Exscientia PLC (EXAI) 2021 Q4 法說會逐字稿

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 Fourth Quarter and Full Year Ended 2021. (Operator Instructions)

At this time, I'd like to introduce Sara Sherman, Vice President, Investor Relations. Sara, you may begin.

Thank you, operator. A press release and Form 20-F was issued yesterday after U.S. market close with our fourth quarter and full year 2021 financial results and business updates. 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 on Slide 2 that we may make forward-looking statements on our call. These may include statements about our projected growth, revenue, business models and business performance, including with respect to our technology platform and pandemic preparedness program. 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; and Garry Pairaudeau, Chief Technology Officer; Ben Taylor, CFO and Chief Strategy Officer; and Dave Hallett, Chief Operations Officer, will also be available for the Q&A session.

And with that, I will now turn the call over to Andrew.

Thank you, Sara, and thank you to everyone who joined us today. 2021 was a remarkable year for Exscientia. We strategically scaled the company, and we expanded our capabilities.

As you can see on Slide 3, we have significantly grown our pipeline year-over-year, adding 11 programs and advancing 2 programs into late discovery and 3 into IND-enabling studies. The press release issued last night includes an exhaustive review of our 2021 accomplishments. Let me recap the most notable and a few recent highlights.

Hiring key talent and expertise, tripling the size of our global workforce and add into our U.S. footprint with a new Boston office and office expansion in Miami; completed our acquisition of Allcyte, integrating the world leader in patient tissue platform into our end-to-end system and gaining a tremendously talented team; listed on NASDAQ and raising over $510 million in gross proceeds from our IPO and private placement.

We ended 2021 with approximately $759 million in cash or cash equivalents and we are well positioned to deliver on our strategic imperatives.

Announcing one of the industry's largest AI-powered drug discovering development deals to date, with our $5.2 billion collaboration with Sanofi, with a $100 million upfront payment; successfully executing our partnerships, as we've seen by the expansion and work with 3 of our major partners, BMS, Sanofi and the Bill & Melinda Gates Foundation, with BMS in-licensing an AI designed immune-modulating drug candidate.

The successful application of artificial intelligence and machine learning to reduce our industry's failure rate and produce better, more effective medicines has long been recognized as transformative potential. We are now working to put that promise into practice.

In the last several months, we've seen some of the world's largest drug makers announced the largest deal to date in the AI-powered drug discovery. The back-to-back announcements by titans in biotech and pharma represents the industry's fullest embrace of AI to date, and we think are an inflection point in the evolution of AI-powered drug discovery and design.

It should come as no surprise that there is mounting dissatisfaction with the time it takes to deliver new medicines, particularly when we are faced with urgent health crises, such as the global pandemic. Even more frustrating is that most of this time is spent trying to fix problems as they arise, and with currently a lengthy step-by-step process falling out over the course of 10 years. It's striking when you consider no other consumer products are made by this way.

By the time the drug reaches the patient, the underlying research is dated by 10 years and the science is likely significantly advanced. Can you imagine if any other technology products were made in this way?

The founding team and I set out to build a completely new type of company to reengineer the drug discovery and design process. Today, that's best illustrated in the near equal splits in our team between drug discovery scientists and technologists, which you might be surprised is an anomaly in our industry.

By bringing together these 2 seemingly disparate disciplines, our scientists are able to tackle new problems with the power of our AI systems. Whilst our technologists encode these learnings, working towards a day when we can achieve full automation.

Today, our Chief Technology Officer, Garry Pairaudeau, will talk more about our technology and how we're using it to design and develop better molecules and the deep investments we've made in technology.

What is incredible about this is how our underlying technology and AI platforms may have the potential to achieve feats we have never seen before in drug discovery. Our AI platforms can make decisions based on analyzing thousands of different parameters in parallel, enhancing creativity with generative algorithm, working in a computational space far beyond the ability of any one scientist or team of scientists to consider.

Our drug design process from the AI generation of the first novel molecules to the design of a development candidate has averaged about 1 year versus industry standard of 4.5. Our AI-driven methods lead to the nomination of drug candidates after the synthesis, on average, of less than 1/10 of the number of compounds versus the industry average. This efficiency enables us to concurrently evolve more than 30 programs.

I'd like to think of AI as supercharging our amazingly talented drug discovery teams. This is a combination of human and machine that has enabled us to begin to crack foremost areas that have truly purified medicine, an area of our industry has been talking about more than 25 years.

Today, as seen by results published in Cancer Discovery, where our platform was the first to successfully guide treatment outcomes for late-stage cancer patients, achieving a 55% ORR. This gives us confidence that the models we're developing may translate to potential patient benefit in the clinic. This is an area that I'm personally very enthusiastic about, and I look forward to seeing where we can take the platform next, including ovarian, lung and breast cancers.

And as we look at what's ahead in 2022, we're driven by the possibility of how much we can advance powered by this AI-led approach. We anticipate continued expansion of our pipeline by not only adding new discovery programs, but by also continue to nominate new drug development candidates and progress spend towards our clinic.

So we're building out our clinical capabilities and infrastructure, increasing validation of a platform through additional data, including data on our pipeline programs, EXS-21456 and GTAEXS-617. That will be presented in April at the upcoming AACR Congress and throughout 2022.

Proof in our mission to fully automate drug creation for the opening of our laboratory automation suite in Oxford. Today, Garry, our CTO, will be focusing on just one aspect of our tech, how do we design better drugs. The technology team is up to some incredible work this year, including an opening and operationalizing a new 26,000 square foot automation suite that will bring us one step closer towards fully automating the chemical synthesis and analysis of our small molecules in our discovery process.

And I'll now turn over the call to Garry to walk through our technology platform.

Thank you, Andrew. Today, I would like to give you a high-level overview of our technology platform. So let me bring to life how the underlying technology at Exscientia is differentiated from what others in the industry are doing.

There are several fundamental ways in which I believe we stand apart. But perhaps the easiest way to explain it is where we start, with the patient, as you can see on Slide 6.

We think about drug discovery as a learning cycle, a cycle that begins with the patient, fueled by our AI platforms that enables us to learn from every new piece of data and bring more information to bear through every step of drug creation.

In a conventional drug discovery project, it may take years before a potential new drug candidate is tested in humans. With our AI precision medicine platform, we are able to bring this process much, much earlier into the discovery phase.

On the next slide, we show how we are identifying the right target. This is possibly the most important decision for a drug discovery program. We use Centaur Biologist, which integrates literature, along with genomic and transcriptomic data into our knowledge graph to identify connections and predict target-to-disease associations. This process is disease area agnostic with application to date across oncology, immunology, immuno-oncology and rare diseases.

Our precision medicine platform utilizes primary human tissue samples and we align our early target identification activities to leverage this platform, capturing new insights from drug action on patient cells along with transcriptomic and genomic data. All of this gives us increased confidence in the relevance of our targets to actually make a meaningful difference in improving the outcomes for patients and having the ability to better understand the potential impact long before we reach the clinic.

Once we've established the desired target, we've rigorously defined our objective, the target product profile, or TPP, which describes in detail the properties we desire in our optimized drug molecule. Once we have rigorously defined the target product profile on Slide 8, we now take this set of objectives and encode them as a reward bundle for our algorithms to optimize towards. This enables our design systems to create structures meeting those criteria.

For example, we may want to design a brain penetrant drug that has a low human dose, good selectivity, but in particular, avoids having efflux issues. We can include that specific set of objectives, potency, selectivity, efflux, et cetera, through normal structures generated drive towards these criteria.

As you might imagine, we use and generate a lot of data when doing this, illustrated on Slide 9. For each project, we generate the initial hit structures algorithmically from integrating any public data with proprietary data from fragment or focused screening, which is developed in-house.

As you've heard us talk about, half of our company are drug discovery scientists, generating proprietary assays and data at our Vienna and Oxford labs that we can bring to bear within our projects. In addition to that proprietary data, the platform can also scour existing data, going back years to search for anything that might be relevant. For example, data extracted for a 20-year-old patent or a recent nature paper can all be integrated with data generated in our labs this morning to help serve the models.

One of the great powers of our AI platform design is that we can use any type of data to drive the design process, meaning it does not require a specific data type like 3D, critical structures or high-content images, but we can use any and all of these types of data, plus many others, that will enable us to triangulate towards designing drugs that meet complex design requirements.

This diversity of data is required to precision engineer a novel chemical series that we anticipate will have a robust treatment effect in patients. In order for our system to generate potential molecules, we need models to predict all of the properties that we require. This could include potency, ADME, selectivity, physical properties and many, many more.

We have extensive model-building capabilities that stand a full range of skills and technologies from quantum mechanics and molecular dynamics to exploit structural information, to machine learning and computer vision to interpret pharmacology and cellular imaging.

Going back to our earlier example where we highlighted that we are trying to design molecules that meet specific project requirements, for example, the right level of selectivity and potency but that doesn't have unwanted issues. We are now at the stage where we have identified the desired TPP, and we have an initial set of models that will help guide us on that journey, as you can see on Slide 10.

We can now apply generative design, which is an AI-driven process of molecular ideation. Our system is exploring nearly the entirety of chemical space and creating molecules to meet our desired criteria, scoring them and learning, learning from the score is how to create better molecules.

Using evolutionary algorithms or reinforcement learning, the system rapidly and efficiently explores chemical space, creates a population of novel molecules that are predicted to meet our criteria. At the end of each situation, usually a population of tens to hundreds of thousands of molecules are created. These molecules are driving towards the criteria that we desire.

On the next slide, from this large population, we apply a detailed filtering process that may involve more sophisticated and compute-intensive models to reduce the set and then we apply a process called active learning.

We want to make a few molecules as possible because it's time consuming and expensive. Usually, we make 10 to 20 molecule per design cycle. Therefore, we want to synthesize and test the compounds that will help us learn faster to improve our models and to take us forward towards our objectives. It is by learning faster to navigate across a potentially vast chemical landscape that gives us the industry-leading productivity metrics that we have been demonstrating.

Our active learning algorithms ask which molecules will provide us with the most information to improve our models in a certain dimension. In short, what should we do next in order to learn the most and to select this set of molecules in an unbiased and mathematically rigorous way so that they enable us to learn the most at each cycle.

Now on Slide 12. The selected molecules are synthesized and tested. We profile each molecule in detail so that we can update our models with new information and learn the maximum amount from the laboratory work. We have extensive biology capabilities in our labs in Oxford, including structural biology, biophysics and pharmacology screening.

We can then visualize the project telemetry, the progress of the project in an unbiased way using what we call a merit score as a representation of the desired target product profile, as you can see on Slide 13.

Each dot is a novel compound synthesized and tested. The X axis is the sequential progress of the project in terms of compound numbers and the Y axis is the multiparameter optimization score, with 1 being the ideal score across multiple objectives. Each design cycle is colored from red through to blue.

As the project progresses, the system moves from exploration, where we are exploring a range of different chemotypes. Once there is promising phase identified, we move into an exploitation phase focusing on a particular area of chemical state. At this stage, molecules are consistently fulfilling most of the key project goals. And we will rapidly close down on a candidate molecule integral for preclinical testing. As we learn through each cycle, we can track the learning as the project progresses towards its desired criteria.

On the next slide, you can see that the AI algorithms are refining the final designs in order to achieve the project's potency, selectivity, bioavailability and safety requirements in a final candidate molecule. Hopefully, I've shown you how we design differentiated molecules, just one aspect of our end-to-end platform.

On Slide 15 is our learning loop. By starting and ending with the patient, we can apply the platform to produce new candidate medicines with attributes that we predict will lead to better treatment benefits. We are also using our precision medicine platform in biomarker discovery and in patient stratification as we move forward. You will hear more about this later in the year. So there's no better way to showcase the true value of our design capabilities than with an example.

I'll now turn the call over to Andrew to talk more about our design process with one of our programs in development as part of our pandemic preparedness efforts.

Thank you, Garry. Today, we want to highlight how our platform could truly overcome complexities and design challenges in efforts to create molecules that fit the desired properties we are seeking.

I'll start on Slide 17. We are showcasing our objectives for designing the drug against Mpro, a critical virus protease enzyme target for SARS-CoV-2, the coronavirus responsible for COVID-19. Mpro is a key enzyme of coronaviruses, and has a pivotal role in immediate in viral replication, making it an attractive drug target.

Importantly, we started this project less than 9 months ago in the summer of 2021. And with a clear target product profile, we've been able to design and synthesize promising compounds that are starting to meet our objectives in in vitro studies. We entered into a collaboration with the Bill & Melinda Gates Foundation in September 2021, and we've accelerated our efforts in pandemic preparedness.

We've not yet nominated our development candidate for this target, but for us, this was an important example to showcase our design capabilities and share some emergent early discovery data coming from our platform. So here you can see what our design objectives are, namely to develop a once daily, orally bioavailable covalent protease inhibitors with pan-coronavirus activity.

Turning to Slide 18. We've highlighted our process to design a potential candidate. This process is still ongoing and the in vitro data we will be highlighting today is illustrative of our design capabilities for an important target. Our design cycle utilizes generative design, as Garry mentioned, with a focus on improving key parameters and prioritizing the most promising compounds for synthesis and testing.

Importantly, we recently brought on Ian Goodfellow, Professor of Virology at the University of Cambridge, as our new Vice President of Antivirals. Ian is a leader in the field and he'll advance our efforts in developing our wholly-owned antiviral platform, including pandemic preparedness. Ian has already provided invaluable insight and we're pleased to welcome him to Exscientia.

Here on Slide 19, you can see we're looking at the potency of 2 of our lead molecules, as measured by equilibrium association content by surface plasmon resonance, or SPR. Compared to nirmatrelvir, the Mpro inhibitor given in combination with ritonavir to form Paxlovid, the first approved SARS-CoV-2 protease inhibitor.

In our head-to-head preclinical study, we are comparing the potency of 2 of our designed and synthesized compounds that have emerged from our AI design process. To be clear, we believe Paxlovid is an incredibly important drug that provides benefit to patients suffering from COVID-19. Our focus today is on how we can design an optimal antiviral, with the potential to be dosed once daily over without the need to be co-administered with ritonavir, which can result in adverse events and reduce the metabolism of other medications the patient may be taking.

What we're showing here is the progression of our design segment, how we can continue to learn and improve. Compound EXS-68 was a 68 compound sympathized and an early leader we designed to show superior enzyme-binding affinity based on SPR viable assay. Compared to nirmatrelvir, there was an elevenfold improvement in potency, as measured by enzyme-binding affinity and a potential to improve oral bioavailability.

One of our latest compounds, that's still undergoing profiling, compound 161, has shown a marked improvement of activity, being the most important compound in our series at the KD of only 3 pM. About 200-fold more potent than nirmatrelvir, as seen in the graph in this in vitro assay head to head.

As part of our pandemic preparedness efforts, we are focused not only in potency against SARS-CoV-2 that causes COVID-19, but other variants and coronaviruses to be able to design a molecule to have the potential to be useful in a future pandemic.

So on the next slide, in the chart, the lower the full variation, the more potent the molecule is against other coronaviruses in the legend as tested in the functional enzyme assay. The higher the bar, the more likely the compound is to lose effectiveness and need higher dosing against these other coronaviruses.

For some background, we wanted to look at coroviruses that were identified to cause severe disease, such as SARS-1 and MERS, both Betacoronaviruses with mortalities of approximately 10% and 34%, respectively as well as common respiratory viruses, as illustrated by 229E and NL63, both Alphacoronaviruses, and HKU1 and OC43, both Betacoronavirus.

The Exscientia compounds importantly showed broad spectrum activity in vitro across diverse coronaviruses. We believe that this activity, combined with biophysical potency on target observes in vitro and SPR, will be critical properties necessary to retain antiviral activity against emerging coronaviruses.

Turning to Slide 21. This ties to what Garry walked us through earlier. Our approach of designing against multiple objectives has allowed us to create a molecule that balances potency with our desirable properties. We were cognizant of the need of our compounds to be designed to avoid an off-target impact given their potency.

We have been able to design compounds where the increase in potency against the viral proteases did not come at the cost of inhibiting human system proteases with EXS-161 showing a greater than 1,000-fold selectivity and in vitro biochemical assays against human proteases.

So to summarize, on Slide 22, we have already been able to meet most of what we set out in our target product profile, in our efforts to develop a once-daily oral antiviral. Our platform was able to integrate viable target protease analysis with our state-of-the-art biophysical screening capabilities to design potent SARS-CoV-2 Mpro inhibitors, with selectivity of human proteases while still showing pancoronavirus activity.

We have designed and simplified compound with good drug-like properties, including promising antiviral activity and preclinical pharmacokinetics. We believe we have designed a molecule that shows, based on safety and relevant human cell lines, better potency compared to nirmatrelvir, potential broad-spectrum coverage and the ability to be dosed on its own, that with the properties that will allow it for co-dosing in the face of resistance.

We look forward to continuing to synthesize and design against our target product profile and to share with you more data on this important program later in 2022.

And with that, we will open the call for questions. Operator?

(Operator Instructions) Our first question is from Chris Shibutani with Goldman Sachs.

This is CJ on for Chris this morning. Congratulations on all the results and progress of the last quarter and year.

I was wondering if you could give us a sense of whether we should expect the AACR presentation for the adenosine receptor antagonist to give us a sense more of what the patient-specific expansions are going to look like when we get to the patient phase of the trials? Or should we wait for the kind of the top line or the more detailed -- healthy volunteer data to have visibility to that?

And maybe could you also give us a sense of sort of business development priorities for the year? I saw that you've promoted a business development chief at this point. How should we think about priorities there? Will it be more deals like the Sanofi deal? Or is there going to be a shift in some way?

Thank you, CJ. Thanks very much as well for your comments on the quarter.

In terms of answering the question on the AACR data, I'm going to hand that over to Dave Hallett, our Chief Operating Officer, to explore that. And then I'll come back on and talk about business development strategy for 2022. Dave, do you want to introduce how -- what we're thinking about introducing to the world at the AACR?

Sure. Thank you, Andrew. So the -- all 3 posters that we're going to present at AACR are really focusing on translational aspects and patient selection. And also one of the posters is actually touching on the design aspects that kind of Garry outlined, how they were applied to CDK7.

So going back to A2a specifically, think about timing of information this year. The AACR poster itself will focus on ongoing functional and multiomic work, which is to identify both novel and robust patient stratification methods ahead of a forthcoming clinical study that we're anticipating we will start in patients in the second half of this year.

The Phase I information that you referred to, we'll be looking to release that towards the end of the first half, and that will cover information such as pharmacokinetics, safety and tolerability, but also recommended Phase II dose based on a pharmacodynamic biomarker that we have in place.

And I'll pass you back to Andrew to address the question around business development.

CJ, yes, we've been incredibly active in business development, as you might have noticed, over the past 6 months or so. And that expansion for extra work not with just with joint ventures, but with BMS and Sanofi. We're keen as well also to make sure we do balance that with our business model. And key to that actually is how we think of our business development, particularly for 2022.

What we are thinking about also is ensure then that we build out our capabilities and showing that as we expand new ways of doing things that we're able also to bring along partners to do that. And in fact, you've already seen elements of that with the Sanofi deal. A big difference between that collaboration and the BMS collaboration was the inclusion of the precision medicine platform. And I think that gives you an example that as we develop new technologies, we then look to see how we can also work with partners at an early stage actually to ensure that core technologies are on the right track in terms of understanding real patient needs and real needs in the marketplace.

So one thing I would expect this year actually is to think about how we do technology deals as well as doing pipeline deals.

And also, I'd like to bring our Chief Strategy, Ben Taylor, as well, CJ, just to add some more color to that.

CJ, so just a quick note on the AACR poster. I think although we'll probably save most of the enrichment data for around when we're starting the actual clinical trial in the next phase. What is really exciting about -- what we think is really exciting about the A2a poster is you'll see some evidence of how you can have a functional ex vivo IO model.

And remember, A2a is not a direct cytotoxic agent. So we really have to have that immune interaction, which you're not going to see in most all of the current translational models. And that's why IO has really suffered from having good translational models. So this is a really exciting potential model that could be used not only for A2a, but hopefully other IO agents in the future that might be more directly relevant to the patient environment.

Absolutely, Ben. And that really underlines the importance of these models more generally to the company. And the way we think about it, CJ, is that it's not just the human data from the Phase Ia on the molecules. It's actually the work being done in parallel on the ex vivo human data in defining the patient selection approach, which we are taking. And there's two bits of work coming together then into designing the Phase Ib2.

Great. That ex vivo assay has certainly been a big gap. So looking forward to seeing that data.

Our next question is from Michael Ryskin with Bank of America.

I want to start on the Mpro inhibitors you talked about just because we spent the least amount of time on that. Given how quickly the COVID pandemic is evolving and given the presence of Paxlovid and other agents out there, I'm just wondering if you could talk about the potential to accelerate the development of a final candidate? And could you talk a little bit more about your commercialization strategy, sort of the next steps beyond that assuming you're able to develop a solid candidate?

Excellent. Thanks, Mike. I'll give a bit of introduction to the question, and then I'm going to hand it over to Dave again, actually, to give you a lot more detail on how we're thinking about it.

The first thing, of course, is this project that we showcased today, I think, is a really good example of the ability of the company to rapidly design and develop high-quality drug molecules. And I think the data we're getting through now actually really places that in context.

The other thing that's important to take on board is that all this started with our collaboration with the Gates Foundation and the private placements that came in at the IPO. And it's a case then of really giving us that key partner going forward and ensuring that we are ambitious in sort of the target product profiles that we're going after, particularly compared to some of the competitor molecules that are out there. And when we look at our markets, the ideal TPP we think about is how do you identify something that could be low dose and long acting, that really has the protection against future variants that we're seeing.

And to give you some more context now about how that program is developing and how we're thinking about it, I want to introduce Dave now to the table.

Thanks you, Andrew. I think in terms of the first question about specific timing, yes, it's important to note that we continue to kind of synthesize molecules and explore the lead series we have in place. And we're looking to select a development candidate in the second half of this year, obviously, clearly aware of the timing and the need for additional agents.

In connection to your first question about Paxlovid in a wider market. I think it's interesting that the kind of current climate is that the data is around both vaccines and small molecules tells us a lot of things around that there is waning resistance kind of following vaccinations. We continue to live in a global environment where vaccine uptake around the world differs by geographies.

So in some areas, vaccination uptake, particularly in high-risk areas, is still very low. And even more recent data in a Nature paper that just come out showing how people actually infected with the Omicron variant is they actually generate a really low immune response and is unlikely to kind of generate this kind of herd immunity that, I guess, everybody is hoping for.

So I think it highlights, as is often the case with viruses, is that the combination of the vaccines, the potential for resistance to agents on the market. I think it just highlights once again the importance of having multiple new therapy options available for everyone, not just developed nations. And also thinking about how we might apply these in combinations as we saw successfully applied, for example, think about HIV therapies and the importance of multiple therapies, but against different mechanisms of actions. So those are the kind of things that we're looking for over the next couple of years.

And that's why first as well, Mike, it was important to design an agent that doesn't need co-dosing with metabolism inhibitors such as ritonavir. And if we're going to create combinations with co-dosing, we think it's -- we'd rather a strategy where ultimately you combine in 2 agents acting on COVID hitting at different mechanisms. And as Dave says, we've seen with HIV to be a very successful approach in the long term.

Okay. That's helpful. And then a follow-up, just sort of on investment priorities for 2022. You've got a very healthy balance sheet exiting the year. And you also have the upfront payment from the Sanofi collaboration from January.

So how do you think about expanding investment priorities this year, given the relatively neutral cash flow from operations last year? It seems like you should be able to support a lot more investment in expand products. I mean, you kind of -- you indicated, I think, 30 programs concurrently floating that you will be able to run. But as you move into discovery and IND enabling, where should we -- what should we expect the incremental spend to come in and sort of what's a good runway as we think about progressing through 2022 for that?

Excellent. Thanks, Mike. I want to introduce to you Ben Taylor, actually to take that question as CFO and Chief Strategy Officer. Ben?

Mike, so if you saw our financials for this year, we had an operational cash burn of about $9 million. A lot of that is because we can offset so many of our expenses with cash flows from partnerships. So we brought in a little over $85 million, so just a little bit above the guidance that we've given earlier on in the year for our cash flows from our partnerships.

I would expect that to continue into the coming year. So we've already had the $100 million upfront from Sanofi come in. That will probably hit the actual balance sheet in the second quarter, but we did sign the contract in the beginning of this year. We've had a number of other smaller milestones come in as well. So we're going to have a nice cash flow from collaborations this year again and have meaningful growth over the cash flow from collaborations last year.

So even though we will be growing our operations significantly, and in a second, we can talk about what that means. There should also be a nice balancing from those inflows. So that will maintain a very balanced business profile of growing our business in a way that matches our growth in our partnerships as well.

So just a quick note on some of the areas that you asked about investment. We continue to grow our platform capabilities, and I'm going to turn this over to Garry in a second to talk about it, along with all of our projects. The project growth -- you'll see our pipeline grow as well. But remember, our partnered programs pay for themselves ahead of time. So that actually reduces the net burn considerably.

And then we will continue to make some capital investments. We're growing some of our offices around the precision medicine out in Vienna. We opened up a 50,000 square foot facility. We've got an automation lab, that Garry can talk about, in the Oxford area as well.

So we will have some increase in cash outflows, but I'd imagine that it will stay in a very reasonable neighborhood. Garry, do you want to pick it up from there?

Sure. Thanks, Ben. I think -- I mean you hit on 2 key expansions for us. I mean, obviously, we're building out our technology platform going really deep into the AI capabilities that we described earlier and that cover our kind of end-to-end platform.

But 2 things I'd call out that we're really excited about is obviously the automation lab that I think we've mentioned previously, 26,000 square foot just south of Oxford. And we're deep into the design and specification and building all the equipment to go into that. So -- and that's going to be fantastic, end-to-end synthesis, purification, screening capability, which can really bring a transformational benefit to timelines in drug discovery.

And then the other new exciting area that we're starting to look at is we recently announced that we hired Professor Charlotte Deane, which is a super exciting hire into the tech group and into the organization. And she's going to be looking at developing our biologics capability and how we can apply. It's a really tight synergy with the sort of work we're doing at the moment, and we can -- how we're going to apply AI to the design of biologics.

Great. Just to confirm. A quick one. So are you still sort of projecting about 5 to 6 years of cash runway? I think it's something you commented on earlier.

Yes. We haven't given specific guidance, but I think we feel very comfortable with our number of years of cash flow runway. So part of that is also, remember, under our control as we determine the flow between internal pipeline and partnerships, but our business model expectations would certainly meet around what you're talking about.

Our next question is from Peter Lawson with Barclays.

Thanks for the update and the detail on the call. Maybe just a follow-up question for Ben. Just as we think about the build-out you're undergoing at the moment, how should we think about your needs for expanding when you move into Phase II and Phase III clinical trials as well?

So we've actually factored that into a lot of our current thinking on the growth. So we're building up our clinical team and doing it in what we think is a very balanced data-driven way. But I wouldn't expect that side of the business to really become the major cost center until some of the drugs get into late Phase II or Phase III. And then obviously, the clinical trials get more expensive.

So in the near term, it's not going to have a dramatic impact on our overall cash expenses. And we're able to manage that much more. So I don't see that as being a substantial line item right or a driving line item.

Got you. Is there anything in that kind of later stage development that you can improve on as well, whether it's from the AI side of things? Or is that kind of most like a bolt-on of existing approaches?

Well, now, Peter, you're getting to where we get really excited. So we hope so is the right answer. We're intending to take a very data-driven approach to clinical trials, too.

So as we mentioned earlier, a lot of what we designed for is actually better clinical trials. And so what we need to do is match those clinical trials to the drugs that we're producing. So if you think about that, any clinical trial, whether it's Phase I or Phase III, is all about statistics.

And so the more powerful you can make your statistical analysis, the smaller the trial, the faster the trial, the better the results that you can get to. So by designing more targeted clinical trials, it actually has the follow-on effect of potentially making them smaller and faster and less expensive.

Then just on the kind of the near term on AACR, what should we be looking for in the CDK7 preclinical data?

I'd like to Dave to take that question, Peter.

So the key information you'll see in New Orleans is ongoing work around -- as well as the design of the molecule and some in vitro data and in vivo data is showcasing the kind of qualities of the development candidate we have is ongoing data we're generating in primary patient tissue, which is helping us to identify not only which cancer types, but also within those kind of specific kind of site. So for example, ovarian, which patients are likely to respond better and why? And so producing kind of signatures that we can then use prospectively in a future patient study to kind of -- to guide exactly what Ben has pointed out is to highlight which patients are likely to respond to our drug and which aren't and understand why. And ultimately, that will drive very specific recruitment, should allow us to actually run smaller clinical studies and therefore, actually get earlier and more successful readouts.

Got you. And then I guess the final question, just around A2aR, just that as a single agent, are you getting a sense of what percentage of patients could share a response as with a single agent A2aR?

I think so. I think you'll see some of that in the poster. I think it also depends on the cancer type. So if you look at the data that others have published and that we're building upon, is that sort of particularly looking at which subjects do you see this high adenosine signature and also where do you see kind of high expression of important enzymes like CD73 and other components that are likely to respond. It varies across cancer types. So it can be as high as, say, 15%, 20% in some areas. It can be much lower than that in other areas.

And so as part of the work that we'll present at AACR, you'll start to get a sense of the cancer types that we're focusing on, that will actually form the basis of both the dose escalation and the expansion. So again, with this kind of concept of narrowing down on a smaller patient kind of subsets as we go into the clinical trial, again, identifying which patients are likely to respond.

So it depends, is the answer to your question. But I think the key thing is actually is knowing and having data available and a more important kind of biomarkers and tools to actually identify those patients before you take them into the clinical trial.

Our next question is from Vikram Purohit with Morgan Stanley.

So I had 2 both kind of on the platform. So first, is there any color you could provide at this point on the targets that have been identified through the collaboration with Sanofi? I understand it's early days and you may not be able to share much about targets in particular, but any context you might be able to give around the process for identifying these targets and then prioritizing them? That would be very helpful.

And then secondly, for the precision medicine platform highlighted by the EXALT-1 data, where specifically do you think you could apply this functionality next? And how do you see it being weaved through your current pipeline programs over the coming months and years?

Thanks so much, Vikram. Great questions, actually. It gives us a chance to talk about the expansion of our end-to-end platform. And that's actually a real key feature of the Sanofi deal. In fact, it's moving upstream into using our target ID approaches and downstream into using precision medicine into patient stratification. And in fact, those 2 things do come together in how we're thinking about identifying new targets for Sanofi.

So to give you a bit more color on that, I'm going to bring Dave into the conversation, as it's been his team who's been identifying targets in (inaudible) using the platform.

Thank you, Andrew. I think there's some key components to this is that -- so the first thing to appreciate is that the sort of the areas or spaces kind of cover oncology and inflammation. And so what we're able to do there is kind of a few ways of approaching kind of target selection and target validation. One of obviously the critical ones and it's kind of at the heart of the -- it's one of the reasons that Sanofi did a collaboration. It's to kind of -- we give them access to our patient-driven approach to target identification, which comes back to this kind of critical story of placing the patient at the center of both the target discovery, but also the kind of translational aspect.

And so we are -- what that looks like in practices is that we're assembling data sets that's provided to us by Sanofi because their therapy area has obviously been thinking about this for a while and the kind of targets they want to work on. So they give us access to their proprietary data. We can actually then add on, as Garry described earlier, in terms of the rich history of kind of public literature, both patents and peer-reviewed information from the last 20-plus years.

And then add on to that the information that we're getting from our platform in Vienna which again, it's kind of experimental data over the function level, but also genetic and transcriptional. And we basically bring all that information together and to ask questions around how strong is the relationship between a particular target and the disease of interest and then able to kind of narrow that further down into kind of subtypes of cancer.

So it's still early stages, but I think the power that we have from the kind of proprietary data that Sanofi had brought coupled to our own is kind of -- stands in good study in terms of both identifying novel targets, but also kind of prosecute to them as we go over the next few years.

A good example of this pipeline discovery, Vikram, is actually will be presented at one of the AACR posters, which is how we can show that a deep learning approach to prime mutation tissues is actually being used to discover novel mechanisms of action, that particular one is ovarian cancer, which we'll talk about in a few minutes. But it gives us, I think, a textbook example of how we're using the platform then to start off with the patient tissue material and then use it then for a novel target discovery. So once our poster is out, actually, we will tell you more details about that. But I would recommend looking at it.

In terms of then the broader application beyond EXALT-1 and the precision platform. Firstly, we are incredibly pleased with the EXALT-1 paper that was published in Cancer Discovery and the results of that trial. It's the first time that an AI-based system has shown improved outcomes in oncology. That's an important thing to note. The other really important thing to note was of course the results of our trial. The hazard ratio of 0.53 and an ORR of 55%.

And if you look at sort of where the patients with (inaudible) 1 or less, they have even seen more significant benefits within that by the sort of assay AI guided sort of therapeutic approach. So that was, of course, a trial in hematologic cancers. And of course, we are now got a clinically validated approach in that and hematologic cancers are things we are exploring now in terms of wider internal precision oncology sort of pipeline as we go forward.

But what we are doing now is looking to rapidly expand the range of cancers that we can then apply the same methodology to in developing a lab-based high-content AI-driven assays and also then looking to run both sort of observational and of investigational trials in other cancers along the lines of EXALT-1. We are advancing those stages now. We're looking at ovarian cancer, breast cancer, lung cancer, we're actively developing. You'll see developments around (inaudible) taking place, and evidence of that already gathering. So what we're looking to do really is build this out for a range of other cancers.

The key to doing that really is also the other part of our precision medicine platform, which we'll talk about in a future earnings call, which is how we expand our clinical network and our biobanks. The range of clinicians that we can interact with, you'll be hearing about a lot more sort of collaborations in that space as that network expands, form sort of Central and Eastern Europe at the moment, across other continents, hopefully, including the U.S. and Asia as well.

And what we see from that then is that expanded network of [patients] providing then underlying material and data from most patients, which allows us then to expand our biobank. And what's really exciting, of course, is the depth of analysis we extend into. Not just high-content approaches, but also a much wider range of all mixed approaches now taking place, including transcriptomics, single cell sequencing and basically trying to extract as much deep information as we can and deep profiling on every hard one biobank sample that we've managed to gather.

And that's what's really exciting about this. And that then also provides new data into the target validation platforms as well, of course, it provides us with much more sophisticated approaches to think about patient stratification, where we consider then as a multiomics approach far beyond even uniomics.

Maybe just kind of wind up in the -- maybe I'll just -- it's important we just signed a really exciting kind of collaboration with Sanofi. We've actually already identified a few targets. And we just initiated the operational relationship around that. So we'll keep informed us the progress of that over the coming months and years.

We have no further questions at this time. I'll turn the call over to Andrew Hopkins for any closing remarks.

Thank you, Chris. And thank you to everyone who joined us today. As (inaudible), it can be easy to solely focus on the exciting new chemistry and biology in the creation of a new medicine. However, I hope that today, we've illuminated how with our technology systems, that can truly take the best of science and accelerate it, helping to move us towards a world when we see that all medicines might be designed with the extraordinary computing power of artificial intelligence and machine learning, enabling all of us in the industry to achieve more in advancing new medicines for patients.

And with that, thank you for your time today, and it's been a pleasure.

Ladies and gentlemen, this concludes today's conference call. Thank you for participating. You may now disconnect.