Good morning, everyone, and thank you for joining us today at Needham & Company's 24th Annual Healthcare Conference. My name is Ethan Markowski, and I'm a member of the biotech research team here at Needham. Joining me today is Eran Ophir, Chief Scientific Officer of Compugen, who will provide a presentation on the company, followed by a Q&A session with the remaining time. With that, I'll go ahead and turn it over to Eran. Thank you, Ethan, and thank you to the Needham team for inviting us to your conference. Compugen is a clinical stage computational discovery company that focuses on the field of immuno-oncology. We use our computational platform called Unigen to identify novel drug targets in the field of immuno-oncology. The platform is validated by multiple targets, making them all the way from computational discovery up to clinical testing, and also validated by our partnership with a few pharma companies. Our pipeline currently is eligible for up to more than $1 billion of potential milestones and tiered royalties. Using our computational platform, we identified TIGIT around the same time Genentech did, and then a few years later, we identified PVRIG. We have in our pipeline two fully owned assets: COM701, which is a potential first-in-class anti-PVRIG antibody, and COM902, which is a potential best-in-class anti-TIGIT antibody. We also licensed the right to use our potential best-in-class TIGIT antibody as part of the TIGIT PD-1 bispecific rilvegostomig with AstraZeneca. They are now moving in multiple Phase 3 studies with this molecule, and they are expected an estimation of sale of more than $5 billion. A potential blockbuster for this molecule derived from our TIGIT antibody. More recently, we used our computational platform, Unigen, to identify another first-in-class potential asset and a really novel way to harness cytokines for immunotherapy. We developed an antibody that was licensed last year to Gilead, called now GS-0321. We'll elaborate a bit later about this unique asset and the deal terms. We have some other undisclosed early-stage assets, all of them derived from our computational discovery platform. We have a solid cash balance to the end of last year, $103 million, that should enable us to runway into 2027. Let's discuss a bit our pipeline. I'm going to discuss first our first-in-class potential asset, COM701, which is a PVRIG blocker target, again, that we identified computationally. As you can see, there are some movers that follow us, some with monoclonal, some with bispecific antibodies. We are the first to move, and we have a very unique clinical program because we rely on the fact that not only we identified PVRIG as a novel checkpoint, but we also realized that PVRIG has a very different biology from all the other checkpoints, seems to be dominating other stages of a cancer immunity cycle. Therefore, the activity we have seen for COM701 across different indications and across a few clinical trials seems to be very different from what we see from other checkpoints due to that unique biology that has the potential to enable checkpoint or PVRIG blockade in indications which are typically not responsive to other checkpoints. One of these indications is ovarian cancer. Even before we started, we knew that PVRIG has a very high expression in the pathway of PVRIG, has a very dominant expression in ovarian cancer. Given its unique biology that enables or has the potential to enable checkpoint blockade in less inflamed tumor types, we started by testing COM701 in monotherapy and then moved into combinations. The signals we have seen there supported activity in platinum-resistant ovarian cancer in last-line treated patients. We saw a signal that was above the signal we have seen previously from other checkpoints, from what was shown for PD-1 or even PD-1 plus TIGIT. We think that we have a signal for PVRIG blockade in platinum-resistant ovarian cancer. We made a decision after publishing the results at the end of last year in SITC. We had a few options. One of them would continue in the last-line patients. After we substantiated the signal in the last-line patients, we decided to move earlier. We have seen that in our trial, the patient who did respond had a very good durability, very good safety profile. We thought that this is exactly the kind of drug that could be relevant, actually, in earlier settings. We're going now into the platinum-sensitive settings, in which patients have a more potent immune system because they are less exhausted by all the chemotherapies. We are taking patients who receive, after second or third line, received the platinum chemotherapy, six cycles of platinum. These patients have now lower tumor burden. Also, the platinum have shown to sensitize the tumor to the unique MOA of COM701. Again, due to the fact that COM701 is well tolerated and associated with durable responses, we thought that to start with COM701 monotherapy in that settings, so it's after the platinum as the maintenance settings. Obviously, also, the platinum settings is less competitive. ADCs, which are now starting to dominate the platinum-resistant settings, are trying also to make a move into the earlier lines. Most of the trials are more to take these relatively toxic agents, the ADCs, to replace maybe the platinum chemotherapy. The maintenance settings is still a place of unmet need that we think that COM701 could fit. Looking at these trials in a bit more details, we're talking about second or third-line patients, which are still platinum-sensitive. They receive the platinum after they already received BEV or PARP as maintenance in earlier settings, or maybe they're not eligible because of different comorbidities or other reasons to receive BEV and PARP. They receive now the platinum, six cycles. Now there's no standard of care for these patients. There's really an unmet need. This is exactly the place we're going to start with our adaptive trial design. We're starting with COM701, 2:1 randomization versus placebo. Again, this is the standard of care. There is no standard of care for these patients. We're going to report these results in H2 2026. We're looking for progression-free survival and obviously to continue to see the safety. The historical control for this patient population is quite solid. Seems like around 5.5 months. We expect to see something similar also in our internal placebo control. We think that if we'll be able to extend that by three months PFS, this will be a go to continue to other parts of this adaptive trial design. Again, this we're going to report in H2 2026, and we're going to start the trial in Q2 this year. After seeing the results, assuming they are positive, of course, now we know that COM701 in monotherapy has activity in these exact settings in which it was not tested yet. We have a few options. One, since we talk about a randomized trial, which is blinded, we can, if the results are promising, we can continue and enroll and start thinking about accelerated approval. After substantiating the single agent activity, we have the flexibility in the trial design to now add BEV or to add triplet of PD-1 and TIGIT and move forward in these settings, go earlier potentially in another trial if the results are promising. This is for our COM701. This is the next milestone for COM701. Let's talk a bit about COM902, our potential best-in-class anti-TIGIT blocker. We all know that TIGIT started in the clinic very promising. We had the CT-SCAPE data with very promising hazard ratio in randomized study. There are a few other randomized phase two studies who show that TIGIT is active. Then in the large trials, multiple phase three trials from some of the, I would say, the first movers in the field, Merck, Genentech, we did not see these Phase 2 positive readouts translating into improved overall survival or even PFS in large trials. Of course, this is disappointing and shed a cast on all the TIGIT assets out there. We need to remember that not all TIGIT assets are the same. I think one of the biggest differences is the FC tail of the antibody. Most of these first movers that had the disappointing results in the phase three had an active FC, meaning that in addition to blockade of the receptor of the TIGIT antibody, they also have an active FC tail that could deplete cells, could manipulate cells, could do other things that obviously the companies who went forward with the molecule thought are beneficial, but also have potential side effects. With the caveats of comparing across trials, we do see a strong trend that TIGIT antibodies with active FC tail have a worse safety profile. About the efficacy, it could also harm the efficacy by depleting the TIGIT positive cells. Again, difficult to compare across trials. The safety profile looks differentiated compared to the non-active FC TIGIT antibodies. We're moving into large trials. If you have patients with high rate of discontinuation that are staying less time on the trial, obviously, it's more challenging to achieve this kind of overall survival benefits in large trials. We always said that the right choice, from our perspective, for a TIGIT antibody, like the PD-1 antibodies, by the way, is to block it without any other manipulation, to use a non-active FC and to block TIGIT without any undesired side effects or maybe even positive outcomes. We thought that eventually, you will not have a positive outcome by manipulating the FC receptor, the FC receptor binding capacity of the antibody. We are one of the few companies that have such an antibody with reduced binding to FC receptors, but also some others. I think these are Arcus-Gilead and AstraZeneca with our own TIGIT PD-1 bispecific antibody. They are now ongoing in Phase 3 trials. We are really eager to see, and we think that to see better outcomes with the TIGIT assets from our class, the non-FC active. Obviously, being one of the few companies out there with a non-active FC TIGIT antibody, these kind of results in big trial and Phase 3 trial are going to probably bring back interest into the TIGIT landscape and specifically in COM902. Now, COM902 is differentiated not only by being a potential best-in-class and an FC reduced binding antibody, but also due to our clinical strategy. We know from others that PD-1 and TIGIT probably show activity mostly in inflamed settings, in non-small cell lung cancer, in gastric cancer. If there is sufficient amount of T cells and inflammation, then PD1 and TIGIT blockers combined show activity. Our strategy, actually, is to combine it with the PVRIG blocker COM701. What we have shown, initial signals clinically and practically quite extensively, is that this unique biology of PVRIG enables us to use COM701 to potentially open less inflamed tumor types to TIGIT PD-1 blockade. COM701 will sensitize the tumor environment, increase T cells, also in less inflamed tumor types. The triplet could have activity also in places that until now we didn't see activity from TIGIT and PD-1 blockers. As mentioned before, AstraZeneca have licensed the rights to use COM902 as part of the TIGIT PD-1 bispecific. They are moving now into—they have now eight Phase 3 trials open across different indications. They reported single-arm studies results that look promising. The activity of these assets looks definitely favorable compared to historical control of PD-1 blockers alone. What is also interesting to mention is that, first of all, as mentioned before, we think that TIGIT is active, and especially AstraZeneca, which also chose a non-active FC. We expect them to see activity even compared head-to-head versus Pembrolizumab as they do in some of these trials. Having a bispecific antibody, actually, the contribution of components that might be asked by regulators is less of a concern. In some of the trials, let's look at TROPION-Lung10, for example. If you combine Rilvegostomig, the PD-1 TIGIT bispecific, with a TROP2 antibody-drug conjugate and compare to Pembrolizumab, if the risk-benefit profile of the combination of these two agents is better than Pembrolizumab alone, this trial could get a win or an approval even in the absence of directly proving that TIGIT is active. As mentioned, we think that TIGIT is active, and we hope to see success upon multiple of these trials. Even if the activity will not translate into direct ability to beat Pembrolizumab, some of these trials could have a win even in the absence of that. We have already received $30 million milestone from this collaboration. We're eligible for additional $170 million. More importantly, maybe the fact that we are eligible also for mid-single-digit royalties. If this drug will be indeed eventually a blockbuster, this will be a great win for Compugen as well and will enable us to further support our pipeline using the income coming from this collaboration. Now, a more recent collaboration. Using our computational platform, we're looking at the tumor environment for resistance mechanism by tumor-associated macrophages. We're not looking for cytokines at all. What we identified there took us into the world of cytokines. As you may know, cytokines, I mean, they're approved for 30 years, but people are really trying to manipulate them and engineer them and to overcome the inherent toxicity of cytokines and the challenging therapeutic window. Until now, with not much of success, and there are still clinical trials ongoing, so we'll see. The observation we made computationally was that in the tumor environment, actually, you already have a cytokine, which is called IL-10. It's induced by inflammasome. It's upregulated in tumor environments. It's low in the periphery. Actually, you have a cytokine there that should be active by itself. This cytokine is naturally inhibited with another cytokine or a soluble molecule called IL-18 binding protein that binds this cytokine IL-10 and prevents its activity. You have this complex floating in the tumor environment in an inactive form. We developed an antibody non-engineering. It's a very good antibody, high affinity antibody, but just a clever approach, which is not sophisticated with all these difficulties of engineering. We use a blocker antibody, a drug moiety, to inhibit the inhibitor to unleash IL-10 for its natural activity. What we have shown preclinically is that this really seems to overcome many of the challenges that the cytokine therapeutics has. We saw preclinically no systemic side effects. We have seen a really strong, potent activity only in the tumor environment. We licensed this program to Gilead. We initially received $60 million upfront on this asset. More recently, when we submitted the IND, we received additional $30 million. A total of $90 million to date. We are now starting the clinical trials. We started with those the first patient of the dose escalation in the beginning of this year. We are eligible for additional $758 million milestones and up to low double-digit royalties. What's also interesting is that even though Gilead licensed the assets, we're the ones moving the asset into phase one. It's kind of a unique collaboration approach. We have the know-how. We have the capabilities. Luckily, Gilead team thought that our clinical development track is sufficient for them to trust us and to enable us to put their asset in our hands to continue into the Phase 1. We are now starting. As we said, we dosed the first patients. The part one of the study is dose escalation in monotherapy and in combination with ZIM, the anti-PD-1. In the part two, we're going to do expansion and specific indications based on some preclinical data and also, obviously, based on what we'll see in the clinical study itself, again, in monotherapy and in combination. This is a very good collaboration that is going on. Obviously, in the next advanced stages, Gilead will take this asset. This is the first time we're doing this kind of collaboration. We think there's a very strong potential for synergism, us taking it in the first stages and then Gilead moving in the advanced stages into the advanced clinical trials. As mentioned before, we also have other assets. We have a computational platform that continues to feed the pipeline. For now, we're not disclosing the name of these assets, but definitely, the future of Compugen will be bringing more and more assets like COM701, like COM503 to clinical testing and hopefully eventually to patients. By that, I can stop and take any questions Ethan you may have. Great. Yeah, no, thank you for the presentation, Eran. As a reminder, please feel free to use the question feature to submit any questions. I'll be keeping an eye on the live feed there. Maybe to kick things off, and I know you touched on it during the presentation, but TIGIT, as you mentioned, obviously, a lot of, I guess, disappointment in general, more broadly, obviously, FC active versus FC inactive is a lot of caveats. Do you find that investors are receptive when you kind of explain where things may have went wrong and why your program's different? Or do you think that's still a pretty high bar to hurdle right now and that they're just waiting to see better data at this point? Yeah, definitely waiting to see data. I think that I totally understand that. I mean, after all these phase three failures, it's difficult to expect that the investors will have any trust in TIGIT. I think that one needs to look into the fine details as in many things. Yes, the data does show that the FC non-active have better safety, have much less discontinuation rates. About efficacy, difficult to say without doing Phase 3 trials. I think that eventually, and it's maybe not a coincidence that most of the companies who dropped their TIGIT antibody are the ones with the FC active and the ones with the FC non-active are still running in Phase 3 trials. Yes. Also, we need to remember that this also reflects—I don't think it should, but it does reflect also in PVRIG because it is related to TIGIT in some way. We've showed quite extensively it's very different from TIGIT, being active in other indications. Eventually, from an investor perspective, it's also, I think, underestimated. We'll work hard to generate the data and convince. Yeah, yeah, that's all you can do. Kind of another sort of TIGIT, or I guess more related to your bispecific, there's been a lot of excitement recently behind a slightly different mechanism bispecific, which is like the PD-1, PD-L1, VEGF bi-specifics. I want to see, first of all, if you had any thoughts more broadly on that space. And then if you think that success maybe in that area will help bring more interest to other checkpoint bispecifics. Now, I know PD-1, TIGIT, it's a different mechanism, but curious if you have any thoughts there. I think we have to wait and see. I mean, PD-1 and VEGF, by principle, mechanistically definitely have synergistic potential. This didn't translate enough, I would say, with Atezo and BEV, for example. Yes, the data looks not bad. We'll have to see if this translates into larger clinical trials, if the safety is indeed as good, because we need to see if the PD-1 localization indeed is sufficient to solve the safety issue of BEV. We'll have to wait and see. I think comparing it again across trials to rilvegostomig, for example, I think the efficacy looks quite similar. Indeed, rilvegostomig was not yet tested in randomized study, but just overall response rate in similar settings seems similar. I think that eventually, mechanistically, I believe that the PD-1 TIGIT bi-specific will be safer. All this strategy of AstraZeneca, using it as an IO backbone, combining with all the ADCs, I believe that this will be a better asset for combination. Since IO and all oncologies are going into combinations, I think that we really have to see the safety of the PD-1 VEGF to be convinced they could serve the same role. Yeah, no, that makes a lot of sense. Maybe switching over to one of the last programs you mentioned there, your IL-18 or anti-IL-18. Is there any interest there in studying that program in hematological malignancies? Yeah, it's an interesting question. We didn't look much into hematological, mainly because our observation is that in the solid tumor environment, because of inflammasome activation and cell death and all of that, there is a really high amount of IL-10, which is naturally there, blocked by IL-10BP. Therefore, the opportunity to use COM503 to block it, displace it, and have this unique activity we have shown, we didn't study it. But I'm not sure that in hematological settings, you will have the same type of tumor environment with inflammasome and IL-10 naturally there. Maybe for recombinant cytokines, IL-10 or CAR T cell secreting IL-10, this could make a lot of sense. For our program, we didn't test it much, but we're starting with the solid tumors due to that rationale. Yeah, no, that makes sense. One aspect which doing a little research noticed, you have this collaboration with Ultima Genomics. I wonder if you could maybe talk a little bit about that collaboration and how you're leveraging that technology to enhance your own platform. Yes, thanks. Maybe a few words about our platform. We have a computational platform called Unigen, in which we use different tools, including AI-based tools, to identify novel drug targets. This is quite unique. I mean, people are using AI and computational to do different aspects of R&D. But we are one of the few companies which are focusing on the first stage of really bringing new drug targets to immune oncology. And for sure, we're one of the very few companies which have a validated platform because we're able to do it again and again, to use a computational platform, bring novel drug targets, collaborate with pharmas, take it into the clinic. So in that aspect, our platform is really unique in all this AI-dense landscape. Specifically about that collaboration, Ultima Genomics have a great sequencing technology. And we collaborate with them to sequence patient tumor samples mostly. And by that, we feed our own database, which is the base for every AI is having a good, large, strong database with data of patient samples. We are using Ultima to collaborate on that. Also, we identified that using their technology, and this is what we published in a recent conference last month, we could identify at the single-cell level specific alternative splicing events. We are using our algorithms to employ their technology to use these kinds of approaches to identify novel drug targets. We are both feeding our pipeline and using the technology for specific queries, looking at novel potential targets in immune oncology. Very interesting. Like you said, your platform definitely drives the company. There is a lot there. Maybe one more for me, and I will keep an eye on the live feed as well. You are going after platinum-sensitive cancer in the maintenance setting. Obviously, refractory has been more or PROC has been kind of more of the popular place to study drugs these days just because high unmet need. It sounds like there is a very significant high unmet need in the sensitive setting as well. Why do you think more companies either aren't going after sensitive or have been unsuccessful so far? Is it just a safety therapeutic window, or what do you think? First of all, I think that every new drug normally goes into the last-line patient. This is where you start. You want to look for your signals. This is what we did as well. After substantiating the signal in the last-line settings, we had some patients in monotherapy, including responses in monotherapy and then the combinations. Moving into the earlier platinum-sensitive made a lot of sense. It seems that there is an unmet need. There were a few attempts, maybe not in this exact population that we identified, but there are a few attempts for maintenance settings with checkpoints. I mean, also other thought is a good idea and with not much success to date. As mentioned, we think that PVRIG blockade should and could and should be different. This is because of the really unique biology of PVRIG compared to all the other checkpoints, which there is a biological reasoning why it could be active in ovarian cancer in general. Also because we've seen the signal. The signal seems to show differentiated activity. Having a safety profile of a checkpoint, but your unique activity in these settings should enable us or could enable us to show activity in the platinum-sensitive settings. As mentioned, the ADCs, which are dominating or starting to dominate the PROC settings, are trying to move now also in the PSOC, the platinum-sensitive. Again, most of the studies, and probably rightly so, we still talk about the toxic agents, are aimed to replace the platinum or the chemotherapy. Yeah, in the maintenance, some of the PUA attempts for PARP after PARP, BEV after PARP, and all these attempts, but there's definitely an unmet need of maintenance. I think that we have a really interesting potential for COM701 there. No, and the adaptive study design and where you could start monotherapy and then work your way to combination if necessary. I think it's definitely a unique way to address the issue. I think taking a look at the chat, I don't see any further questions at this time, but I just want to thank you for coming on today and taking the time to talk a little bit more about Compugen. And yeah, thank you for attending our conference. Thank you.