
No, because it was a board, so some of them hadn't even been opened by the customers. We were able to replace a board in an electronics board in the platform, so no, it didn't have a material impact on our margin.
Perfect. Thanks, Gordon. That's it from me. Thanks for taking the follow-ups again.
Yeah. I think that is, we're at the top of the hour. Thanks for joining us today. A replay of this call will be available on our investor relations website shortly, and customer presentations will also be, being added. Hopefully, you'll get a chance to look at those as well. Last but not least, we are hosting our first Capital Markets Day on the nineteenth of October. Please log on and register your interest for that as well. Thanks, everyone. Thanks for your time today. Thank you, Rosemary.
Thanks, everyone.