Hello everyone, my name is Chris, and I'll be your conference operator today. At this time, I'd like to welcome everyone to Exscientia's business update call for the Q1 2023. All lines have been placed on mute to prevent any background noise. After the speaker's remarks, there will be a question and answer session. If you'd like to ask a question during this time, simply press * then the number 1 on your telephone keypad. To withdraw your question, please press * 1 again. At this time, I'd like to introduce Sara Sherman, Vice President of Investor Relations. Sara, you may begin. Thank you, operator. A press release in 6-K were issued this morning with our Q1 2023 financial results and business update. These documents can be found on our website at www.investors.exscientia.ai, along with the presentation for today's webcast. Before we begin, I'd like to remind you that we may make forward-looking statements on our call. These may include statements about our projected growth, revenue, business models, preclinical and clinical results, and business performance. Actual results may differ materially from those indicated by these statements. Unless required by law, Exscientia does not undertake any obligation to update these statements regarding the future or to confirm these statements in relation to actual results. On today's call, I'm joined by Andrew Hopkins, Chief Executive Officer, David Hallett, Chief Scientific Officer, and Ben Taylor, CFO and Chief Strategy Officer. Garry Pairaudeau, Chief Technology Officer, and Mike Krams, Chief Quantitative Medicine Officer, will also be available for the Q&A session. With that, I will now turn the call over to Andrew. Thank you, Sara. Today we're going to talk about a differentiated approach to personalized medicine, how we use complex primary patient tissue samples as preclinical models. Combining this with our in-house multi-omics capabilities, we can go from target identification all the way through to the clinic. 2023 is off to an exciting start as we continue to advance our pipeline and strengthen our business. We've made significant progress across our internal and partner programs, including advancing 2 molecules into the clinic, EXS4318 and EXS21546. An additional molecule, DSP-2342, was advanced by Sumitomo Pharma, which was a result of an early collaboration with Exscientia but is now complete. This marks our 6th novel molecule created through Exscientia's generative AI platform to enter the clinical stage. We've expanded our precision oncology pipeline by initiating IND-enabling programs for EXS74539, an LSD1 inhibitor, and EXS73565, a MALT1 protease inhibitor. More recently, we presented multiple posters at the AACR annual meeting, highlighting research that continues to validate our end-to-end approach and demonstrates the potential of a platform to rapidly advance high-quality drug candidates towards the clinic. Our team's commitment to strong execution has enabled us to rapidly move programs from discovery through to the clinic. We have achieved a number of milestones already this year. In March, we announced 2 new wholly owned precision design molecules, an LSD1 inhibitor, 539, and a MALT1 inhibitor, 565. Both programs continued to progress through IND-enabling studies. We expect to provide an update on clinical development plans in the H2 of this year. We remain on track to meet our target of 4 candidates with meaningful economics for Exscientia in clinical development by 2024. In February, Bristol Myers Squibb initiated the 1st in-human study of EXS4318, our potential first-in-class selective PKC theta inhibitor. 4318 was designed by Exscientia and is currently in phase I clinical trials in the United States. Earlier this month, the 1st patient was dosed in IGNITE-AI, our phase I/II trial evaluating EXS21546 or 546, our A2A receptor antagonist. This was the 1st AI-designed immune oncology drug in the clinic. We remain on track to dose the 1st patient in a phase I/II study of GTAEXS-617, our precision design CDK7 inhibitor, co-owned with GT Apeiron in the coming weeks. We also remain well-capitalized with $553 million in cash at the end of the quarter. This provides us with several years runway to advance our near-term programs without the need to raise external capital. On today's call, we'd like to provide more detail on our approach of combining precision design with personalized medicine. Before handing over to David Hallett, our CSO, I want to highlight a recent scientific present at this year's AACR meeting. We presented data further validating our ability to efficiently design high-quality drug candidates and to identify and predict the right patient populations that may benefit the most from treatment. Firstly, for 546, we presented research on our adenosine burden score or ABS. It showed that 546 reverses the effect of adenosine analogues ex vivo in patient tissue samples and other complex models. The ABS has been validated in our ongoing IGNITE-AI phase I/II clinical trial of 546 and will be discussed further today. IGNITE-AI was designed based on extensive simulations to enable the most effective Continuous Reassessment Method settings to predict and accurately evaluate the anti-tumor oval effects of 546 in combination with checkpoint inhibition. The team also presented preclinical data on EXS74539, our precision-designed LSD1 inhibitor. We designed 539 to optimally target LSD1 in future oncology and hematological patient populations. These preclinical data demonstrate that 539 has the potential to overcome significant safety limitations of other LSD1 inhibitors through its differentiated profile, combining reversibility and brain penetrance. Lastly, we highlighted the benefits of using data generated with Exscientia's precision medicine platform in combination with its proprietary methodology for multi-omics and multi-modal mapping. By better understanding disease mechanisms, these tools combined can be leveraged to improve patient outcomes by uncovering clinically relevant drug targets already at the discovery stage. We'll go into more depth in this topic shortly. In summary, we have 5 programs of economics that are either in the clinic or in IND-enabling studies, all are a testament to the power of our platform and our approach. We are thrilled in our recent advances and look forward to sharing more details of our clinical development plans in the H2 of 2023. Today, we would like to focus on how we're advancing towards our goal of increasing the probability of success within drug discovery and development through an end-to-end patient-centric approach. In our pipeline to date, we have developed precision design compounds with a patient-driven data approach in a faster and more efficient way than existing methods. I'll now hand over to Dave to walk through how we are working towards predicting clinical responses pre-clinically. Thank you, Andrew. We incorporate the concepts of patient-centric drug discovery and development as early as possible in our efforts. Through the use of complex primary patient tissue samples as preclinical models, we are able to leverage our clinically predictive functional imaging platform, especially in translational research. While cell lines and organoid models are scalable and useful in design and development, they do not capture the complexity of actual disease biology, nor do they represent the diversity of patients seen in the clinic. As you can see here on slide 6, there is a clear difference in the images of the homogeneous cell line compared to the heterogeneous primary patient material we use. We believe that the heavy use of cell lines as translational models has contributed to the high rate of clinical failure we typically see in our industry. Our answer is to strategically leverage primary patient material for decision-making purposes before entering the clinic. By getting as close to the actual patient as possible, we can embrace both the heterogeneity and complexity of disease biology using our patient-derived model systems coupled with AI-driven technology. In our preclinical studies, we utilize primary material to create complex model systems that better reflect disease and represent patient diversity. These elaborate models are deployed with the goal of identifying indications as well as subpopulations likely to respond to treatment, uncovering patient enrichment and non-invasive pharmacodynamic biomarkers, understanding the potential for resistance, combination effects, and more. Depending on the program, we take advantage of our precision medicine platform, which has successfully predicted which drugs will work for a given patient, as shown in the EXALT-1 study published in Cancer Discovery in 2021. Functional endpoints in our complex systems allow us to simultaneously quantify what a drug or combination of drugs is doing to cancer, immune, and non-transformed cells at the single-cell level. We can measure anything from cell size to cell death, through to pathway activity, depending on what we want to quantify. We then combine this functional data with omics readouts from the same patient samples, such as genetic mutations, expression, fusion, and transcriptional events. The omics data provides a molecular understanding of the observed phenotypes. The union of technologies, functional and multi-omics, combined with years of knowledge of how to interpret these datasets in multi-modal programs, drives a deep understanding of disease biology and population heterogeneity. Exscientia's unique proposition is that these data are derived from primary patient samples. This provides a preclinical understanding of how and why a drug is, or just as importantly, is not working in a given patient sample, thus enabling patient enrichment hypothesis generation and the generation of molecular signatures. Today, we will describe 2 ways in which we are combining the use of our functional precision medicine platform with our omics datasets. 1 to gain an understanding of the effect of adenosine on the cancer microenvironment ahead of a clinical trial in patients, and the other for target discovery. We'll first highlight progress for our A2A receptor antagonist, 546, which specifically blocks the recognition of adenosine by immune cells within the cancer microenvironment. Adenosine is an immunosuppressive metabolite produced at high levels within the tumor microenvironment. Adenosine limits the functionality of multiple protective immune infiltrates, including T cells, while enhancing the activity of immunosuppressive cell types. Reversing the effects of adenosine driven through the A2A receptor with our antagonist 546, should therefore release the immune system and also help those patients who have become refractory to immune checkpoint inhibition. For patients to benefit from such an approach, 2 critical attributes are required to be present. 1, high levels of adenosine in the microenvironment. 2, an immune system primed but suppressed by adenosine. To date, there has been no robust way to measure both immune potential and adenosine levels within the tumor microenvironment. We believe other drug candidates for this target have not achieved clinical success because they failed to enrich for those patients most likely to respond to A2A receptor pathway inhibition. Leveraging our precision medicine platform and scalable in-house omics capabilities, we have identified a patient enrichment biomarker that correlates with adenosine levels in the tumor microenvironment. We call this the adenosine burden score, or ABS. This was found through a detailed examination of multiple primary samples at baseline and after perturbation with adenosine pathway activation. All this work has been done in an effort to maximize the probability of success of 5-4-6 in the clinic. On this slide, we show 3 different data sets, 2 from human databases and 1 from mouse data. These include The Cancer Genome Atlas, or TCGA, and the Reactome database. TCGA is a landmark cancer genomics program from the National Cancer Institute and National Human Genome Research Institute that characterize, at a molecular level, over 20,000 primary cancer and matched normal samples spanning 33 cancer types. Reactome is an expertly curated database of biological pathways. At the top in the TCGA dataset when filtering for patients with a high ABS, we observe that these same patient samples are low for public signatures related to inflammation, such as the Tumor Inflammation Signature or TIS. The TIS has been used to predict anti-PD-1 efficacy. In the middle panel from the Reactome dataset, the ABS anticorrelates with the PD-1 signaling pathway, indicating that where adenosine is high, as measured by the ABS, PD-1 signaling is low, thereby nullifying anti-PD-1 effects. The last chart is an expert-curated mouse dataset called TISMO, or Tumor Immune Syngeneic Mouse dataset. This shows that mice considered resistant to checkpoint inhibitor therapy were also enriched for higher mouse ABS, highlighting the rationale for combination therapy in our 5-4-6 clinical trial. Taken together, we believe we have discovered a robust, specific, and sensitive biomarker for adenosine pathway activation within the tumor microenvironment. This represents a method for enriching patients likely to respond to our selective A2A receptor antagonist, EXS21546. Comparing the left and right panels, we can see that compared to other disclosed signatures, ours is much more robust and reproducible across samples. Our signature is comprised mainly of B-cell genes towards the later stages of B-cell and plasma cell maturation. Similar to that of data from another molecule recently presented at AACR that was discovered retrospectively after a phase I-B clinical trial. Our work was done pre-clinically and will be validated alongside the IGNITE trial. What we have shown here is that we can generate data ahead of clinical trials using primary patient samples that our peers can only do in the clinical setting. We believe this is a key differentiator for Exscientia as we advance additional programs and have implications well beyond our A2A program. Since our founding, we have aimed to be a learning company with a goal to constantly increase our knowledge from and to reuse all of the data that we produce from discovery through to development. We've just shown you an example of how we can pre-clinically identify patient enrichment biomarker hypotheses using a combination of functional and omics data. I'll now take a moment to highlight how we leverage this same approach in our discovery efforts to understand more about disease biology and target discovery. Using the data sets from preclinical studies, which will be supplemented with information from our clinical and precision medicine studies when available, we can work to understand a disease computationally. I will highlight how we use functional and multi-omic data from our primary models to help identify novel targets and druggable pathways for future projects, some of which we believe may help overcome resistance. Here we show an overview of some of the data inputs we use to triangulate and prioritize novel targets. We start with our proprietary data from various programs that take advantage of our functional precision medicine platform and next-generation sequencing unit. All of this data is from patient tissue models and this differentiates our approach from others. We then combine this with well-annotated public data such as known drug to target annotations, taking into account a drug's polypharmacology and protein-protein interactions in a custom unified and extensible computational framework. While the use cases of a program that captures the complexity of a disease in silico are vast, the example I want to describe today is focused on target identification. Our patient-centric multi-omic platform has the potential to uncover targets with high clinical relevance at the discovery stage, as well as support target validation and biomarker discovery. At the bottom of the slide, we see our functional layer of data, target annotations, and the interactome come together to prioritize targets. Using drug sensitivity and protein-protein interactions as a guide to identify convergent targets. Here we put everything together. I want to first show you a diagram of how this data is represented. We use our precision medicine platform to collect functional and multi-omics data from patient tissues, in combination with proprietary methodology for multi-omic and multimodal dataset mapping. We integrate it using our computational framework. The outer layer represents the standard of care drugs we use as tools to probe the potential target landscape. Drugs are connected to their known targets, including off targets on the next layer. Known targets are embedded in a curated protein-protein interaction network, allowing us to identify novel targets at the focal points of successful therapies. More than that, we are also able to corroborate and refine our findings using a rich layer of multi-omics data, such as phosphoproteomics and single-cell RNA-seq, generated under treatment conditions from the same samples. This approach has the potential to uncover targets with high clinical relevance at the discovery stage and lead to improved patient outcomes. What you see here is an example of functional screen performed in 20 ovarian cancer patient tissue samples. We wanted to understand the cancer-specific cytotoxic effect of drugs with well-annotated targets. You may recognize this data from one of our recent AACR posters. On the left, we have identified numerous novel sensitivities to a subset of tyrosine kinase inhibitors or TKIs, signified by large dark purple circles within a subset of samples. What's important to appreciate here is that the effects we observe for many drugs in patient tissues, the left panel, are not recapitulated in publicly available cell line sensitivity data indicated on the right. This demonstrates how the use of cell lines and other cultured model systems may obscure targetable pathways. This is likely due to oversimplification of tumor biology, since the cell lines lack a complex and diverse cancer environment. Instead, our primary model system incorporates multiple cell types and avoids immortalization or amplification in order to better capture the complex biology of the original microenvironment. What this does not yet tell us is why specific drugs are having an effect and what they have in common, complicated by the fact that many of them have known polypharmacologies. Overlaying our unique functional endpoints with multi-omics data, we use drugs as tools while also mapping sensitive and insensitive pathways across multiple molecular layers and begin to reveal novel biology and target spaces. Here we show the actual data with the targets blinded. First, we use network integration of patient tissue functional data to triangulate convergent targets. We add a layer of data from multi-omics measurements that lets us further prioritize them by factors such as disease-specific expression, mutation profiles, or novelty. The diagram from outer to inner circle shows firstly, global compound sensitivities, then known primary targets, and finally predicted downstream targets. These targets are not impacted by community bias, highlighting first-in-class potential. Keep in mind, this is data from real patient samples, grounding us in complex human biology. This means that we can combine real-time multi-omics data with the functional biology readouts to directly measure drug response from multiple angles on every sample. This helps us identify novel targets with demonstrated biological activity that we would not have been able to find with more artificial models or database screening. We already have some targets identified from this approach, going through tractability and validation internally, and we look forward to keeping you updated on our truly differentiated platform. As I mentioned earlier, Exscientia is a learning company, not just in practice, but also through the reuse and redeployment of collected disease modeling datasets. Here we use functional profiling as a guide to build computational disease models for target ID. We are also working to redeploy data for target validation, faster patient enrichment, biomarker discovery, and combination prediction. We've provided examples here on how complex disease-relevant models, combined with a smart analysis and interpretation of many levels of big data, can reveal mechanisms of adenosine pathway activation for us to identify patients that may be sensitive to 546 treatment. We're also working on predicting combinations and identifying resistance-breaking characteristics for our CDK7 inhibitor, 617. We plan to present 617 data towards the end of this year, and we'll be adding more data to these models as our pipeline grows and as we recruit patients into our clinical studies. With that, I will now turn the call over to Ben to walk through financial highlights. Thank you, Dave. I'll now take a minute to close with highlights from our financial results. Full results are detailed in our press release in Form 6-K. I'll review the results in US dollars using the March 31st, 2023 constant currency rate of GBP 1.24 to the pound. We ended the quarter with $553.3 million in cash, equivalents, and bank deposits. We believe this provides us with several years of cash runway and the resources to continue investing in our growth. As Andrew noted earlier, we continue to successfully advance our internal and partnered projects. At the same time, we have also been executing cost efficiency programs that are expected to save over $20 million during the course of 2023 and more in 2024. This has been a combination of automation through technology and narrowing the focus of our operations on core activities that have a differentiated commercial profile. We remain cautious in the current macroeconomic environment and intend to continue our cost control efforts through the end of the year with a focus on optimizing workflows and automation. We have a robust business development dialogue and maintain our guidance of 2 new deals this year. Earlier in the year, many of the large pharma had substantially slowed their decision-making process for new partnerships as they conducted pipeline reductions and budget cuts in response to the IRA and other well-noted industry trends. Recently, we have seen a renewed energy and excitement from our potential partners, especially in our core technologies such as personalized medicine and generative AI. It is important to note that we have never stopped investing in new technologies. While we are being intelligent about burn rate, we continue to see substantial technology advancements even on a quarter-to-quarter basis. Dave discussed how we had taken a strong phenotypic translational platform and invested to add multimodal data that now can produce personalized cellular signatures at every stage of discovery and development. This is only 1 example of our growth. We have over 200 people in our technology group focused on improving the capabilities and predictive powering of our AI across the company. This is how we intend to stay in our current leadership position. With that, I will turn the call back over to Andrew. Thank you, Ben. During our presentation today, we've highlighted the progress of our clinical and preclinical programs. We are bringing new molecules into the clinic and building out our AI-powered precision medicine platform. We are confident that our differentiated tech-enabled approach will yield strong outcomes. To finish, let me add just how proud I am to lead a global team this talented and determined, who help us do everything in our power to deliver on Exscientia's promise to transform the way the industry discovers and develops effective medicines and to deliver the best possible outcomes for as many people as possible around the world. We'll open up the call for questions. Thank you. As a reminder, if you would like to ask a question, please press * then 1 on your telephone keypad. Our 1st question is from Alec Stranahan with Bank of America. Your line is open. Hi, guys. Thanks for taking our questions. I have 2 higher-level 1's. I saw an interesting quote, I think from Gary, that by the end of this decade, design of all new drug candidates will be augmented by AI. What do you see as being the key points that need to be addressed today for this future to be realized either at the basic science level programming or regulatory levels? As a follow-up to that, maybe for Andrew, how does a company such as drive the most value for shareholders if this is the direction that the industry is going? Is it through more design as a service such as your collaboration with Sumitomo or driving pipeline assets through approval and commercialization yourself? Any directional commentary would be helpful. Thanks. Thank you so much for your excellent questions, Alex. Really great actually, and very topical point as well. Actually, for the 1st question, as you did actually direct that to Gary, I'm actually gonna have Gary to outline as CTO what he sees actually as sort of the key further challenges to really expand AI's use in pharma for all drugs eventually to be designed by AI. Gary. Cool. Yeah, thanks. Thanks, Andrew. I think, I mean, the first thing is we're incredibly proud at Exscientia that we've now enabled 6 clinical candidates using AI. That, you know, kind of really shows the promise and the power. You've only got to pick up a newspaper or look anywhere really to see how the entire world and the entire world of drug discovery is starting to embrace the use of artificial intelligence and broader computational methods. I think there is a natural evolution. I think for us, what's really important to us is how do we stay at the forefront of that. I think the activities that Exscientia is building out at the moment, particularly in linking AI design to physical automation, robotics synthesis, robotic screening, is really closing the cycle and enabling us to drive our projects even more quickly in the future. I think it's developments like this that are gonna enable more broad acceptance and utilization of these kind of technologies in drug discovery. Let's be honest, it has to be a fantastic thing, doesn't it? You really want to bring medicines to patients faster and more effectively as we're demonstrating technology can do. Thank you, Garry. Really, I want to underline Garry's answer actually in how we think about things. To answer your second part of the question, Alec, the way we think about it is that we're incredibly pleased to see that sort of our design prowess now and bring in 6 molecules have used generative AI approaches now into the clinic. As you said, actually, the latest one actually being with Dainippon and Sumitomo Pharma, which was with an earlier business model called Design as a Service. We're always open to doing, you know, many kinds of deal structures, as you've seen, actually, I think our business development prowess for the past few years has actually shown that. The way we see that AI is gonna create real value is to think about what that product of the future looks like, what that sort of AI-enabled drug starts to look like. What we see as the hallmark of an Exscientia drug is a drug that uses advanced compute, machine learning, AI, and physics-based methods to design, precision design a high-quality molecule. Also then use in our deep learning multimodal approaches that Dave was talking about earlier to really define the patient selection strategy. Bringing those 2 together in a model-driven adaptive learning approach to learn about the drug, that's what we see as those 2 pieces of key IP, the molecule being designed by AI and using AI then to design the biomarker. Those coming together is what we think is the hallmark of Exscientia drug, and that's where we believe in the long term, the high value wealth can be created by effectively creating highly effective medicines with high response by actually designing the best molecule and targeting the right patients. Great. Thank you. The next question is from Vikram Purohit with Morgan Stanley. Your line is open. Hi. Thanks for taking our question. This is Steve for Vikram. I wanna ask about the A2A program. Could you discuss the prior treatment history for the patient you are enrolling into the trial? When can we expect to see the initial data, and what's your expectation about the readout? Thank you. Thank you very much, Steve. For that question, actually, I want to hand the stage over to Mike Krams, our Chief Quantitative Medicine Officer, who's actually leading our clinical development work here. Mike. Yeah. Thank you very much for the question. We have recruited our 1st patient into this program. It's a phase I and II study. We use simulation-guided clinical trial design to come up with an approach where we initially have a dose escalation, aiming to make the correct decision at the earliest time point as to what the dose and treatment regimen is that we will take into a dose expansion phase. We're going to learn about the operating characteristics of the investigational compound. At the same time, we are qualifying the adenosine burden score, as Andrew pointed out, as our tool to identify which are the correct patients who might benefit from an A2A receptor antagonist in conjunction with a checkpoint inhibitor. As to when, data will become available, this is a phase I/II study in early development and oncology as many others, so it's really very similar to other programs, and we are going to provide further guidance, as time progresses. Thank you. That's star 1 if you'd like to ask a question. The next question is from Peter Lawson with Barclays. Your line is open. Peter Lawson with Barclays, your line is open. Please go ahead. Okay, we will move on to the next question, which is from Chris Shibutani with Goldman Sachs. Your line is open. Hi. It's Roger on for Chris. Just a quick question on 565, the MALT1 inhibitor. You're likely aware that J&J, they debuted their phase I data for their MALT1 inhibitor in NHL and SLL. I was just wondering if you'd comment a little bit on the inhibition of UGT1A1 and, you know, where do you expect 565 to come out in terms of differentiation, noting the competitive landscape. Thanks. Thank you much, Roger. It's a great question, actually. It has been a key point of how we have been designing a differentiated molecule. I'm actually gonna hand this sort of question over to David Hallett, our Chief Scientific Officer, to give you some more color on it. Thank you, Andrew. And thank you for the question. I think the publication of that abstract I think is coming out ahead of a European oncology symposium was very timely. If you recollect the information that we put out very recently around the design criteria around our MALT1 inhibitor, and specifically the topic of hyperbilirubinemia and driven by inhibition of UGT1A1. If you remember the takeaway story from that is that we strongly believe that our molecule is differentiated from J&J and most likely quite a few other competitor molecules, and that it has little to no activity at that particular transporter, and is therefore unlikely to kind of to drive that particular side effect. If you actually look into, even into the abstract details, it's pretty apparent from J&J as we would have predicted, that they do see hyperbilirubinemia in the clinic. They've had to take account of that in their, with their recommended phase II dose. I'm sure they would have preferred not to have done that. So I think we stand by. I think that original assertion is that was a really important differentiation criteria. I think it will, our molecule we believe, should be, should be free of that particular potential toxicity. More importantly, as I think as we highlighted, is that we must remember it's very likely that a MALT1 inhibitor will be used in combination with other agents like BTK inhibitors, and therefore you need a, as clean as possible a safety profile so that you could dose that molecule as high as possible. No, it was. I think I wish J&J well. I think, and obviously, as they take that compound forward into patient studies. I think it's, I think it supported our notion about the differentiation angle of our own compound. Thank you. The next question is from Peter Lawson with Barclays. Your line is open. Hi, this is Shay on for Peter. Thanks so much for taking our question. Just wanna touch base on the biologic side of your platform and maybe some progress there and how you're thinking about balancing your biologics versus small molecule development and maybe even when we could see the first antibody program going into the clinic. Thanks so much. Excellent. Thank you very much. I'm going to hand over this question actually to Gary, whose team has the algorithms for developing sort of biologics by design, by discovery are currently being developed. Gary. Yeah, thanks. Thanks for the question. We're, I mean, we're really excited about the way that we can introduce biologics into our AI design platform. Professor Charlotte Deane has been working to build out the algorithms and all the technology to actually drive that forward. We're still at the point where we're developing a robust process, and we're starting to run our 1st pilot project. I think we're a little bit away from talking about a molecule in the clinic right now. What I can tell you is we are developing, actually, I'd say world-leading capabilities in the areas of predicting structure and being able to do generative design into the antibody space. In terms of growing the pipeline, we certainly are now looking to think about how we might bring forward sort of our first programs going out and actually how then we start to map then the antibodies, the capabilities we've then built and actually to sort of our key therapeutic area sort of focus. 1 exciting thing is that we've already demonstrated is that our precision medicine platform actually also works with antibodies as well as small molecules. That's a key thing then, because it allows us then to think about how then as we head towards the clinic, we can also bring to bear precision medicine technology. I think that's gonna bring a unique differentiator as well, actually, in this particular field for these modalities. Great. Thank you. The next question is from Steve with Morgan Stanley. Your line is open. Good morning, everyone. This is Gospel, on for Vikram. I have a question regarding your PKC program. For the PKC theta program in partnership with BMS, I was wondering how much visibility and control do you have now into the path forward for this molecule and how it might progress through early-stage development. Thank you. This is David Hallett. Thank you for that question. In terms of public visibility, because BMS in-licensed that particular program, they both now control the clinical development of that project, but also obviously kind of public disclosures that are related to that. As a trusted partner and part of the JSC, we will receive kind of updates on that program ourselves. Just to reiterate to everyone who's on the call is that that particular asset has begun a healthy human volunteer study in the United States in the early part of this year. We look forward to kind of receiving updates from BMS as they progress. Thank you very much. We have no further questions at this time. We'll turn it back to the presenters for any closing remarks. Thank you, Chris. As Exscientia CEO and founder, I am proud to see our company maturing into an end-to-end precision medicines business, spanning from discovery into early development and supported at each stage by our innovative technology platforms. Our goal is to be as innovative in the clinic as we have been in discovery. Our remarkable progress to date is a testament to the strength of the company. Thank you to everyone today on the call for your continued support and on our journey and for joining us today. We look forward to continuing to share our progress with you throughout the year. Operator, you may now disconnect. Thank you. Ladies and gentlemen, this concludes today's conference call. Thank you for participating. You may now disconnect.