Welcome to Wednesday at the J.P. Morgan Healthcare Conference. It's my pleasure to introduce you to the Molecular Partners session. I'm Richard Vosser, European Pharma and Biotech Analyst, and with Molecular Partners, we have Patrick Amstutz, the CEO. Before I hand over to Patrick for his presentation, I'd remind people that when we reach the Q&A period, you can put your hand up, and I'll come and find you with a mic. Or you can introduce questions through the portal. Patrick, welcome to the conference. Thanks, Richard, for the kind introduction. And also, actually, there is a little special anniversary for us, and I know you know. It's our 10th time here. So we went public 10 years ago. And since then, Richard has been on our side and has been going through the ups and downs and lefts and rights and has had to learn about new therapeutic areas as we went there. And a really big special thanks for supporting us on all those movements. So big thanks for that. My name is Patrick Amstutz. I'm the CEO of Molecular Partners, and we have now 40 minutes to go through the story. It is a really special year for us, 2025. I know many of you had investor meetings, and the atmosphere is not all that good. But I will quote Seth, who is our head of IR. He says, "We were so often wrong in the last years when we thought, 'Now the change happens. Now it's a bit more gloomy.' So maybe now the change will happen." So that's his take on the mood. At Molecular Partners, we're actually in a very good mood. And the reason is we are back on the map in oncology, and we have two clinical key readouts. And one is in radiotherapy with 712, our DLL3 DARPin. And the other one, and it came a bit back on the map, is our AML drug 533. And in the next 30 minutes, we will go through those topics, and I will introduce you a bit further into what we're doing. Here's my disclaimer, as always. And for those who are maybe not quite that familiar or are new to the story, I see some familiar faces from earlier times. Hi. We will definitely go through this. What has not changed in the last 20 years is really the focus on the patient value. Whatever we do is a therapeutic and is gauged for value for patients. We have built key capabilities and skills over the last year. This really was a long build over those 20 years where the technology evolved. We know DARPin is more a therapeutic modality. Now we talk about our technology platforms on radio and the immune cell engagers. The team is capable of moving that, and we have also partners to support us where needed. On the financing side, we're listed in Switzerland. We're also listed on the NASDAQ, and we have CHF 150 million in cash, which brings us well into 2027 and to key value inflection points. Just a few words on strategy and DARPins, starting with DARPins. Over the last two years, and also in the radio space, there is sort of a bit of new word coming up. When we started, DARPins would have been classified as scaffolds. Now they're called mini proteins, which I actually like quite a bit. The whole field of AI and mini proteins is taking off. DARPins definitely belong in the field of mini proteins. They're a tenth the size of an antibody. We have so far had seven clinical compounds on DARPins and treated 2,500 patients. We know quite well what we're doing. We also know what DARPins should do and what they should not do. On the strategy, it's clear we will build on the DARPin. We don't want to do me-toos or drugs to antibodies or others, but really pioneer a space, transform a space based on the DARPin. We go for true patient value that translates into high medical need. These are life-threatening diseases, and the early clinical readout allows us to get value upfront, and it translates into single-agent activity. As I said before, we don't have to reinvent the wheel. If there is a partner out there, we will approach them or they approach us, and so we have been working over the years with many, and in this point, it's especially Orano Med. I will come to that quite a bit, and also our partner, Novartis. For those who do like pictures, this slide on the two platforms we are working on. On the left-hand side, the radiotherapy DARPin, where you see the DARPin delivers an isotope that then kills the tumor. While on the right-hand side, the immune cell engager, and it stands for a BiTE too, you see three binders to the malignant cell, and the green one is a CD3 binder to bring in the T cell that then kills the cells. And these are obviously single-agent active, and both of them are hot fields. And I think we were well positioned, and we did foresee that at this point in time, these two platforms would be really sought after in our industry. This translates into our pipeline, where you see the radio space on top. We have four arrows depicted. There are more programs ongoing. The two that are written there, it's in small cell lung cancer, DLL3. That's a target we introduced a year ago. Today, we add mesothelin, and I'll come to that, why mesothelin, and we expanded the Orano Med collaborations. I'll come to that. But there we have actually in total 10 candidates or 10 products in the end that can be replaced, so many more targets possible. On the lower side today, I will be focusing on 533 and the switch platform. Let's move into the radio space first. Here, I think most of you will know, and the idea is old. I mean, radiation is maybe one of the oldest ways to treat the tumor. It is also a way where there is very little resistance if you do it well, especially with an alpha emitter that causes double-strand DNA breaks. The whole trick here is that you have a target on a tumor, you have a specific vector, and then a linker chelator and an isotope. And so site-specific, you deliver the isotope that kills the tumor. And the picture here is from an image from Novartis showing a PSMA-positive prostate cancer patient that has many metastases. And after three treatment cycles, all are gone. And that level of success obviously sparks interest, and there is a huge group of followers here and innovation being done. And the one problem to single out is that the word there is PSMA, where there are like 60 programs on PSMA, but it's much more difficult to take this idea and bring it to other targets. And that's where our space is. And so how do we go about that? And this is our construct. That's what we're working on today. That's the DARPIN on the left side. And the DARPIN has intrinsic properties, the high selectivity, affinity, small size. And here, especially the broad target space. So against almost any target that you will select, you can make a DARPin. The DARPin itself had to go in a transformation to become a radio-DARPin. And that is the surface engineering and then the half-life extender or half-life extension that we add to make it the ideal vector. The surface engineering is tuned to get the kidney down. So that's what we do, too, and I will show some data on that, how we can reduce the kidney uptake. The half-life extender brings the tumor up, and we will also see data on that. And then that is linked to a chelator where we add the isotope. On the isotope side, we partner. So we don't have our own supply. And with whom we partner is Orano Med, and they work on lead. We like lead because, A, it has a proven efficacy. They're in their phase three towards registration with their lead program. You have a lot fast and high energy deposition in the tumor. It is safe, and you have a good waste management, and so just on Sunday, we announced that we have added to that collaboration, so we now can do 10 products, or we have access to 10 products with lead with Orano Med, and Orano Med, I'll just quickly back up there, they have 22,000 drums of thorium, which is the precursor of lead, and so they have quasi-unlimited supply, and that is obviously strategic for us, but it is more than that. It's not only the supply. It's also that they can have a great hot lab, an animal facility, and together, the culture is also great to work with them. We can rapidly cycle candidates to find the best ones that we then want to move forward. Sometimes in the beginning, it was up to 10 cycles we had to do for one candidate. This we're bringing down to maybe two or three that we become much faster and really use the platform approach here. The other point here is that Sanofi made a big investment in the lead program of Orano Med. Orano Med has, in that collaboration, gained access to EUR 400 million to build late-stage capabilities and the supply. All of that will obviously be very helpful for us when we come with our first product, which hopefully then will be our DLL3 targeting MP0712. DLL3 is in small cell lung cancer, a cancer that is very chemo-resistant and progresses fast. DLL3 is specific there, and it is rapidly internalized, and it is validated by Tarlatamab, a T cell engager. Tarlatamab has followers, especially with ADCs and obviously then radiotherapy. We get the question like, will one of them solve the problem? Our feeling is much more it will need all of these drugs that will be given in some sort of sequence as the disease has shown to be highly resistant or build resistance to different types, especially chemo, which also will pull in the ADCs. Let's look at some data from what we have here. What we published early last year was the green part here. That is the MC38 model, which is a high tumor expression model, as you see in the middle. You see we have engineered a bit longer half-life. That's on the bloodstream. You see the 14% and 23% in the kidney. So we brought down, and you should know we started at around 150% in the kidney, tenfold reduction in kidney while the tumor is almost tenfold up and now reaching 21% and 58%. So we have a positive, a very positive tumor-to-kidney ratio. The question we were then following up on was more, can we also translate this into a low expression level that is more matched to the human situation that you see in the NCI-H82 model? And yes, in this IV model, we can show very nice that we have almost the same distribution as in the high-expressing model. And our theory is that it is because of the high internalization of the target that we get a lot also inside the cells. This translates into very good efficacy on this model. I will not go into much detail, but we kill the tumors and they keep down. Maybe more importantly, one slide on safety. As we do have a bit longer half-life, there will be a decay of lead in the blood. So yes, the body weight is important, but our focus was on the blood cells where you see a little dip. What we were looking for is that that dip will come back. So we see, yes, blood cells will be impacted. We will likely or for sure see that also in the clinic, but they recover fast. From the doses, our partner Orano Med, who has the clinical experience, reassures us that those 10-30 microcurie are exactly the space you want to go. What are our plans for this year? Last year, I had to say next year. Now it's this year. We will start with Phase 0 imaging. The beauty of radiotherapy is that you actually can do imaging with lead-203, a low-energy compound. You get the full dosimetry. You will know where the compound is. That can directly then lead to an understanding what you will see in the phase one efficacy trial. We will have a very early value inflection point from the phase zero. This data we will get this year that will guide, A, the selecting the dose for phase one. It will also guide the value creation and will be a real value inflection point for us. It's a typical phase one that then leads into a phase two that will then happen for sure next year. Now let me introduce also the next target. Our research teams at any point in time are working at around four to six of these targets in parallel. And then we show those where we actually have the breakthroughs. I'll start with ovarian cancer. Ovarian cancer is a difficult-to-treat cancer because it is also chemo-resistant. And the T cells, they are there, but they seem to be not very active, possibly exhausted or inhibited. So radiation is a good and valid way to attack the problem. Mesothelin is an interesting target. It's expressed with high prevalence on many of these cells, also on high levels. The problem is it is shed. It is cut from the membrane. So you have a shed form. You have a full-length bound form, and you have a membrane proximal form. If you now have a binder that we would then call the distal binder that would bind the shed form, that's not helpful. So we came up with binders that only bind the full cell-bound form and the membrane proximal epitope. If you now take an experiment, and we actually have obviously binders for both and actually many more for the distal form. So 99% is binding distal while only a few bind the proximal. You see the distal part here, the yellow one is inhibited by free soluble and shed mesothelin while the proximal binds the cell. So we can now weave through this high excess of soluble mesothelin and target the cells. And that will be our next product. And it is being advanced with our partner Orano Med, and we will give an update at around AACR. This is sort of the outlook slide for this part where you see we have the tumor cell binder there, the shed part that was an outlook that we have now fulfilled. And we're working on others. I think the point to highlight here is the two-in-one DARPIN that we will also see in other sections. There we can actually then target two different targets on tumor cells, maybe even tumor cells and fibroblast cells or tumor stromal cells to get a better coverage of tumor killing. Let's come to 533. In 533, we had maybe underestimated a bit in our development. It came a bit as a Christmas present last year when we had a DRC. We actually looked at some of the new responders, and we had a nice surprise that I want to share with you today. 533 is a tetraspecific T cell engager. You see three binders on the malignant AML cell. It's an AML. You have CD33, 1 2 3, and 70 targeting CD3. That's the blue DARPIN. The idea behind this is that because these targets are not 100% clean, that you actually get an avidity-driven mechanism where you preferentially kill AML cells while you don't kill the healthy cells. We started the program, and this is what was in public domain. This was published at ASH. We did a dose escalation, dose range one to seven. You see we had sporadic responses, but you really didn't see a lot. Here my reference to Monty Python. If you go to the pet shop and see this, you think the parrot is dead. In all fairness, I think we were about to leave that pet shop when we had a great meeting with our KOLs at EHA. We also put that in perspective to the PK data. What we found is we were heavily underdosing. And if you see here in green, that's the predicted therapeutic range. You see we were barely scratching the therapeutic range. And the idea, and this was we saw less exposure than calculated because we didn't have or we were underestimating the TMDD, the Target-Mediated Drug Disposition. So we saw less. So what do you want to do? In this case, you want to up the dose. So we added an orange cohort there. We have a steeper uptake, one additional dose. And here I will talk about that. But in principle, we want to go to daily dosing. So the green is what we're excited about. The data I will show you now is on dose range eight and was, I call it a bit of surprise to us on the magnitude of impact. First about the PK data. So now here you see the blue. You were scratching the therapeutic range. With that additional steeper ramp-up, an additional dose at day 12, now we're hitting the therapeutic range quite often. Not perfect, so green likely will be better, but yes, we are there, and this has actually now translated into a very different responder profile. At data cutoff, December 16th, we had eight patients in. It was safe. It was manageable safety. We had five by on-cycle one, and three of the five have now developed into complete response, which is very different than we had seen before, so we think we definitely have a signal there that we need to follow. This is, again, it's far from victory, but it is a signal that we need to follow. We and our investigators are highly encouraged to now, A, expand dose range eight and go to the green, dose range nine, 10, and 11 to find out how we can dose and develop this drug in AML, the most deadly liquid tumor. Working on that product, we were deep in the immune cell field, so in the T cell engager field. We were following the field and the team around Anne, who will also come up here for the Q&A. They were analyzing what would be needed to bring the whole field forward. They came up with three dimensions. First, lack of tumor-specific targets. We have sort of addressed that with the tetraspecific. Narrow therapeutic windows, so T cell engagers are highly active. Can we change that by masking or conditionally activating the DARPIN? Yes. And others, especially here, I would like to point out Janux and also VR have nice data showing in conditionally activated settings. You can dose more, and you get a higher effect. And still, maybe that's not enough. Maybe you need a co-stimulation effect on top of this. And that is work that has been done by many, including CD28 by Regeneron, for example, or CD2 that we also like. Can you add that? And all of these dimensions, you have to or can combine two molecules that look like this. You have gray half-life extender, so you don't have to dose too frequently. You have a tumor-associated antigen one in green. Then you have the trick, which is the masking DARPIN. It binds either/or. A tumor antigen can be one or two. That will be an AND gate. Then you have block, the CD3, and you add a co-stimulation DARPIN. In our case, we like to use CD2. So in circulation on healthy cells, nothing happens. The construct is in off mode. If you now go to a setting where you encounter a tumor cell with the two targets, in this case, we took Mesothelin. We took EpCAM as a co-express target. And then the construct will open, and you see killing. And in this model, in vivo, let me quickly walk you through what we did see. Let's start with the red line, which is the conventional T cell engager, Mesothelin, CD3, the TriTAC. That's what you would expect in this model, a medium inhibition or killing of tumor cells. We'll now go to the green line. That's the Switch-DARPin. You see the CD2, CD3 adds on top. Obviously, it's a difficult model. We add activity on top. We kill better with this model. And at the same time, and this was sort of an internal control, we didn't quite foresee this. The purple line is the unswitched construct. We see no activity. Why is that the case? It is the case because the CD2 will cross-link T cells with CD3, and they will kill each other, fratricide. And that's shown on the blood sample there. There's simply no T cells anymore because you have so effective killing. And others will then tune down the CD2 or CD3 activity to find a way to tune that. And in our case, we don't have to take any compromise, but we can actually go all in in full activity on these cells and don't have to compromise. And this data is going to help us select a program for ourselves, but it is also a platform where we are opening and in discussions with potential pharmaceutical partners. So let's take a deep breath, re-emerge from the machine room, back to the outlook, and see kind of what is ahead of us to come this year. Let me start with 712, and I will keep it simple. 712 radio is all about the image. Everything we're doing is to get that image that will guide further development and, if you want, validate the platform. Next to that, we will expand the portfolio. We will update on mesothelin and hopefully then also on additional programs and strengthen the collaboration further with Orano Med, which really gives us that strategic access to lead. 533, I think there's two prongs there. It's about this cohort eight. Is the cohort eight already where we want to go, or will we want to go to a cohort nine and beyond? Cohort nine will be started in February. So second half, we will have data for sure. And in first half, we will have much more data on cohort eight, deeper dose response rates, and especially also we want to follow the duration of response. And as I said on the Switch-DARPin, it's about selecting the first candidate or program and looking into potential partnerships. I'll end the outlook with we're well-financed, so we're not at this point in time stressed for funding. Our cash brings us well into 2027. And with this, I do want to thank you all for your showing up here, especially also the people online. We will be open for your questions. I do want to thank all the coworkers. We have our 20 years logo there. We started with six. We're now roughly 150, and many have come and left the game. And we are in very close contact to all of the alumni and everybody who has brought Molecular Partners and DARPINs from where we started to where we are today. A big thanks from my side in the name of all the co-founders, but also the coworkers there. Then obviously to our partners, I mean, Orano Med, Novartis, great partners to have. Then our investors who have supported us all the way. And most important, obviously, the physicians and the patients in our trial, which really give us this energy to stand up, get up every day, and pioneer this space. And special thanks to Richard for being with us all the time. And with that, I would ask my colleagues, Dani Steiner and Anne, to join me here on podium for the questions. And we have another, I think, 15 minutes for that. Excellent. So first of all, any questions in the room? Maybe I'll start. Maybe we could go with 533 to start with. And obviously, great to see some of those responses. Maybe you could just talk a little bit in terms of how you see put that into context with what you need to see to move forward. Do you need to see more consistency? Do you need to see duration? What do you need to see in cohort nine? So the bar overall for an AML product is roughly a 30% response rate and well beyond three months duration, ideally six plus months. So that has not changed from the beginning. That's why my reference to the partial is dead if you don't see that. So we didn't see it. Now, if the cohort eight already brings us there, we'll have to find out. And now we're then going into this even higher dosing frequency where the protocol is a bit open. It does allow up to daily, but you don't have to treat daily. So there is a level of flexibility. And I must say, we also spoke to a few of the pharma companies, and big kudos to our team. This is proper development. We're not rushing this and just throwing out the dose and declaring victory. We really want to understand what is the best way forward to then have the best success. And maybe the daily is better than what we see now. That would be great. What we will want to see is the 30% response rate over a bigger cohort and the duration. Can you monitor the level of drug in real time to see what the coverage is or? Well, not real time. No, no, but we will go back. And what I maybe left out a bit is we are not optimizing the peak dose. You saw dose range seven was not better than dose range eight and even five. It's really about hitting the right dose consistently. Then, and as I was referring to the target-mediated drug disposition, once the disease is gone, we actually have much better exposure. So over time, you then go after cycle one, you could call that a loading dose. You go to weekly or even biweekly dosing schemes. And how's the tolerability in the cohort eight? The tolerability in cohort eight is very comparable to cohort six, where the peak dose is matched. So far, that makes sense. And I think, maybe Anne wants to comment on that, but I think what you actually are trying to not have is boosts, is individual peaks, because in T cells, that means go and then stop and go. And the stop and go causes IRRs and CRS. But maybe Anne, as the T cell expert, better to have a constant activation than an overactivation? Yeah, I think in terms of reactivity, most certainly, but also in terms of activity against the tumor, it's a balance to find basically between the overactivation and the reasonable activation of the T cells. And I think we're getting closer to that with cohort eight. And you talked about interest with pharma companies. Let's assume that cohort eight gets that 30% response rate or cohort nine. What are the next steps after that? It definitely will depend a bit on the signal. I mean, if it is groundbreaking, we are well above 30%, and we have a clear fixed dosing regime, there is a path forward for accelerated approval. There is such a high medical need. Our investigators are stressing that in every meeting that we have. Don't drop the ball, and that would be even a path we could for the fast approval do ourselves. Then maybe look for a partner even while we're doing that. I will there connect it to the other platform. So we will have readouts from both sides. If both work, there will be different strategic decisions to be taken. Then if radio maybe underperforms, this looks great. This becomes then maybe more the center of investment, so the ability to have these options is what gets us very excited this year. Talking about Radio-DARPin 712, good to see the improvement in terms of the kidney ratio, but maybe you could put that into context for us in terms of how good is it? What do you need to see for that to be viable? Yeah, so the real viable number is a one to one, but I would love to hand over to Dani, who is by now a real expert in this field, has spent all the time, knows all the competitors, and how he sees this number where we are and maybe also in the context of lead. Happy to follow up on that question, so it's like starting from it's very well understood how much energy you need to deposit on the tumor. There, people are talking about like 50 gray will really be the energy needed to kill the tumor, to really get rid of the tumor. Then from this number where you say, I need a 50 gray on the tumor, you know what is the most sensitive organ? This is the kidney. So if you are at the one to one, one to one, two to one, you are really in a good space, in a good starting space in the preclinical model since the kidney has a dose limit. Today, they say for bone marrow to meet the 30 gray, but most people say this will probably in the next years to come definitely go up. What people have been seeing from the preclinical model to the clinical side, normally you get even a shifting in a positive direction in this tumor to kidney ratios. That's the assumptions we are taking and the field is taking in that context. Makes sense, and blood as well. Can blood be a barrier here? So blood, here we have to be very specific on the isotope. If you're talking about beta-emitter, blood is nothing you want to have because it has a far pathway and you will kill a lot of blood cells. If you h ave an FC, or if you're antibody-based, you don't want blood. You bind to so many immune cells that you will kill. That's not good. So we have to look at blood in the context of lead and HSA binding. And maybe Dani can follow up on that. Yeah, happy to start. So starting from what we've been seeing preclinically, so the molecule is safe in terms of toxicity, safe and well-tolerated up to the 30 microcurie. This is roughly three times above the human equivalent therapeutic dose. So 10 is the human equivalent dose, what you would need in the therapeutic setting. So three times higher, still well tolerated. And you see this full recovery of the blood after the 30 days. And from our partners, we know that if you have a molecule that causes heme tox, this dip will be much more pronounced and much longer. So the duration will be much longer. And I think that the context why this is happening is like the first one that Patrick mentioned. We are talking about an alpha-emitting particle. So the transfer from the energy from what you have decaying in blood to the bone marrow is very low because the distance of this radiation is short. The second thing is like we have a shorter systemic half-life than the antibodies that are in clinics in combination with alpha or beta-emitting particles. And the third element there is we're using albumin binding for extending the half-life. So there is no Fc involved. And it's well known for antibodies that the FC receptor interaction with the hematopoietic system in the bone marrow leads to an enrichment of an interaction with these cells, which is of course very disadvantageous if you use alpha-emitting particles that then decay on the hematopoietic system basically and kill that. So I think that what we're seeing preclinically is well supported by the theoretical argument. And of course, we will see in phase 1 how these turns then into reality. That's something we will need to follow closely. You talked about the obvious imaging, and you can see because you'll be able to see it in the kidney. You'll be able to see it around the body and measure it. How fast can you go if you see the right profile into sort of treating patients with lead? I think you still have to run the full trial. You just have a very strong value inflection point there, which allows you to be more confident on your investment. I don't think you can be much faster. It's a clinical trial. What you can do is obviously imaging of each patient before you treat. That actually is done. So you do image a patient to make sure they have the target and how much target before you treat. Then it is a simple dose escalation. I do think obviously the starting dose might be a higher one. So you might have to have less step-up doses. That can be faster. But in general, you'll have to do a normal phase 1 trial. I'll also say that this is a very new field and the regulators are very open for dialogue. And I personally am very close in Switzerland, but there is a real interest to collaborate with industry to find the best regulations to improve also clinical development. And can you remind us with the Orano Med collaboration in terms of, I mean, they're going to do manufacturing, but maybe the economics, how this is going to fit together, how commercialization might work. I know we're getting a long way ahead, but. Yes. And I think let's say on the commercialization, if we have a drug, we'll find a way to commercialize and monetize. I think the sticky point there is the supply chain and manufacturing. And that's why I like to point out that this obviously has not been done for lead. And Orano Med is likely the first group to do it with their AlphaMedix product that is now with Sanofi. So Sanofi will be in charge of commercialization, but they are building the full supply chain. And I do think that we have as a partner the alpha supplier of choice with deep pockets, very long understanding of radiation. Orano, the mothership, is the French nuclear power company. So they know how to do this. They have the deep pockets. And I do think we have a great partner for it. It still has to be done. And that is where I personally have my focus on following that field. And are there any nuances with targeting mesothelin for the second product relative to DLL3? So in relation. Yeah. I think the target choice was in a way similar, but DLL3 can be approached also with other modalities, including smaller entities as we know that Novartis has a peptide from DLL3. For mesothelin, yes, you will find those, but not for the membrane proximal part. And maybe Dani can talk a bit about our campaign where it was a lot of work to find those. And the only other modality that is described are antibodies. And antibodies are not the way to go here. So from the differentiation point of view, we see mesothelin as more differentiated on the binding properties. Maybe Dani. Yeah, I can add on that one. And And maybe I start with a bit of a zoom out. So it's the question regarding target choice in general. What are we going for as a company? We have many of the parameters that many companies will follow is like selectively expressed on tumor, clean on non-tumor, high medical need. All these aspects are common. Where we are looking at it from a different perspective is like, is the target we are addressing, is there something that we can do in a specific way? Is the target difficult for other modalities? And here, especially in the radioligand space, we're looking, is it easy for small molecules or cyclic peptide? If this is the case, it's a target we're not going after. Because we say, if it can be covered by cyclic peptide, by small molecules, this is the perfect choice. They should cover those targets. We're looking for targets that are really challenging for those modalities. And that's where, for example, mesothelin came in because it's a very tricky epitope of mesothelin to target. What we're knowing for cyclic peptides, they need to be able to screen thousands of cyclic peptides to find those that behave well, those that give the desired biodistribution profile. Now, if you narrow that down to a very tricky epitope, you have only a couple of hundred if you're lucky. The chance of success for a cyclic peptide on something that challenging will be very low. That's the way how we're looking at target and picking target to really bring in that DARPin differentiating aspect to ideally work on things where we say we can deliver benefit where it's difficult for others to deliver that. Building on that, the structure that you have created for DLL3 in terms of the DARPin and the kidney ratio, does that then translate very well across to the Mesothelin DARPin? Is that translatable? I can take that question. So it's like it will be. So maybe I start here. It's like a lot of the learnings we've been doing on DLL3 and programs we've been killing on the way. There, of course, have been other programs for whatever expression level, model reasons we have been stopping on the way. A lot of the learnings we figured out are transferable. So there were a lot of similarities in the mesothelin program, but of course, we've been doing improvements on the way and we're applying these improvements. So I expect in the candidates to come a continuous improvement in terms of the profiles we are reaching. And that's what we're aiming for, continuously integrate and bring it to a next level. But overall, I say the architecture of the molecule, including the surface engineering, the half-life extension, that probably will be a part of the architecture of the molecule that will stay. Excellent. I'm afraid we're out of time. Thanks very much. Thank you. Thank you.