Recursion Pharmaceuticals Inc (RXRX) 2025 Q1 法說會逐字稿

Hi, everybody. My name is Chris Gibson, Co-Founder and CEO of Recursion, and I'm delighted to welcome you to our earnings call this morning. We're going to go ahead and get started.

大家好。我叫克里斯·吉布森 (Chris Gibson)，是 Recursion 的聯合創始人兼首席執行官，我很高興歡迎您參加今天上午的財報電話會議。我們將繼續並開始行動。

Perfect. So of course, important to note that we're going to be providing forward-looking information today, so please understand all of these important caveats.

完美的。因此，當然，值得注意的是，我們今天將提供前瞻性信息，因此請理解所有這些重要的警告。

So I want to begin by just talking a little bit about Recursion's mission, which is to decode biology to radically improve lives. And unlike a traditional biotech, where learnings from a program typically work within that specific program, those learnings could improve a program, or the scientists from a given program might take some of those learnings onto their next program.

所以我想先簡單談談 Recursion 的使命，即解碼生物學以徹底改善生活。與傳統生物技術不同，傳統生物技術中從一個項目中獲得的經驗通常在該特定項目內發揮作用，這些經驗可以改進一個項目，或者特定項目的科學家可能會將其中一些經驗運用到他們的下一個項目中。

At Recursion, we're trying to do something different. We're trying to build a learning system, a Recursion operating system, that learns from every program to make the next generation of programs better and better, and that requires some scale. And what you're going to see today is that we're taking decisive action to make sure that we can continue to both take our internal pipeline forward, our partnerships forward, and also that we can continue to run this critical experiment for the biopharma industry, and that is to build a firstrate tech bio company.

在 Recursion，我們正在嘗試做一些不同的事情。我們正在嘗試建立一個學習系統​​，一個遞歸作業系統，它可以從每個程式中學習，以使下一代程式越來越好，而這需要一定的規模。今天您將看到，我們正在採取果斷行動，確保我們能夠繼續推進內部管道和合作夥伴關係，同時確保我們能夠繼續為生物製藥行業開展這項關鍵實驗，即打造一流的科技生物公司。

I want to talk a little bit about our earliest version of the platform, Recursion 0.1. This was a platform built on top of phenomics and siRNA and repurposing. And today you'll hear a bit about how some of those programs have done well, like our FAP program, with preliminary efficacy data and safety data we'll share soon, and also how some of those programs have not turned out the way we hoped, such as our CCM program. But building on the learnings of that first generation, we built an improved Recursion operating system, Recursion 1.0 that added transcriptomics that allowed us to go after new chemical entities and use advanced tools like CRISPR. That's allowed us to take forward incredible programs like our RBM39 program and our C.

我想稍微談談我們平台的最早版本 Recursion 0.1。這是一個建立在表型組學和 siRNA 之上並可重新利用的平台。今天，您將了解其中一些項目如何取得良好進展，例如我們的 FAP 項目，我們將很快分享初步療效數據和安全性數據，同時也了解到其中一些項目如何沒有像我們希望的那樣取得成果，例如我們的 CCM 項目。但在第一代技術的基礎上，我們建立了改進的 Recursion 作業系統 Recursion 1.0，它增加了轉錄組學，使我們能夠追蹤新的化學實體並使用 CRISPR 等先進工具。這使我們能夠推進 RBM39 項目和 C 等令人難以置信的項目。

diff program, some of these programs advancing, and others we're holding back for strategic reasons today. And ultimately, Recursion 2.0 is what we are now. Post the combination with Exscientia, we've seen the power of combining our 2 platforms, the Multimodal data, the Compute and the Active design, these are allowing us to generate a new generation of early stage discovery programs that we think are extraordinarily exciting. And all of this work is enabling us to demonstrate these leading indicators of success. We're able to validate our hypotheses more quickly.

不同的項目，其中一些項目正在推進，而其他項目我們今天由於戰略原因而擱置。最終，我們現在看到的就是遞歸 2.0。與 Exscientia 合併後，我們看到了結合我們的兩個平台、多模式資料、計算和主動設計的力量，這些使我們能夠產生我們認為非常令人興奮的新一代早期發現程式。所有這些工作使我們能夠展示這些成功的領先指標。我們能夠更快地驗證我們的假設。

We're able to generate candidates with fewer molecules synthesized. We can spend less and go faster. And through each generation of Recursion's operating system, we expect to improve on these kinds of parameters. And today, we're sharpening our focus, sharpening our R&D portfolio, because we committed to doing that with the combination of Recursion and Exscientia, 2 of the leading tech bio companies, because we've seen the power of our Recursion OS 2.0 platform, and we want to make sure we can double down on the winners and also make sure that we're the kind of company that can decisively move away from programs that don't meet our mark. And finally, we understand the challenging macroeconomic environment, and we want to be absolutely sure that in this kind of uncertain environment, we're making disciplined and thoughtful decisions to ensure that we can deliver on our long-term mission to decode biology, radically improve lives.

我們能夠透過合成更少的分子來生成候選分子。我們可以花費更少，進展更快。並且透過 Recursion 作業系統的每一代，我們都希望能改進這些參數。今天，我們正在加強我們的重點，完善我們的研發組合，因為我們致力於透過合併兩家領先的生物科技公司 Recursion 和 Exscientia 來實現這一目標，因為我們已經看到了 Recursion OS 2.0 平台的強大功能，我們希望確保我們能夠加倍投入到獲勝者身上，同時也確保我們是那種可以果斷放棄不符合我們標準的項目的公司。最後，我們了解充滿挑戰的宏觀經濟環境，我們希望絕對確保在這種不確定的環境中，我們能夠做出嚴謹而深思熟慮的決定，以確保我們能夠履行解碼生物學、從根本上改善生活的長期使命。

So today, we're unveiling our go forward pipeline, more than five clinical and preclinical programs that we believe have a much higher probability of success. We believe these are programs that are worth taking the shot, and we're doubling down on them, and we're going to hear a lot more about each of these a little bit later today. But before we do, I want to also share that it's not just about our internal pipeline. It's also about our partnerships. We brought in more than $450 million earned through these 4 collaborations to date.

因此，今天，我們推出了我們的前進之路，其中有五個以上的臨床和臨床前項目，我們相信這些項目的成功機率要高得多。我們相信這些項目值得一試，我們正在加倍努力，今天晚些時候我們將聽到更多關於這些項目的消息。但在此之前，我還想說的是，這不僅與我們的內部管道有關。這也與我們的合作關係有關。迄今為止，我們透過這四次合作獲得了超過 4.5 億美元的收入。

Today we're also sharing that we've received our fourth program option from Sanofi, part of that collaboration, and I believe that through our continued work on programs like those that we're advancing with Roche Genentech and with Sanofi, Recursion is not only going to continue building its Recursion operating system, it's going to continue learning to improve not only our internal pipeline, but all of the partnership programs that we advance together in the future. So with that, I want to turn it over to our Chief R&D Officer and Chief Commercial Officer Najat Khan. And I just want to say a huge thanks to her and the team has done incredible work to help us make these important decisions for the future of the company. And with that, over to you. Najat.

今天，我們還將分享我們從賽諾菲獲得的第四個項目選項，這是雙方合作的一部分，我相信，透過我們與羅氏基因泰克和賽諾菲共同推進的項目的持續合作，Recursion 不僅將繼續構建其 Recursion 操作系統，還將繼續學習，不僅改進我們的內部管道，而且還改進我們未來共同推進的所有合作項目。因此，我想將這個權力交給我們的首席研發長兼首席商務官 Najat Khan。我只想向她和她的團隊表達衷心的感謝，他們做了出色的工作，幫助我們為公司的未來做出了這些重要的決定。好了，就交給你了。納賈特。

Thank you, Chris. Good morning, good afternoon, good evening, everyone. Thank you so much for joining our Q1 2025 earnings call. As Chris mentioned, over the next 40 minutes or so I'll walk through some of the key pipeline updates, delivering on our commitment to sharpen our focus following the culmination of Exscientia, I'll also highlight the programs we're advancing with the potential for greatest impact, and also programs that we have thoughtfully chosen to discontinue. And in addition to that, I've rounded out -- so to go back to the next -- prior slide.

謝謝你，克里斯。大家早安，下午好，晚上好。非常感謝您參加我們的 2025 年第一季財報電話會議。正如克里斯所提到的，在接下來的 40 分鐘左右的時間裡，我將介紹一些關鍵的管道更新，履行我們在 Exscientia 結束後加強關注的承諾，我還將重點介紹我們正在推進的可能產生最大影響的項目，以及我們經過深思熟慮後選擇停止的項目。除此之外，我已經完成了 — — 因此請回到下一張 — — 上一張投影片。

Chris shared this slide. I just want to double down on a few more points. So, first of all, 3 key points to consider. One is our pipeline really reflects the strategic application of Recursion OS and AI, where it matters the most. So as I go through each of these programs, I'll talk about places where we have novel biological insight, or areas where we're engineering and designing differentiated molecules, as well as areas that we're driving precision based development, a really key theme that we're doubling down on further.

克里斯分享了這張投影片。我只是想再強調幾點。因此，首先，需要考慮 3 個關鍵點。一是我們的管道確實反映了 Recursion OS 和 AI 的戰略應用，這是最重要的。因此，當我介紹每一個專案時，我會談論我們擁有新穎生物學見解的地方，或者我們正在設計和設計差異化分子的領域，以及我們正在推動精準開發的領域，這是我們正在進一步加倍努力的關鍵主題。

But every single asset you see on this page is done with one end in mind, which is programs that aim to create differentiated medicines that patients are waiting for. The second point, in terms of sharpening our focus, we're doubling down in these programs, both in oncology and complimenting it with a focused effort in rare diseases. As Chris mentioned, we're advancing over 5 internally developed programs with first or best in class potential, each targeting unmet needs with a clear and efficient path to development and potential launch. So now the third point, look as part of having portfolio, since we did the integration, the portfolio has grown, and we said that we would actually make disciplined decisions to sharpen our focus. This is based on both data, and I'll talk through that, but also strategic considerations.

但是，您在本頁上看到的每一項資產都是以一個目的為目的的，即旨在創造患者所期望的差異化藥物的計劃。第二點，在加強我們的重點方面，我們正在加倍投入這些項目，既在腫瘤學領域，也在罕見疾病領域中重點關注。正如克里斯所提到的，我們正在推進 5 個內部開發的項目，這些項目具有一流或一流的潛力，每個項目都針對未滿足的需求，並有清晰有效的開發和潛在啟動途徑。現在第三點，作為投資組合的一部分，自從我們進行整合以來，投資組合已經成長，我們說過我們實際上會做出嚴謹的決定來加強我們的關注。這是基於數據（我將對此進行討論）以及戰略考慮的。

And it includes deprioritizing 3 programs, NF2, CCM and C. diff. We're also placing LSD on a strategic pause as we assess opportunities for more differentiated TPP, and we've also made some choices in our pre-clinical programs, a balanced approach, both pre-clinically, in research and also in development. You will see through the presentations that this move, and these moves reflect our clear commitment to a high bar on differentiated medicine following the special combination, while also contributing to capital efficiency by reallocating these precious resources towards the highest potential opportunities. Next slide.

其中包括降低 3 個程序的優先級，即 NF2、CCM 和 C. diff。我們正在策略性地暫停 LSD 項目，以評估更具差異化的 TPP 機會，並且我們在臨床前項目中也做出了一些選擇，即在臨床前、研究和開發過程中採取平衡的方法。透過示範您將看到，這項舉措反映了我們在特殊組合之後對差異化藥物高標準的明確承諾，同時也透過將這些寶貴的資源重新分配給最具潛力的機會來提高資本效率。下一張投影片。

So, let's start with some of the go-forward programs from the prior slide. I'm going to go through further details, but I just want to hit a few key points as a summary. So CDK7, starting with our selective and reversible CDK7 inhibitor, REC-617, was precision designed using our Recursion 2.0 platform with a remarkable 136 novel compound synthesized; compared to thousands that are typically made. It was designed to optimize that therapeutic index with the goal to improve safety and efficacy compared to our competitor molecules. Early clinical data, which we shared in December from our monotherapy shows there's encouraging monotherapy activity, including a partial, confirmed partial response to the platinum resistant ovarian cancer patient, and so far, a manageable safety profile.

那麼，讓我們從上一張幻燈片中的一些前進計劃開始。我將進一步介紹細節，但我只想總結幾個關鍵點。因此，CDK7 從我們的選擇性和可逆性 CDK7 抑制劑 REC-617 開始，使用我們的 Recursion 2.0 平台進行精確設計，合成了 136 種新型化合物；與通常生產的數千個相比。它旨在優化治療指數，目標是與我們的競爭對手的分子相比提高安全性和有效性。我們在 12 月分享的單一療法早期臨床數據顯示，單一療法具有令人鼓舞的療效，包括對鉑耐藥性卵巢癌患者的部分、確認的部分反應，並且到目前為止具有可控的安全性。

As we committed to before, we will open combination studies in first half of 2025, and further details about that will be announced upon study initiation. Now turning to our potential first in class, RBM molecular grouppdegrader, REC1245, this emerged from our phenotypic insight that revealed RBM39 as a potential novel mechanism that is functionally linked from our phenotypic platform to CDK12, a historically challenging oncology target. By degrading RBM39, REC1245, is designed to potentially move the downstream effects of CDK12 inhibition, disrupting RNA splicing to down regulate cell cycle checkpoints, DDR networks, etc. This triggers cell stress and apoptosis. We advanced this program from Target ID to IND enabling in less than 18 months, showcasing the speed and precision of our learning discovery and Recursion 2.0 platform.

正如我們之前承諾的那樣，我們將在 2025 年上半年啟動聯合研究，有關研究的更多細節將在研究開始時公佈。現在轉向我們潛在的同類首創產品，RBM ​​分子基團降解劑 REC1245，它源於我們的表型洞察，揭示了 RBM39 是一種潛在的新機制，其在功能上與我們的表型平台和 CDK12（一個歷史上具有挑戰性的腫瘤學靶點）相聯繫。透過降解 RBM39，REC1245 旨在潛在地轉移 CDK12 抑制的下游效應，破壞 RNA 剪接以下調細胞週期檢查點、DDR 網路等。這會引發細胞壓力和細胞凋亡。我們在不到 18 個月的時間內將專案從目標 ID 推進到 IND，展示了我們的學習發現和 Recursion 2.0 平台的速度和精確度。

The clinical development program, as we noted earlier, is focused on a biomarker defined set of solid tumors and select lymphomas, with preclinical studies supporting its PK/PD relationship and anti-tumor activity. The study is now in monotherapy dose escalation. Turning to the next one, MALT1. This is our selective best in class MALT1 inhibitor, which recently entered full Phase I dose escalation last quarter with the first patient now dosed. REC-3565, is being developed for relapse refractory B cell malignancies.

正如我們之前提到的，臨床開發計劃專注於生物標記定義的一組實體腫瘤和選擇性淋巴瘤，臨床前研究支持其 PK/PD 關係和抗腫瘤活性。該研究目前處於單一療法劑量遞增階段。轉到下一個，MALT1。這是我們精選的最佳 MALT1 抑制劑，該抑制劑於上個季度進入 I 期全面劑量遞增階段，目前已為第一位患者給藥。REC-3565 正在開發用於治療復發難治性 B 細胞惡性腫瘤。

Again, another example of a molecule that was designed by a Recursion 2.0 platform, especially the generative AI piece, which integrated hotspot analysis and molecular dynamics to enable best in class profile with improved potency, selectivity and safety. And one piece I want to point out is that the molecule was designed to avoid meaningful inhibition of UGT-101, a known off target liability seen in this class of molecules that can drive rive hyperbilirubinemia. Next, PI3K. This is our most advanced, one of our most advanced pre-clinical programs. It's a PI3K alpha inhibitor designed to highly selectively target the H-104 seminar mutation, which is a driver alteration presented about 14% of breast cancers and 4% of all cancers, while PI3K of course, is a crowded landscape, this molecule was again developed using our generative AI platform to optimize selectivity from the mutant as well as over wild type.

再次，這是由 Recursion 2.0 平台設計的分子的另一個例子，特別是生成 AI 部分，它整合了熱點分析和分子動力學，以實現同類最佳的配置，並提高效力、選擇性和安全性。我想指出的一點是，該分子的設計是為了避免對 UGT-101 產生有意義的抑制，這是已知的脫靶責任，存在於可導致高膽紅素血症的此類分子中。接下來是 PI3K。這是我們最先進的、最先進的臨床前項目之一。它是一種 PI3K α 抑制劑，旨在高度選擇性地針對 H-104 研討會突變，這是一種驅動變異，在約 14% 的乳腺癌和 4% 的所有癌症中出現，而 PI3K 當然是一個擁擠的領域，這種分子再次使用我們的生成 AI 平台進行開發，以優化對突變體和野生型的選擇性。

So, about 100x more selectivity over wild type and 10x greater selectivity from some of the wild type sparing inhibitors that you've seen recently. And to date, in our pre-clinical models has not shown any sign of hyperglycemia or GI tox. Early preclinical data is something I'll share in greater detail today that is showing 2 regressions at low doses supporting a potential therapeutic window, a broader therapeutic window that limits liability, such as hyperglycemia, with the goal to also improve the potential for efficacy. Next, I'll move to some of our targeted rare disease programs, FAP. So FAP, this is an allosteric MEK1/2 inhibitor in development for the orphan disease FAP.

因此，與野生型相比，選擇性大約高 100 倍，與您最近看到的一些野生型保留抑制劑相比，選擇性高 10 倍。到目前為止，我們的臨床前模型尚未出現任何高血糖或胃腸道毒性的跡象。我今天將更詳細地分享早期臨床前數據，該數據顯示低劑量下的 2 次回歸支持潛在的治療窗口，更廣泛的治療窗口可限制責任，例如高血糖症，目的也是為了提高療效的可能性。接下來，我將介紹一些我們針對的罕見疾病項目，即 FAP。因此，FAP 是一種針對孤兒病 FAP 開發的變構 MEK1/2 抑制劑。

Just as a reminder, there are no approved therapies, and the unmet need is very high. This is a potential first in disease program that originated, as Chris mentioned, from the earliest iterations of the Recursion OS platform, showing an unexpected and novel insight between MEK and APC and FAP disease modulation. 4881 is currently an ongoing Phase II open label signal safety study. As you know, some of the initial data, including safety, tolerability and preliminary efficacy of REC-4881, 4 milligram was presented yesterday by our investigator, Dr. Joel Samadar, in a late breaking podium session at the DDW Conference in San Diego.

需要提醒的是，目前還沒有批准的療法，未滿足的需求很高。正如克里斯所提到的，這是一個潛在的首個疾病項目，它起源於 Recursion OS 平台的最早迭代，展示了 MEK 和 APC 以及 FAP 疾病調節之間意想不到的新穎見解。 4881 目前正在進行 II 期開放標籤訊號安全性研究。如您所知，我們的研究員 Joel Samadar 博士昨天在聖地亞哥 DDW 會議的閉幕式上介紹了 4 毫克 REC-4881 的一些初步數據，包括安全性、耐受性和初步療效。

I'll walk through some of the similar data to share with you in terms of the pipeline update, as well as the next steps for the program. ENPP1, moving to our second program in rare diseases. REV102 is an orally bioavailable small molecule ENPP1 inhibitor being developed jointly with Rallybio for hypophosphatasia, which is a rare metabolic bone disease. I'll leave some of the details, but I'll share more in terms of our, how we leverage our platform in designing a highly selective ENPP1 inhibitor, potential first and best-in-class, and also some of the pre-clinical data that we have seen recently for this compound. And LSD1 lastly, as noted earlier, we are strategically pausing development, a potential best in class CNS penetrant, reversible LSD1 inhibitor to ensure some of the internal and external data that would be important to have a competitive TPP.

我將介紹一些類似的數據，以便與大家分享有關管道更新以及該計劃的後續步驟的資訊。ENPP1，轉向我們的第二個罕見疾病計畫。REV102 是一種口服生物可利用的小分子 ENPP1 抑制劑，正在與 Rallybio 聯合開發用於治療低磷酸血症，這是一種罕見的代謝性骨病。我將保留一些細節，但我將分享更多關於我們如何利用我們的平台設計高選擇性 ENPP1 抑製劑、潛在的首創和同類最佳藥物的信息，以及我們最近看到的有關這種化合物的一些臨床前數據。最後，如前所述，我們正在策略性地暫停開發一種潛在的最佳中樞神經系統滲透劑、可逆性 LSD1 抑制劑，以確保獲得一些對於具有競爭力的 TPP 至關重要的內部和外部數據。

We may pick this program up later, depending on how some of this data evolves. Now, moving to some of the programs that we are deprioritizing. So, let's start with NF2. After a thorough review of the clinical data by the NF2 team, we believe that the decision to discontinue further development is clear. Although Phase II for NF2 related meningiomas technically passed the futility threshold, and this is what we were waiting for.

我們可能會稍後選擇該程序，具體取決於這些數據的發展。現在，我們來談談一些我們正在降低優先順序的項目。那麼，讓我們從 NF2 開始。在 NF2 團隊徹底審查了臨床數據後，我們認為停止進一步開發的決定是明確的。儘管NF2相關腦膜瘤的II期臨床試驗從技術上來說已經通過了無效性閾值，但這正是我們所期待的。

It was primarily driven by the lower 40 milligram core. The 60 milligram and the combined dose arms did not pass the futility criteria. And the point I want to emphasize the most is we observed limited tumor shrinkage and clinical activity across both 40 and 60 milligram arms. The next program, CCM. For CCM, we initially reported top line data from the Phase II SYCAMORE trial, and we were pleased to see that it was safe and well tolerated.

它主要由下部 40 毫克核心驅動。60 毫克和合併劑量組未通過無效性標準。我最想強調的一點是，我們觀察到 40 毫克和 60 毫克組的腫瘤縮小和臨床活動有限。下一個節目，CCM。對於 CCM，我們最初報告了 II 期 SYCAMORE 試驗的頂線數據，我們很高興地看到它是安全的並且耐受性良好。

While the early data did show some promising trends, potentially in exploratory efficacy endpoints at 400 milligrams, this was both for lesion volume and the Rankin score. There were negative trends in the efficacy of 200 but these signals and data were not statistically significant. One of the things we noted in terms of next steps was looking into the LTE as well as additional regulatory engagement. So, what are the new findings from the long-term extension studies, we do not see promising trends in MRI, a decline or functional outcomes. And I want to emphasize a couple of things.

雖然早期數據確實顯示出一些有希望的趨勢，可能在 400 毫克的探索性療效終點方面，但這既適用於病變體積，也適用於 Rankin 評分。200 的效力存在負面趨勢，但這些訊號和數據並不具有統計意義。關於下一步，我們注意到的一件事是研究 LTE 以及額外的監管參與。那麼，長期擴展研究的新發現是什麼，我們沒有看到 MRI 的良好趨勢、下降或功能結果。我想強調幾點。

One that we were paying attention to a placebo to 400 milligram crossover, where each patient served as their own baseline. We did not see any trends there, across any of the end points. And for 400 milligrams to 400 milligrams arm, we did not see the continuation of prior trends. And further, it was not distinguishable from natural history. So unfortunately, based on the totality of the data, and this was important to actually have an LTE for a first in disease program to give us confidence.

我們關注的是安慰劑與 400 毫克交叉的情況，其中每個患者都作為自己的基線。我們沒有看到任何趨勢，無論是在哪個終點。對於 400 毫克至 400 毫克的手臂，我們沒有看到先前趨勢的延續。而且，它與自然史沒有區別。因此不幸的是，根據全部數據，對於疾病治療計畫而言，實際上擁有一個 LTE 來給予我們信心是非常重要的。

We looked at the totality of the data and it supports the discontinuation. And now turning to C. diff from a platform perspective, we saw novel insight. This is a new mechanism of action in terms of how we can tackle recurrent C. diff, a highly potent and orally viable C.

我們查看了全部數據，結果支援停止該操作。現在從平台角度轉向 C. diff，我們看到了新穎的見解。就我們如何對抗復發性艱難梭菌（一種強效且可口服的艱難梭菌）而言，這是一種新的作用機轉。

diff toxin, B selective inhibitor. However, as with any of these programs, we're constantly tracking the external landscape, and the recurrent rate, aka the unmet need of C. diff with some of these programs that are further along, is now reducing to almost around 5%. And without a clear differentiation, this comes to the point we made earlier, without a clear differentiation profile, we've decided not to pursue further development, internal development, and take those precious dollars to double down on areas where we have scientific, commercial and technical promise. So, I just want to round by saying for our clinical trial transparency policy, we intend to make all clinical data publicly available in a peer reviewed journal, following appropriate review.

diff 毒素，B 選擇性抑制劑。然而，與任何此類項目一樣，我們也在不斷追蹤外部形勢，復發率（即一些進展較快的項目中艱難梭菌未滿足的需求）現已降至約 5% 左右。如果沒有明確的差異化，這就回到了我們之前提出的觀點，如果沒有明確的差異化概況，我們決定不再進行進一步的發展，不再進行內部發展，而是把寶貴的資金投入到我們有科學、商業和技術前景的領域。因此，我只想總結一下，根據我們的臨床試驗透明度政策，我們打算在經過適當的審查後，將所有臨床數據在同行評審期刊上公開。

And a huge, huge thank you to all of the investigators and patients for supporting us, for being part of the studies. And research and development is one of the hardest things we do to make new medicines for patients, and it's with deep gratitude, thank you for being part of our journey. So, I just want to take a moment, we talked about go-forward programs. We talked about programs from where we are making data driven and strategic decisions. What is that go-forward portfolio strategy as a learning organization?

衷心感謝所有研究人員和患者對我們的支持和參與。為了為患者研發新藥，研發是我們最困難的工作之一，我們懷著深深的感激之情，感謝您參與我們的旅程。所以，我只想花一點時間，我們討論一下前進的計畫。我們討論了基於數據做出戰略決策的計劃。作為一個學習型組織，未來的投資組合策略是什麼？

It has to be grounded in scientific rigor and disciplined capital allocation. This reflects a raised bar for what we choose to pursue, programs that meet the highest standards of differentiation, address significant unmet need and leverage the unique strength and the full power of the Recursion 2.0 platform. So, as you can see, we will, we are evolving to a more focused, product oriented strategy, leveraging the full power of the Recursion 2.0 platform. And that tech stack is not just biology, not just chemistry, but also in clinical and I speak to that more as we go through some of the programs. Number 2, we're focusing, and we'll continue to focus on medicines for patients that are differentiated at the time of launch.

它必須以科學嚴謹性和嚴格的資本配置為基礎。這反映了我們選擇追求的目標的提高，即滿足最高差異化標準的項目，解決重大未滿足的需求，並充分利用 Recursion 2.0 平台的獨特優勢和全部功能。因此，正如您所看到的，我們將發展更專注、以產品為導向的策略，充分利用 Recursion 2.0 平台的全部功能。而且該技術堆疊不僅涉及生物學、化學，還涉及臨床，在我們進行一些專案時，我會更多地談論這一點。第二，我們專注於並將繼續專注於在上市時針對特定患者提供差異化的藥物。

Those can be first in class, those can be best in class. And third, the how; we are applying discipline execution through rapid, data driven and resource efficient go, no go, decisions. In house, AI driven design, really important and clear, differentiated target product profiles to accelerate a proof of concept and maximize R&D resource and time efficiency. So, with that disciplined framework, what comes next? Let's go to the catalyst slides.

這些可以是班上的第一名，也可以是班上最好的。第三，如何；我們透過快速、數據驅動和資源高效的決策來執行紀律。在內部，人工智慧驅動的設計非常重要且清晰，差異化的目標產品配置可以加速概念驗證並最大限度地提高研發資源和時間效率。那麼，有了這個嚴謹的框架，下一步該怎麼做呢？我們來看看催化劑幻燈片。

We anticipate meaningful readouts and catalyst across our internal pipeline in 2025 and 2026. So the first half, as I mentioned earlier, the initiation of our combination study in advanced tumors, building on the monotherapy insights from December of last year for CDK7. Second half of this year, we will have additional Phase I data from the monotherapy as that program matures in the second half for CDK7. In addition for FAP we also have additional patients enrolling in our 4 milligram core, so we'll have additional data as well. And as I mentioned with PI3K we expect development candidate nomination the second half of this year, so we can start the all-important IND enabled studies for this important program.

我們預計 2025 年和 2026 年我們的內部管道將會出現有意義的讀數和催化劑。因此，正如我之前提到的，上半年我們啟動了針對晚期腫瘤的聯合研究，以去年 12 月對 CDK7 單一療法的見解為基礎。今年下半年，隨著 CDK7 單一療法計畫日趨成熟，我們將獲得更多 I 期數據。除了 FAP 之外，我們還有其他患者加入我們的 4 毫克核心，因此我們也會獲得更多數據。正如我在 PI3K 中提到的那樣，我們預計今年下半年將獲得開發候選人提名，因此我們可以為這個重要專案啟動至關重要的 IND 研究。

Transitioning to next year. So, first half of 2026, which is the early safety and PK data readout from our ongoing monotherapy trial in biomarker enriched RBM39 program for solid tumors. And in the second half, similarly early safety and PK update from monotherapy for our MALT1, study in B cell malignancies, as well as Phase I initiation for our ENPP1 inhibitor in hypophosphatasia. So that was the overall view of the portfolio. What I'd like to do now is go through some deep data for FAP, for PI3K, and ENPP1.

過渡到明年。因此，2026 年上半年是我們針對實體腫瘤進行的生物標記富集 RBM39 計劃的單一療法試驗的早期安全性和 PK 數據讀數。而在下半年，同樣是針對 MALT1 單一療法的早期安全性和 PK 更新、B 細胞惡性腫瘤研究，以及針對低磷酸酯酶症的 ENPP1 抑制劑的 I 期啟動。這就是投資組合的整體情況。我現在想做的是研究 FAP、PI3K 和 ENPP1 的一些深度資料。

I'll also ensure that you have a one pager latest and greatest update on all of the other programs, clinical and pre clinical, and then I'll round it out with some of our latest updates on partnerships. All right FAP. Next slide, please. Lots of words. I'm not going to go through all the words on the slides, but I think the most important elements are, this is an allosteric MEK1/2 inhibitor developed for FAP.

我還將確保您擁有關於所有其他項目（臨床和臨床前）的最新和最好的更新，然後我將用一些關於合作夥伴關係的最新更新來完善它。好的 FAP。請看下一張投影片。很多話。我不會詳細介紹投影片上的所有內容，但我認為最重要的元素是，這是針對 FAP 開發的變構 MEK1/2 抑制劑。

FAP is a rare, inherited condition with no FDA approved therapies. Next slide, we leveraged, how did we leverage the Recursion OS platform? We analyzed cellular models of APC gene loss, which is the root cause of FAP, and we identified MEK1/2 inhibition as a novel therapeutic intervention of this mutation. This insight drove the discovery of REC-4881, which is a molecule that we in licensed. And next step was to see did this insight, which was unknown before play out in preclinical.

FAP 是一種罕見的遺傳性疾病，尚無 FDA 批准的治療方法。下一張投影片，我們利用了，我們如何利用 Recursion OS 平台？我們分析了 APC 基因缺失的細胞模型，這是 FAP 的根本原因，並且我們確定 MEK1/2 抑制是這種突變的一種新的治療幹預措施。這項見解推動了 REC-4881 的發現，這是一種我們獲得許可的分子。下一步是看看這種見解是否能夠在臨床前發揮作用之前發揮作用，這是未知的。

So in preclinical models, REC-4881 demonstrated significant reductions in both polyp count and high-grade adenomas, outperforming celecoxib, which is one of the off-label drugs that are used today and not approved to date. In addition to that, we then decided to go into clinical studies. If we go to the next slide. The trial is a 2-part study, evaluating REC-4881 in patients with FAP. First, starting with a Phase Ib safety run-in that you're seeing here, which is 4 milligram and placebo and then for the next click advancing into an open-label Phase II Signal-Seeking study.

因此，在臨床前模型中，REC-4881 顯示出息肉數量和高級別腺瘤的顯著減少，優於塞來昔布，塞來昔布是目前使用但迄今尚未獲得批准的非上市藥物之一。除此之外，我們也決定進行臨床研究。如果我們轉到下一張投影片。本試驗分為兩部分，評估 REC-4881 對 FAP 患者的療效。首先，從您在此處看到的 Ib 期安全性試驗開始，即 4 毫克和安慰劑，然後下一步進入開放標籤 II 期訊號尋求研究。

To reduce class-related side effects, which are seen with MEK1/2 inhibitor, the Phase II portion was refined to enroll patients 55 and over. The ongoing Phase II portion is evaluating 2 once-daily oral doses of REC-4881. For the 4 milligram, the primary endpoints included safety, tolerability and preliminary efficacy. And the main efficacy point is really percent change in polyp burden after 12 weeks of treatment. A follow-up endoscopy at week 25 following a 12-week off-treatment period is used to evaluate the durability of the response.

為了減少 MEK1/2 抑制劑類藥物相關的副作用，II 期臨床試驗進行了改進，招募 55 歲及以上的患者。正在進行的第二階段研究正在評估 REC-4881 的每日兩次口服劑量。對於 4 毫克，主要終點包括安全性、耐受性和初步療效。主要療效點實際上是治療 12 週後息肉負擔的百分比變化。經過 12 週的停藥期後，在第 25 週進行內視鏡隨訪，以評估反應的持久性。

And as of March 17, 2025, data cutoff, 6 patients were efficacy evaluable in the 4-milligram arm. And this is going to be the focus of our clinical efficacy data. But before we go into efficacy, I'd like to cover the safety element. So, when you look at the Phase Ib and Phase II, the data you're seeing here is among the 19 safety evaluable patients, I want to show the entire cohort that received 4 mg across both phases. The most frequent treatment-related AEs were grade 1 and 2 with grade 3 being about 16%, and we did not have any grade 4 or above.

截至 2025 年 3 月 17 日資料截止，4 毫克組有 6 位患者的療效可評估。這將是我們臨床療效數據的重點。但在討論功效之前，我想先談談安全性因素。因此，當您查看 Ib 期和 II 期時，您在此處看到的數據是 19 名可進行安全性評估的患者的數據，我想顯示在兩個階段均接受 4 毫克治療的整個隊列。最常見的治療相關不良事件為 1 級和 2 級，其中 3 級約佔 16%，且未出現 4 級或以上不良事件。

Treatment-related AEs reported to date were mostly rash, diarrhea, and there were some left ventricular ejection fraction. When we looked at the Phase II portion, the most commonly related treatment-related AEs were still rash, CPK and diarrhea. And I will say that rash and the decreased left ventricular ejection fraction are both consistent with the reported safety profile of approved MEK1/2 inhibitors and/or class effect. The LVEF did not lead to any discontinuation in the study. Now let's look at the efficacy.

迄今報告的治療相關不良事件大多為皮疹、腹瀉，並有左心室射血分數降低。當我們觀察第二階段時，最常見的治療相關不良事件仍然是皮疹、CPK 和腹瀉。我想說的是，皮疹和左心室射血分數降低都與已批准的 MEK1/2 抑制劑報告的安全性特徵和/或類效應一致。LVEF 並未導致研究中斷。現在我們來看看功效。

Again, I want to preface by saying these are preliminary results and of 6 efficacy evaluable patients. The distribution of polyp burden changes across all efficacy evaluable patients is shown in the waterfall. At the time of data cutoff, Phase II data shows that the 4 milligrams led to a preliminary 43% median reduction in polyp burden in post-colectomy patients that are 55 or over. 5 of 6 patients, about 83%, experienced reduction ranging from 31% to 82%. However, one patient did show a substantial increase in polyp burden number from baseline, which has been shown in prior studies as well, something we need to, of course, investigate and understand further.

再次，我想先說明一下，這些是初步結果，涉及 6 名可評估療效的患者。瀑布圖顯示了所有可評估療效的患者的息肉負擔變化分佈。數據截止時，第二階段數據顯示，4毫克劑量可使55歲或以上結腸切除術後患者的息肉負擔初步減少43%。 6 名患者中，有 5 名（約 83%）的減幅為 31% 至 82%。然而，一名患者的息肉負擔數量確實較基線大幅增加，這在先前的研究也得到了證實，當然，我們需要進一步調查和了解這一點。

3 patients -- 3 of the 6 patients achieved a greater than 50% polyp burden reduction at week 13. Of the 2 patients, 2 maintained a durable greater than 30% reduction even after the off period, the [indiscernible] off period of the drug. If you go to the next slide, we also wanted to look a little bit deeper into the efficacy data. Recall that the patients with FAP often develop polyps throughout the GI tract, so both the upper and lower regions. In the waterfall, you see reductions both in the upper and lower regions with median reductions 50% or higher.

3 名患者－6 名患者中有 3 名在第 13 週實現了息肉負擔減少 50% 以上。在這 2 名患者中，有 2 名即使在停藥期（即 [音頻不清晰] 停藥期）之後仍能維持 30% 以上的持續減幅。如果您翻到下一張投影片，我們也想更深入了解功效數據。回想一下，患有 FAP 的患者通常整個胃腸道都會長出息肉，因此上部和下部區域都會長出息肉。在瀑布圖中，您會看到上部和下部區域的減少量均達到或超過 50%。

These were encouraging and suggest that there's clinical activity across both anatomical sites. We also looked into an important set of endpoints. So, I talked about polyp burden reduction, but also polyp count and Spigelman stage, which is an important classifier of disease severity and the potential for cancer risk down the road. So, the table shown here summarizes Spigelman stage of screening and again at week 13, highlighted in green, you can see about 3 out of the 6 patients that are on the 4-milligram arm experienced a reduction in Spigelman stage with 2 patients showing a full change of about a scale of 2. While this is early data and of course, any time you do Spigelman, it's potentially confounded by biopsy sampling, it is worth noting that prior pivotal FAP trials have included changes in Spigelman stage as part of the composite endpoint to track FAP-related disease progression.

這些都是令人鼓舞的，顯示兩個解剖部位都有臨床活動。我們也研究了一組重要的端點。因此，我談到了息肉負擔的減少，以及息肉數量和 Spigelman 分期，這是疾病嚴重程度和未來癌症風險可能性的重要分類指標。因此，此處顯示的表格總結了 Spigelman 篩檢分期，並且在第 13 週再次以綠色突出顯示，您可以看到 4 毫克組的 6 名患者中約有 3 名患者的 Spigelman 分期有所減少，其中 2 名患者的分期發生了大約 2 級的完全變化。雖然這是早期數據，當然，任何時候進行 Spigelman 試驗，都可能因活檢取樣而產生混淆，但值得注意的是，先前的關鍵 FAP 試驗已將 Spigelman 階段的變化作為綜合終點的一部分，以追蹤 FAP 相關的疾病進展。

So, in terms of summary and next steps, look, taken together, and I will point your attention to the left-hand side first, the initial Phase II preliminary data does show a consistency of insight from the platform, bolstered by what we see in vivo and now encouraging early clinical signals. As we look ahead, enrollment for this Phase II study is ongoing and we expect to share additional data with efficacy and safety in the second half of 2025. I'll now move on to our PI3K H1047 mutant selective program. So, REC-7735 is a highly selective PI3K H1047, a mutant selective inhibitor. 30% to 40% of HR-positive breast cancers have PI3K mutations with the H1047R mutation in the kinase domain being the most common, about 14% of HR+ breast cancer.

因此，就總結和後續步驟而言，請看一下，總的來說，我首先請您注意左側，初始 II 期初步數據確實顯示了平台洞察力的一致性，並得到了我們在體內看到的以及現在令人鼓舞的早期臨床信號的支持。展望未來，此 II 期研究的招募工作正在進行中，我們預計將在 2025 年下半年分享更多有關療效和安全性的數據。我現在將轉到我們的 PI3K H1047 突變體選擇程式。因此，REC-7735 是一種高度選擇性的 PI3K H1047，一種突變選擇性抑制劑。 30% 至 40% 的 HR 陽性乳癌具有 PI3K 突變，其中激酶結構域中的 H1047R 突變最為常見，約佔 HR+ 乳癌的 14%。

So about anywhere from 9,000 to 11,000 patients in the U.S. and EU. This molecule was engineered, as I mentioned before, with 100x selectivity over wild type, and demonstrates strong CNS penetration and from the data we've seen so far, a low risk of metabolic AEs like hyperglycemia. I'll walk you through some of the data that's showing superior efficacy to PI3 and also Capi, an AKT inhibitor with synergy at low doses when combined with CDK4/61 inhibitors and SERD. The study is currently in candidate profiling with a nomination to DC, as I mentioned before, the second half of this year.

因此，美國和歐盟大約有 9,000 到 11,000 名患者。正如我之前提到的，這種分子經過精心設計，比野生型具有 100 倍的選擇性，並表現出強大的中樞神經系統滲透性，而且從我們目前看到的數據來看，發生高血糖等代謝不良事件的風險較低。我將向您介紹一些數據，這些數據顯示出比 PI3 和 Capi 更優異的療效，Capi 是一種 AKT 抑製劑，與 CDK4/61 抑製劑和 SERD 結合使用時，低劑量即可產生協同作用。這項研究目前正在進行候選人分析，並提名華盛頓特區議員，正如我之前提到的，將於今年下半年進行。

So, let's just take a second in terms of how does the platform -- what is the platform insight here? We started by applying molecular dynamics to characterize the flexibility of the mutant PI3K alpha protein. It allowed us to capture key confirmational snapshots that reveal cryptic mutant-specific binding pockets. These insights formed our SAR strategy early on, which is really important, giving us a more precise road map for design. From there, we use AI/ML models to prioritize chemical space predicted to optimize potency and selectivity while also accounting for critical drug-like features such as physiochemical properties and [ AME ].

那麼，讓我們花一點時間來了解一下平台是如何運作的——這裡的平台洞察力是什麼？我們首先應用分子動力學來表徵突變體 PI3K α 蛋白的靈活性。它使我們能夠捕獲揭示隱密的突變體特異性結合口袋的關鍵確認快照。這些見解很早就形成了我們的 SAR 策略，這非常重要，為我們提供了更精確的設計路線圖。在此基礎上，我們使用 AI/ML 模型對預測的化學空間進行優先排序，以優化效力和選擇性，同時考慮關鍵的類藥物特徵，例如物理化學性質和[ AME ]。

At any point, I'd love to go through more detail on this. This tightly integrated AI-driven approach enabled us to progress our target concept to differentiated candidate in 18 months. Again, another example of this learning platform and how we're trying to get better molecules designed faster. Now if we look to the next slide, which is focused on some of the preclinical data that I just mentioned. Chris mentioned this slide and some of this data in the prior earnings call.

無論何時，我都樂意詳細地討論這個問題。這種緊密整合的人工智慧驅動方法使我們能夠在 18 個月內將目標概念發展為差異化候選人。再一次，這是該學習平台的另一個例子，以及我們如何嘗試更快地設計出更好的分子。現在我們來看下一張投影片，它重點介紹了我剛才提到的一些臨床前數據。克里斯在先前的財報電話會議上提到了這張投影片和一些數據。

So, REC-7735 shows a dose-dependent tumor regression, in the CDX models, the PI3K1047R CDX models compared to a better than standard of care PI3K inhibitors. We have PI3K here, which is one of the first-generation wild-type the [ scorpion ] and also the Loxo compound. This supports a differentiated safety profile while maintaining efficacy, we do not see an increase in hyperglycemia markers, as you can see to the right side. Now if you go to the next slide, I just want to share some of the newer data head-to-head with Capi. Here, you see REC-7735 at 18.7 mgs per kg, which is a medium dose, significantly outperformed high-dose Capi in tumor regression.

因此，REC-7735 顯示出劑量依賴性腫瘤消退，在 CDX 模型中，PI3K1047R CDX 模型與標準護理 PI3K 抑制劑相比效果更佳。我們這裡有 PI3K，它是第一代野生型 [蝎子] 之一，也是 Loxo 化合物。這在保持療效的同時支持了差異化的安全性，我們沒有看到高血糖標記的增加，正如您在右側看到的那樣。現在，如果您翻到下一張投影片，我只想與 Capi 直接分享一些較新的數據。在這裡，您可以看到 REC-7735 的劑量為 18.7 毫克/公斤，這是一種中等劑量，在腫瘤消退方面明顯優於高劑量 Capi。

Even the low dose 6.25 mg per kg -- per kilogram dose was comparable to Capi. And in terms of tolerability, we did not have any weight loss observed, which we have seen with some of the other agents to date in these mouse models. We also looked at some additional data in combination with endocrine therapy. REC-7735 significantly enhances the effect of Fulvestrant, CDK4/6 as well as SERD inhibitors in the CDX models. At low doses, REC-7735 alone was more effective than the standard of care combo.

即使是每公斤 6.25 毫克的低劑量也與 Capi 相當。就耐受性而言，我們沒有觀察到任何體重減輕，而到目前為止，我們在這些小鼠模型中使用其他一些藥物時已經看到了這種情況。我們也研究了一些與內分泌療法結合的額外數據。REC-7735 顯著增強了氟維司群、CDK4/6 以及 SERD 抑制劑在 CDX 模型中的作用。在低劑量下，單獨使用 REC-7735 比標準治療組合更有效。

And when it's added, we see some further deepening of response. So again, preclinical early data, but a rationale -- compelling rationale for REC-7735 in combination regimens for HR+ HER2-negative breast cancer. So next steps, as I mentioned, again, we see that arm, right? The biological insight, but I would say much more focused on the biological insight in terms of greater selectivity by going after the most common mutation in PI3K, the design elements that I just mentioned, novel scaffold and then also some of the in vivo data that we're seeing, which is encouraging for us as a potential best-in-class therapy. DC nomination is planned for second half 2025 with again, the goal of addressing a clear patient need in a genetically defined population.

當它被添加時，我們看到響應進一步加深。因此，再次強調，這是臨床前早期數據，但有理由相信——REC-7735 在 HR+ HER2 陰性乳癌聯合治療方案中具有令人信服的理由。那麼接下來的步驟，正如我再次提到的，我們再次看到了那隻手臂，對嗎？生物學見解，但我想說的是，我們更專注於生物學見解，即透過追蹤 PI3K 中最常見的突變、我剛才提到的設計元素、新穎的支架以及我們看到的一些體內數據，從而實現更高的選擇性，這對我們來說是一種潛在的最佳療法。DC 提名計畫於 2025 年下半年進行，目標同樣是滿足基因定義族群中的明確病患需求。

Next, I will share a little bit more data about our ENPP1 inhibitor shifting to rare disease. REV102 is a potent highly selective ENPP1 inhibitor we're developing with our partner, RallyBio for HPP, a rare debilitating bone disorder with limited treatment options, particularly for adults. The treatment options today are injectables, 3 to 6 injectables a week, and that is challenging. And so we think that there is a significant patient population that would benefit from a first oral non-immunogenic disease-modifying therapy that can reduce the burden and cost of lifelong enzyme replacement therapies, which is the current standard. So again, platform insight here.

接下來，我將分享更多關於我們的 ENPP1 抑制劑轉向罕見疾病的數據。REV102 是一種強效、高選擇性 ENPP1 抑制劑，我們正在與合作夥伴 RallyBio 共同開發用於治療 HPP，HPP 是一種罕見的衰弱性骨病，治療選擇有限，尤其對於成年人而言。目前的治療選擇是注射劑，每週注射 3 到 6 次，這很有挑戰性。因此，我們認為，有相當一部分患者將受益於第一種口服非免疫原性疾病改良療法，這種療法可以減輕終身酵素替代療法的負擔和成本，而這是目前的標準。再次強調，這裡是平台洞察。

This is another example of the Recursion OS delivering targeted innovation. We identified structural insights into ENPP1 and used our generative AI design and ML-driven optimization to create novel samples with high safety margins and oral dosing potential. This was not easy. This is not for the faint of heart. This took a lot of reps.

這是 Recursion OS 實現有針對性創新的另一個例子。我們確定了 ENPP1 的結構見解，並使用我們的生成式 AI 設計和 ML 驅動的優化來創建具有高安全裕度和口服給藥潛力的新型樣本。這並不容易。這不適合膽小的人。這需要多次重複。

But by modeling human [indiscernible] early, we built confidence in the clinical profile before nominating a candidate. This is how we want to front-load some of the risks in our now go-forward and portfolio approach strategy. So, looking at some of the preclinical data, so I'll show you 2 slides on this. The first one is REV102 in early onset HPP models. In an unpublished early onset knockout model, REV102 significantly extended survival.

但透過早期對人類[音訊不清晰]進行建模，我們在提名候選人之前就對臨床概況建立了信心。這就是我們希望在目前的前進和投資組合策略中預先承擔一些風險的方式。因此，請查看一些臨床前數據，我將向您展示兩張投影片。第一個是早發性 HPP 模型中的 REV102。在未發表的早發性敲除模型中，REV102 顯著延長了存活期。

So, you can see on the left-hand side, survival lines for the early onset HPP models. And when you look to the right, you will see also reduced PPI levels, which is a known biomarker that is critical for bone mineralization and restored bone density close to wild-type levels. You will see on the chart to the very right, the ALPL knockout group was not shown given all of the mice guide around the mean age. These data suggests the compound can address both biochemical and skeletal aspects of HPP in early onset models. But we also wanted to look at late onset models, right, which more closely mimics adult HPP.

因此，您可以在左側看到早發性 HPP 模型的存活線。當您向右看時，您還會看到 PPI 水平降低，這是已知的生物標記物，對於骨礦化和恢復接近野生型水平的骨密度至關重要。您會在最右邊的圖表上看到，由於所有老鼠的平均年齡都差不多，所以沒有顯示 ALPL 敲除組。這些數據表明該化合物可以解決早發性模型中 HPP 的生化和骨骼方面的問題。但我們還想研究晚發模型，對吧，它更接近模擬成人 HPP。

And as you can see to the left-hand side, corrective key skeletal defects have normalized the patella structure. And when you look at the right-hand side, very consistency, it also -- we also observed a clear dose-dependent reduction in plasma PPI, further validating the mechanism. So, lots more work to do in this space, but we see, again, the arc here for the biology design in vivo with Phase I expected to initiate in the second half of 2026. We see a compelling opportunity to address unmet needs in juvenile and adult-onset disease where access, convenience and long-term tolerability for chronic disease that starts early, especially [indiscernible]. So those are the 3 programs where I shared a deep dive.

正如您在左側看到的，矯正關鍵骨骼缺陷已經使髕骨結構恢復正常。當您查看右側時，會發現非常一致的結果——我們還觀察到血漿 PPI 明顯呈現劑量依賴性減少，進一步驗證了該機制。因此，在這個領域還有很多工作要做，但我們再次看到了體內生物設計的弧線，第一階段預計將於 2026 年下半年啟動。我們看到了解決青少年和成人發病疾病中未滿足需求的絕佳機會，特別是對於早期發病的慢性疾病，其可及性、便利性和長期耐受性[音頻不清晰]。這就是我深入分享的 3 個程式。

Now I'm going to do a quick tour, one pager to some of the other oncology go-forward programs. So, the first one, RBM39 degrader, identified, again, as I mentioned earlier, through our phenotypic platform, it mimics closely CDK12 loss phenotypically and induces in our preclinical models, as you can see in the middle, dose-dependent antitumor activity in preclinical models. This is one place where I want to emphasize an area of platform capability build, you'll see across all of these programs, which is precision selection of patients. This is where clin tech and causal AI efforts that we're really doubling down on in our platform is helping us accelerate for the right patients, but then also site selection and enrollment. The study right now is in Phase I monotherapy with an update expected first half of 2026.

現在我將進行一次快速瀏覽，透過一頁介紹其他一些腫瘤學進展項目。因此，正如我之前提到的，第一個 RBM39 降解劑是透過我們的表型平台識別的，它在表型上與 CDK12 缺失非常相似，並在我們的臨床前模型中誘導，正如您在中間看到的那樣，在臨床前模型中具有劑量依賴性的抗腫瘤活性。在這裡我想強調一下平台能力建設的一個方面，您會在所有這些項目中看到，那就是對患者的精準選擇。這就是我們在平台上加倍投入的臨床技術和因果人工智慧努力，它不僅幫助我們加速找到合適的患者，還幫助我們選擇地點和招募患者。該研究目前處於 I 期單一療法階段，預計將於 2026 年上半年完成更新。

CDK7, I've mentioned a little bit about CDK7 before, how we leverage our platform to optimize the PK/PD and therapeutic [indiscernible]. In the middle, you see potent tumor regression. We're not showing some of the data, the clinical data from December. You can find it online, but we also saw one complete PR, confirmed PR, multiple stable diseases as well. Now what we're doing is we're using causal AI and human genetics with some of the [ temp ] data and then also some of the capabilities we're building in-house and cell line panels.

CDK7，我之前提到過一些關於 CDK7 的內容，我們如何利用我們的平台來優化 PK/PD 和治療[音頻不清晰]。在中間，你可以看到明顯的腫瘤消退。我們沒有顯示一些數據，即 12 月的臨床數據。您可以在網路上找到它，但我們也看到了一個完整的 PR、確認的 PR 以及多個穩定的疾病。現在我們正在做的是將因果人工智慧和人類遺傳學與一些[臨時]數據一起使用，然後也使用我們正在建立的一些內部和細胞系面板的功能。

So, we do both, the best of both worlds to guide precision indication expansion with monotherapy escalation ongoing and combination, as I said, initiating first half of 2025. And MALT1, I won't go through the details of MALT1 that I shared before, but improved safety and efficacy profile through structure-based design and hotspot analysis showed both single agent and synergistic activity. You can see in the middle in vivo data and durable response. We're also -- this is a competitive space. So, we're using advanced RWE analytics, clin tech approaches to accelerate recruitment.

因此，我們採取兩種方法，即兩全其美，透過單一療法的持續升級和聯合療法來指導精準適應症的擴展，正如我所說的，從 2025 年上半年開始。至於 MALT1，我不會詳細介紹我之前分享過的 MALT1，但透過基於結構的設計和熱點分析提高的安全性和有效性概況顯示了單一藥劑和協同活性。您可以在中間看到體內數據和持久響應。我們也是──這是一個競爭激烈的領域。因此，我們正在使用先進的 RWE 分析和臨床技術方法來加速招募。

Already, we've identified, in a matter of days, 50 new high potential trial sites in the U.K. and Spain to support efficient steady execution. So back to our principles from before, we can have faster P&C. And I'll wrap it up by talking a little bit about our partner programs. I know we'd shared the milestones, we shared the upfronts, et cetera.

短短幾天內，我們就已在英國和西班牙確定了 50 個新的高潛力試驗地點，以支持高效穩定的執行。回到我們之前的原則，我們可以擁有更快的 P&C。最後，我將簡單談談我們的合作夥伴計畫。我知道我們分享了里程碑、分享了前期工作等等。

But how are we? What is the nature of this partnership scientifically? These partnerships reflect tackling complex targets and generating high-quality molecules across a range of diseases and also reflects strong external validation of our approach, an additional opportunity for us to learn and create medicines, as Chris was saying earlier, and value for the company. So let's start with Sanofi. We achieved 4 milestones to date with Sanofi in a partnership that's about 3 years to date.

但我們怎麼樣？從科學角度來看，這種夥伴關係的本質是什麼？這些合作關係體現了我們針對一系列疾病解決複雜目標和生成高品質分子的努力，也體現了我們的方法得到了強有力的外部驗證，正如克里斯之前所說，這為我們提供了學習和創造藥物的額外機會，也為公司帶來了價值。那我們就從賽諾菲開始吧。迄今為止，我們與賽諾菲的合作已持續約 3 年，已取得 4 個里程碑。

That is multiple challenging targets the team is working on in both immunology and oncology. The collaboration uses the full suite of the Recursion end-to-end platform, in vivo biology, all the way to generative chemistry and active learning to rapidly design and optimize first-in-class and best-in-class compounds. I want to emphasize what's coming up next. Over the next 12 to 18 months, we have development candidate milestones coming up, which is an important milestone, additional milestone for us and a potential opt-in with Sanofi that starts at that phase. With Roche.

這是團隊在免疫學和腫瘤學領域正在研究的多個具有挑戰性的目標。此次合作利用了 Recursion 端到端平台的全套技術，包括體內生物學、生成化學和主動學習，以快速設計和優化一流和最佳化合物。我想強調接下來要發生的事情。在接下來的 12 到 18 個月內，我們將迎來候選開發里程碑，這是一個重要的里程碑，對我們來說是一個額外的里程碑，並且從這個階段開始，我們有可能與賽諾菲達成合作。與羅氏公司合作。

Recursion OS is being used. We've talked about some of the maps, but I just wanted to maybe add a little bit more color. Recursion uses [indiscernible] maps, but Recursion has also developed a slew of disease context-specific maps. For instance, in this partnership, more than 5 pheno-maps and over 5,000 transcriptomes across neuroscience and GI oncology, feeding discovery at a really fast pace. We, as you know, last year in the fall, triggered a 30 million map that was accepted, and then we have more map milestones that are coming.

正在使用遞歸作業系統。我們已經討論了一些地圖，但我只是想添加一點顏色。Recursion 使用 [音訊不清楚] 地圖，但 Recursion 也開發了一系列特定於疾病背景的地圖。例如，在此次合作中，我們獲得了神經科學和胃腸道腫瘤學領域的 5 多個表型圖和 5,000 多個轉錄組，以非常快的速度推動研究發現。如你所知，去年秋天，我們觸發了一張被接受的 3000 萬張地圖，然後我們即將迎來更多的地圖里程碑。

The other piece, and albeit fondly calls it lab in the loop, which we really, really like, model enables not just a tight cycle of AI-driven hypothesis generation, but really important for this partnership, we're pivoting from the maps and the novel biology to now programs, right, experimental validation and the design of the programs to have the potential to make new medicines. So I want to wrap it up by saying we have multiple internal and external pipeline catalysts that are coming up. As you've seen some of the internal ones laid on top, but then also meaningful partnership catalysts with new Phenomap options, program initiations and potential options exercised by some of our partners. So, with that, I'm going to conclude the R&D and pipeline update, both internal and external. Let me close by saying that we are encouraged by the momentum we're seeing, both internally and with our partner programs.

另一部分，儘管我們親切地稱之為“實驗室在環”，但我們非常非常喜歡這個模型，它不僅能夠實現由人工智能驅動的假設生成的緊密循環，而且對於這種合作關係來說非常重要，我們正在從地圖和新型生物學轉向現在的程序，對吧，實驗驗證和程序設計，以有可能製造新藥。所以我想總結一下，我們有多個內部和外部管道催化劑即將出現。如您所見，其中一些內部因素被放在首位，但同時也帶來了有意義的合作催化劑，包括新的 Phenomap 選項、程式啟動以及一些合作夥伴行使的潛在選項。因此，我將結束內部和外部的研發和管道更新。最後，我想說，我們對內部和合作夥伴計劃中看到的勢頭感到鼓舞。

And we remain committed to a disciplined portfolio strategy that prioritizes scientific differentiation, capital efficiency and value creation above all with a clear path to impact. With that, I will turn it over to our CFO, Ben Taylor, for the financial update.

我們將繼續致力於嚴謹的投資組合策略，優先考慮科學差異化、資本效率和價值創造，並制定明確的影響路徑。接下來，我將把財務更新轉交給我們的財務長 Ben Taylor。

Terrific. Thank you, Najat. So, where we wanted to start with when we're going through some of the financials is not only to talk about the pipeline prioritization as Najat just went through and Chris gave a high-level overview at the beginning, but also talk about how we are trying to make data-driven, disciplined decisions across the organization to really maximize our ability to reach all of those milestones that were on the previous slide. So, if you look at how we've been adjusting our operations not only since the merger, but even before and trying to align that to be able to drive our cash runway as long as possible, what you can see is we've really had a focus on adjusting our capacity over our capabilities. And what I mean by that is our capabilities are the platform overall, what we can actually produce out of that platform.

了不起。謝謝你，納賈特。因此，當我們討論一些財務問題時，我們想要開始的不僅僅是討論管道優先級，正如 Najat 剛剛討論過的，以及 Chris 在一開始給出的高層概述，還要討論我們如何嘗試在整個組織內做出數據驅動的、有紀律的決策，以真正最大限度地提高我們實現上一張幻燈片上所有里程碑的能力。因此，如果您觀察我們不僅在合併後，甚至在合併之前就一直在調整運營，並試圖調整運營以盡可能延長現金流，您就會發現，我們確實一直專注於調整產能而不是能力。我的意思是，我們的能力就是整個平台，以及我們實際上可以利用該平台生產什麼。

The capacity would be more of how many can we [indiscernible] and so because we are a tech company, because we focus on automation, we actually have a great ability to adjust our capacity based on the market conditions, based on the pipeline that we want to execute while still being able to enable all of those same capabilities across the platform. And that's exactly what you'll see from us both during the first quarter, but also through the rest of the year. A couple of different points that we wanted to hit on. One, we ended the quarter with $509 million in cash. We will talk a bit about cash burn, and this is something that's really important because it's a little confusing to anyone outside of the company, looking at our financial statements to try and understand what are you actually spending operationally to execute on all of those things that you're doing.

產能更多的是我們能夠生產多少[音頻不清晰]，因為我們是一家科技公司，因為我們專注於自動化，所以我們實際上有能力根據市場情況、根據我們想要執行的管道來調整我們的產能，同時仍然能夠在整個平台上實現所有相同的功能。這正是您將在第一季以及今年剩餘時間內看到的。我們想談幾個不同的觀點。首先，本季末我們擁有 5.09 億美元現金。我們將討論現金消耗問題，這一點非常重要，因為對於公司外部的任何人來說，這都有點令人困惑，他們查看我們的財務報表，試圖了解您在執行所有正在做的事情時實際花費了多少營運資金。

And so, what we've tried to do is key metrics and put them into the context of a cash operational burn. So how are we spending money, excluding the inflows that we get from partnerships, excluding the noncash effects. And so, during the first quarter of '25, as you'll see, that was about $118 million, including all of our cash operating expenses and our capital expenditures, not including anything that was an inflow from our partnerships or financings and excluding transaction costs as a result of the merger. So, in total, we expect a cash runway into mid-2027. Now what's going into that assumption is really based on the 3 different levels of how we drive cash runway.

因此，我們嘗試做的是將關鍵指標置於現金營運消耗的背景下。那麼，除去從合作關係中獲得的流入資金和非現金影響，我們是如何花錢的呢？因此，如您所見，在 25 年第一季度，該金額約為 1.18 億美元，其中包括我們所有的現金營運費用和資本支出，不包括來自合作夥伴關係或融資的任何流入，也不包括合併產生的交易成本。因此，總體而言，我們預計到 2027 年中期會有現金流。現在，這個假設實際上是基於我們如何推動現金流的三個不同層面。

The first is thinking about our partnerships. So we've brought in $450 million from our partners over the [technical difficulty]. In addition, we've also been able to hit on 4 core milestones in the Sanofi partnership over the last 18 months. We hit on a major milestone with Roche as well. And so we are now driving towards continuing to execute on those existing partnerships and really leveraging the cash flows and the partnership milestones available to us there.

首先是考慮我們的合作關係。因此，我們從合作夥伴那裡獲得了 4.5 億美元的資金【技術難度】此外，在過去的 18 個月中，我們還在與賽諾菲的合作中實現了 4 個核心里程碑。我們也與羅氏公司一起創造了一個重要的里程碑。因此，我們現在正致力於繼續執行現有的合作夥伴關係，並真正利用我們現有的現金流量和合作夥伴關係里程碑。

We will continue to look at new business development as well and have the ability to match our operational capacity to the [indiscernible] In addition, as we've historically done, financing is another aspect that we will look at. We intend to follow the same business patterns that we have in the past on that aspect. Finally, what is completely in our control is our cash burn and being able to adjust our cash burn over time based on what our priorities are inside of the business. And so that's where you've seen us take the steps to be able to [technical difficulty] maximize our possible. So, if we can go on to the next slide, what you can see is a pre and a post in some ways.

我們也將繼續專注於新業務的發展，並有能力將我們的營運能力與[音訊不清晰]相匹配。此外，正如我們過去所做的那樣，融資是我們將關注的另一個方面。在這方面，我們打算遵循過去的相同業務模式。最後，我們完全可以控制的是我們的現金消耗，並且能夠根據我們業務內部的優先事項隨時調整我們的現金消耗。所以，您已經看到我們採取措施，最大限度地提高我們的技術難度。因此，如果我們可以繼續下一張投影片，您可以看到某種前後的對比。

So, on the left side of the page is the 2024, this is a non-GAAP measure, a cash burn from the 2 different organizations. What it breaks down to is about $600 million on a combined basis. The company is excluding all of the partnership and financing. What we're looking at for this year is a budget of less than or equal to $450 million in the same terms. How we are doing that?

因此，頁面左側是 2024 年，這是非 GAAP 指標，是來自兩個不同組織的現金消耗。總計起來約 6 億美元。該公司正在排除所有合作夥伴關係和融資。我們今年的預算預計在同樣的條件下不超過 4.5 億美元。我們是如何做到的？

One, we talked about the pipeline prioritization but we've also been able to reach deep into a number of different corporate expenses, adjusting capacity as I talked about and also trying to leverage the fact that we are a technology platform. We should continue to be able to do more with less because we focus on encoding and automating our processes so that we can accomplish more goals the next year than we did the previous one using less resources. You'll see us continue to drive towards that every chance that we possibly can. With that, I'll turn it back over to Chris.

首先，我們討論了管道優先級，但我們也能夠深入了解許多不同的公司開支，調整容量，正如我所說的，並試圖利用我們是一個技術平台的事實。我們應該繼續能夠用更少的資源做更多的事情，因為我們專注於編碼和自動化我們的流程，以便我們明年能夠使用更少的資源來完成比前一年更多的目標。您會看到，我們會繼續盡一切可能去爭取這個機會。說完這些，我會把話題交還給克里斯。

Thanks, Ben. I want to talk a little bit about Recursion 2.0 and the experiment that we are here to run at Recursion. You've heard from Ben and Najat today. And on behalf of them and the entire team, we thank you for your attention. I just want to share with all of you that we believe that Recursion will continue to lead the tech bio space.

謝謝，本。我想稍微談談 Recursion 2.0 以及我們在 Recursion 上運行的實驗。今天您已經聽到了 Ben 和 Najat 的消息。我代表他們和整個團隊感謝您的關注。我只是想與大家分享，我們相信 Recursion 將繼續引領科技生物領域。

And we're going to do that through the sustainable continued growth plan that we shared today. We're going to remain committed to our internal pipeline, though it's going to be more focused than it has been in the past. We're going to continue to execute on our partnerships, and we believe there are substantial milestones that we have the potential to earn over the coming quarters and years. We're going to continue to increase our focus on leveraging AI, not only in drug discovery, but all the way through development with some really exciting build happening in clinical development that we'll share more on soon. And as Ben just shared, we're going to continue increasing the efficiency of Recursion while also never stopping our investment in the Recursion operating system because ultimately, it's that operating system, that learning system that we believe will give Recursion an advantage in the coming years.

我們將透過今天分享的可持續持續成長計劃來實現這一目標。我們將繼續致力於我們的內部管道建設，儘管我們將比過去更加專注。我們將繼續執行我們的合作夥伴關係，我們相信在未來幾季和幾年內我們有潛力取得重大的里程碑。我們將繼續加強對人工智慧的利用度，不僅在藥物發現方面，而且在整個開發過程中，臨床開發中正在發生一些非常令人興奮的進展，我們很快就會分享更多相關資訊。正如 Ben 剛才所說，我們將繼續提高 Recursion 的效率，同時也不會停止對 Recursion 作業系統的投資，因為最終，我們相信這個作業系統、這個學習系統​​將在未來幾年為 Recursion 帶來優勢。

And so with that, a huge thanks to all of you for your attention. I think we're going to go ahead and turn it over to Q&A.

因此，非常感謝大家的關注。我想我們將繼續進行問答環節。

And I will start. It looks like Eric Joseph at JPMorgan has asked, given your stated runway to mid-2027, what burn rate do you anticipate exiting '25 or entering '26 with? From where would you expect incremental efficiencies still to be derived? And do you plan to raise capital? And for that, I'll turn it over to you, Ben.

我這就開始。摩根大通的 Eric Joseph 似乎問過，考慮到您所述的到 2027 年中期的跑道，您預計在 2025 年結束時或進入 2026 年時燒錢率是多少？您期望從哪裡獲得增量效率？您計劃籌集資金嗎？為此，我將把這個任務交給你，本。

Sure, of course. So, we haven't given specific guidance on the runway, but you can imagine if our budget for this year is $450 million or less, we're targeting a runway of less than that. And so we will give additional detail as the year moves forward. I think we're also going to continue to look for different efficiencies across the organization. And let me give you a couple of examples.

當然可以。因此，我們還沒有給出關於跑道的具體指導，但你可以想像，如果我們今年的預算是 4.5 億美元或更少，我們的目標是跑道長度少於這個數字。因此，隨著時間的推移，我們將提供更多細節。我認為我們也將繼續在整個組織內尋找不同的效率。讓我給你們舉幾個例子。

We've been able to, for example, drive better contracts with our partners a more scaled organization. We've been able to integrate different parts of the business where we had high cost on one side previously and low cost and another, eventually reach a lower cost overall. And we'll continue to drive into every aspect that we can extend that runway without impacting our ability to execute and deliver on our pipeline programs, both internally and with our partnerships. So, we'll keep driving on that. As far as raising capital, as you know, we and no one else give guidance on financing, but we plan to just really continue our previous business practices.

例如，我們能夠與合作夥伴簽訂更好的合同，從而擴大組織規模。我們已經能夠整合業務的不同部分，以前一方面成本高，另一方面成本低，最終實現整體成本降低。我們將繼續努力，在不影響我們內部和與合作夥伴執行和交付管道計劃的能力的情況下，延長跑道。因此，我們會繼續努力。就籌集資金而言，如您所知，我們和其他任何人都不會提供融資指導，但我們計劃繼續我們以前的商業慣例。

We'll continue watching the market. We do have an ATM facility, which we have used moderately in the past. And so we'll continue to use our current business practices.

我們將繼續關注市場。我們確實有 ATM 設施，過去我們曾經適度使用過。因此我們將繼續採用當前的業務實務。

Thank you so much, Ben. Next, we've got James and Joe asking a question on partnerships. When can we see an option in on a molecule candidate from one of your four main partnerships? And any further insights on new levers in the OS for accelerating partnership programs to commercialization? I'll take the first part of that.

非常感謝，本。接下來，詹姆斯和喬將詢問有關合作關係的問題。我們什麼時候可以看到你們四個主要合作夥伴之一的候選分子選項？您對於作業系統中加速合作夥伴計畫商業化的新手段還有什麼進一步的見解嗎？我將討論第一部分。

So we've already had 4 programs optioned in our collaboration with Sanofi, another program options in our collaboration with Roche Genentech. And we believe that those programs and many others coming behind them have the potential not only to get those early options, but perhaps to have the potential to go to later-stage options where they might move into our partners' pipelines. And obviously, the economics are significantly higher at those stages. So we're continuing to do that work, and we think a lot of promising progress so far. For the second part, maybe I'll turn it over to you, Najat.

因此，我們在與賽諾菲的合作中已經選擇了 4 個項目，在與羅氏基因泰克的合作中也選擇了另一個項目。我們相信，這些項目以及隨後的許多其他項目不僅有可能獲得早期選擇，而且可能具有進入後期選擇的潛力，從而可能進入我們合作夥伴的管道。顯然，這些階段的經濟效益要高得多。因此，我們將繼續進行這項工作，我們認為迄今為止已經取得了許多有希望的進展。對於第二部分，也許我會把它交給你，納賈特。

Any further insights on new levers in the OS for accelerating partnership programs to commercialization?

對於作業系統中加速合作夥伴計畫商業化的新槓桿，您還有什麼進一步的見解嗎？

Yes. I mean I'll mention maybe 3. One on biology, we talked a lot about the genomics work that is [indiscernible] doing now adding transcriptomics. I think the clinical genomic data that we have that really helps you make the stack multimodal, but not just to understand holistically the biology, but also very early on, start to better understand what the patient population may be, really creating a more differentiated TPP upfront and earlier on. That's one area.

是的。我的意思是我可能會提到 3 個。在生物學方面，我們討論了很多關於基因組學的工作，現在正在增加轉錄組學。我認為我們擁有的臨床基因組數據確實有助於您實現多模式堆棧，但不僅能從整體上了解生物學，而且在很早的時候就能更好地了解患者群體，真正在早期創建更具差異化的 TPP。那是一個領域。

The second, on the chemistry and the design module, you saw some of the examples I shared for internal. It's very, very similar to what we're doing with partners in terms of can we try to model in or model out aspects that we know are challenging with molecular dynamics, QM, you'll see much more coming up in that space. And then also being able to model and predict some of the AME aspects that makes a drug more drug-like earlier on. And the third is more in development. Chris touched on this a little bit.

第二，在化學和設計模組上，您看到了我為內部分享的一些範例。這與我們與合作夥伴所做的事情非常非常相似，就我們能否嘗試對分子動力學、QM 所面臨的挑戰進行建模或建模而言，您將看到該領域出現更多成果。然後還能夠建模和預測一些使藥物更像藥物的 AME 方面。第三個正在進一步發展。克里斯稍微談到了這一點。

As we partner, whether it's not just on a discovery program perspective, but also on potential partnership on an asset perspective, et cetera, we're also going to leverage some of our clin tech capabilities. Again, using multimodal data to really precisely understand the patient population that we just target that would have the highest signal to noise and then also being more rapid in terms of how [indiscernible].

作為合作夥伴，無論是否只是從發現計劃的角度，還是從資產角度的潛在合作夥伴關係等，我們都將利用我們的一些臨床技術能力。再次，使用多模態數據來真正準確地了解我們剛才瞄準的患者群體，這將具有最高的信噪比，並且在如何[音頻不清晰]。

Thanks, Najat. Next up, we've got Vikram from Morgan Stanley, who's asking a question on the pipeline. Your pipeline prioritization leans heavily towards oncology. Do you generally see a pivot away from rare disease for your pipeline and platform? And if so, which aspects of the Recursion operating system underlying approach do you think make onco stronger fit in addition to rare disease?

謝謝，納賈特。接下來，我們請到了摩根士丹利的維克拉姆，他想就管道問題提問。您的管道優先順序主要傾向於腫瘤學。您是否普遍認為您的管道和平台將不再關注罕見疾病？如果是這樣，您認為遞歸作業系統底層方法的哪些方面使得腫瘤學除了罕見疾病之外更適合？

I think at a high level, we believe that both oncology and rare disease are fantastic areas for us to deploy our platform. In both of those areas, we have some genetic markers that often give us sort of an anchor point of biology from which to work from. And so we'll continue to follow the data. And I can imagine us continuing to drive both rare disease and oncology programs forward. The data is going to be ultimately what drives where -- what the balance of the portfolio looks like, but I do not see us abandoning either oncology or rare disease in the near term.

我認為從高層次來看，我們相信腫瘤學和罕見疾病都是我們部署平台的絕佳領域。在這兩個領域，我們都有某些遺傳標記，它們通常能為我們提供進行研究的生物學錨點。因此我們將繼續關注數據。我可以想像我們將繼續推動罕見疾病和腫瘤學計畫的發展。數據將最終決定投資組合的平衡情況，但我認為我們不會在短期內放棄腫瘤學或罕見疾病。

Next up, we'll go to Dennis from Jefferies, Gil from Needham, Alex from Bank of America, Brendan from Cowen and many other folks who are asking questions around the FAP readout. And I'm going to read these off one by one because we've got a whole bunch here, and I'll have Najat answer them. So the very first one, talk about the FAP data shared at DDW and how is that differentiated from other programs that we may have seen in this space?

接下來，我們將採訪 Jefferies 的 Dennis、Needham 的 Gil、美國銀行的 Alex、Cowen 的 Brendan 以及其他許多就 FAP 讀數提出問題的人士。我將逐一讀出這些問題，因為我們這裡有很多問題，我會讓納賈特來回答。那麼首先，請談談在 DDW 上共享的 FAP 數據，以及它與我們在這個領域可能看到的其他程式有何不同？

Yes. Happy to do that. So, when we look at the FAP data, which I just shared, efficacy valuable about N of 6 patients, median polyp burden reduction in the 40s, 43%. But again, early data, right? And I also talked a little bit about the safety where most of what you're seeing there is on target class effect from MAK1/2 inhibitors.

是的。很高興這樣做。因此，當我們查看我剛剛分享的 FAP 數據時，療效價值約為 6 名患者中的 N 名，息肉負擔中位數減少 40 多歲，即 43%。但同樣，這是早期數據，對嗎？我還談了一點安全性，你們所看到的大部分都是 MAK1/2 抑制劑的目標效應。

The 2 other programs that exist, the Celecoxib, as I mentioned, used off-label and then also another program focused on rapamycin and [indiscernible] rapamycin, which is in a competitor's pipeline. Both so far have shown polyp reductions, 20% to 30%. So just from the primary endpoint that we're looking at. The second piece, I think that's also important to note is the change in the Spigelman scoring. And also, we are encouraged by the congruence that we see polyp burden, polyp count and then also in the Spigelman stage.

現有的另外兩個項目是塞來昔布，正如我所提到的，它是非說明書用途的，還有另一個項目專注於雷帕黴素和[音頻不清晰]雷帕黴素，它是競爭對手的研發線。到目前為止，兩組息肉均減少了 20% 至 30%。因此，僅從我們正在查看的主要終點來看。第二點，我認為同樣值得注意的就是史皮格爾曼得分的變化。此外，我們看到息肉負擔、息肉數量以及 Spigelman 分期的一致性，這讓我們感到鼓舞。

Again, early data, but some of the reductions that we're seeing so far is pretty encouraging. I see there's another question in terms of the non-responders; I'm just taking that [indiscernible]. So, I'm talking through the data more holistically. Non-responder with a 6-fold increase in polyps. What do we see in natural history?

再次，這是早期數據，但我們迄今為止看到的一些減少是相當令人鼓舞的。我看到關於無回應者還有另一個問題；我只是拿了那個[音頻不清晰]。所以，我會更全面地討論數據。對治療無反應，息肉增加 6 倍。我們在自然史中看到了什麼？

So in natural history, the polyp burden is increasing for these patients. But there are prior studies and one of our competitor's studies where about 40% to 50% of patients are non-responders in these studies, right? Nonresponders from a polyp burden or from a polyp count perspective. And recent data has shown that even in that 40% of nonresponders to polyp increase, there is anywhere from 1 to 2 to 6x increase in polyp burden. So for our one nonresponder, as I mentioned earlier, we're going to do -- we're doing a lot of work to better understand the reason for that.

因此，從自然病史來看，這些患者的息肉負擔正在增加。但是先前的研究和我們的競爭對手的研究顯示大約 40% 到 50% 的患者對這些療法沒有反應，對嗎？從息肉負擔或息肉計數角度來看無反應者。最近的數據顯示，即使在 40% 對息肉無反應的患者中，息肉負擔也會增加 1 到 2 到 6 倍。因此，對於我們這個沒有回應的人，正如我之前提到的，我們將做——我們正在做大量的工作來更好地了解原因。

And then the work will continue. As we have a larger end, these numbers will evolve, and that's going to be important as we look for more mature data later on this year.

然後工作將繼續。由於我們的目標範圍越來越廣，這些數字也會隨之變化，這對於我們在今年稍後尋找更成熟的數據至關重要。

The next question on FAP was will we continue to dose higher than 4 milligrams?

關於 FAP 的下一個問題是我們是否會繼續服用高於 4 毫克的劑量？

What we want to do first is we're encouraged by the reduction that we're seeing 30% to 80% polyp burden reduction, that's pretty significant. We want to look at some of the data later this year, [indiscernible] of 10. And then next steps would be either if we need to dose higher, but then also discussions with the regulatory agencies on the potential path forward. So first, we want to complete the 4-milligram cohort, really better understand data and then take next steps from there.

我們首先要做的是看到息肉負擔減少了 30% 至 80%，這讓我們感到鼓舞，這是相當顯著的。我們希望在今年稍後查看一些數據，[音訊不清晰] 佔 10%。下一步是，我們是否需要提高劑量，以及是否需要與監管機構討論潛在的前進方向。因此，首先，我們要完成 4 毫克隊列研究，以更好地了解數據，然後再採取下一步措施。

And then the last 2, where do you see the bar for success in FAP, first in terms of FDA approval, but also as it relates to broader uptake among patients?

最後兩個問題，您認為 FAP 成功的標準是什麼，首先是 FDA 批准，還在於它與病人更廣泛的接受度有關嗎？

Yes. I mean that's a great question. Look, bar for success for FAP, what we see with some of the off-label agents that are used anywhere from 20% to 30%, there's nothing approved. So clearly, there's a huge unmet need for these patients because if not, they're doing multiple surgeries throughout their lifetime. There is another agent that's in just starting Phase III.

是的。我的意思是，這是一個很好的問題。你看，除了 FAP 的成功之外，我們看到一些非說明書用途的藥物的使用率在 20% 到 30% 之間，沒有任何藥物獲得批准。顯然，這些患者存在巨大的未滿足需求，因為如果不滿足，他們一生中就需要進行多次手術。另一個代理商剛開始第三階段的研發。

You can see some of the data. Our polyp burden reduction to date is encouraging and higher, but much more to do in terms of learning about the data.

您可以看到一些數據。到目前為止，我們的息肉負擔減少情況令人鼓舞且更高，但在了解數據方面還有很多工作要做。

And any next steps for the program?

該計劃接下來還有什麼計劃嗎？

Yes. So as I mentioned, more patients on the 4 milligram by the end of this year and then conversations with the FDA on the path forward. So far, the endpoints have been a composite endpoint for FAP. And one of the components, as I mentioned earlier, is actually the inclusion of Spigelman score. So we're watching all of those different aspects and more to come [indiscernible].

是的。正如我所提到的，到今年年底將有更多患者服用 4 毫克劑量的藥物，然後與 FDA 討論未來的發展方向。到目前為止，這些終點都是 FAP 的複合終點。正如我之前提到的，其中一個組成部分實際上是納入了 Spigelman 評分。因此，我們正在關注所有這些不同方面，以及未來更多方面[音頻不清晰]。

Thanks. All right. Next up, we have Alex from BofA, who's asking, remind us how CCM, NF2 and C. diff, which are 3 of the programs that we discontinued were initially discovered or developed and how the refined pipeline strategy maybe better reflects the current capabilities of the Recursion platform. I'll take this one.

謝謝。好的。接下來，我們邀請了來自美國銀行的亞歷克斯，他要求我們回顧我們停止的 3 個專案 CCM、NF2 和 C. diff 最初是如何發現或開發的，以及改進後的管道策略如何更好地反映 Recursion 平台的當前功能。我要這個。

So look, CCM, NF2 and FAP were all products of our Recursion 0.1 platform, where we were using RNAi and tools to identify repurposing candidates. And it's exciting to see the FAP program showing us some preliminary efficacy. Our C. diff program came out of our Recursion 1.0 platform, where we started to explore new chemical entities. And we see no reason to date why the science doesn't continue to hold on that program.

所以，CCM、NF2 和 FAP 都是我們 Recursion 0.1 平台的產品，我們在其中使用 RNAi 和工具來識別重新利用的候選物。我們很高興看到 FAP 計劃顯示出一些初步的功效。我們的 C. diff 計畫源自於我們的 Recursion 1.0 平台，我們從這個平台開始探索新的化學實體。到目前為止，我們還沒有發現科學不繼續堅持該計劃的理由。

That decision today was really based on looking at the commercial landscape and the unmet need and making sure we prioritize our investment. Obviously, and are too low to draw any conclusions, but we designed Recursion as a learning platform where each generation of the platform has a higher probability of identifying and uncovering medicines that we think will have an improved probability of success. And again, then on Recursion 0.1 will be too low versus Recursion 1.0. But as we continue to learn and iterate on this platform, I'm confident that on average, the probability of success of our programs is likely to go up. That's what we're here to do.

今天的決定實際上是基於對商業前景和未滿足的需求的觀察，並確保我們優先考慮我們的投資。顯然，和太低而無法得出任何結論，但我們將 Recursion 設計為一個學習平台，該平台的每一代都有更高的機率識別和發現我們認為成功機率更高的藥物。再一次，那麼遞歸 0.1 與遞歸 1.0 相比就太低了。但隨著我們不斷學習和迭代這個平台，我相信平均而言，我們的專案成功的可能性可能會上升。這就是我們來這裡要做的事情。

And Chris, if I could just add 1 point coming in less than a year ago, just looking at the overall programs, just going back to our go-forward portfolio strategy, one of the things you'll see mentioned is really be very thoughtful in terms of how the molecules are designed, and we can do that in-house today and even more so doubling down on that post the integration with Exscientia. CCM and NF2 and [ FPDs ] were all in-license/repurpose programs. So that's number one from a design perspective and from a chemistry -- a biology perspective, Chris mentioned as well. And then the third is also the development strategy, right? I think for a company like us having rapid learnings, rapid go, no go, rapid and clear with endpoints that have some precedence is also going to be important for us.

克里斯，如果我可以補充一點，不到一年前，只看整體計劃，回到我們的未來投資組合策略，你會看到提到的其中一件事就是在分子設計方面非常周到，我們今天可以在內部做到這一點，甚至可以在與 Exscientia 整合後加倍努力。CCM、NF2 和 [FPDs] 都是許可/重新利用的程序。所以從設計角度和化學——生物學角度來看，這是最重要的，克里斯也提到了這一點。那麼第三個也是發展策略，對吧？我認為對於我們這樣的公司來說，快速學習、快速行動、快速止步、快速明確具有優先權的終點也很重要。

So those are some of the other aspects that we're incorporating.

這些就是我們正在融入的其他一些方面。

That's the kind of scaled learning that we're going to get with a scaled portfolio. And in some way related, Gil from Needham is asking if we contextualize the use of AI in clinical development, like for study design and maybe how that's relevant for programs like RBM39. Now this has been a big area of focus over the last year.

這就是我們透過規模化投資組合所能獲得的規模學習。在某種程度上，來自 Needham 的 Gil 詢問我們是否將 AI 的使用具體化在臨床開發中，例如研究設計，以及這與 RBM39 等項目有何關聯。過去一年來，這已成為人們關注的重點領域。

Yes, absolutely. So yes, let's take whether it's CDK7, RBM or even MALT1. So I'll just take CDK7 as an example. There are other programs in the competitor pipeline, which indications do you go after? Solid tumors is very broad with RBM, same thing with others.

是的，絕對是。是的，讓我們看看它是 CDK7、RBM 還是 MALT1。因此我僅以 CDK7 為例。競爭對手的研發線還有其他項目，您會選擇哪些適應症？實體腫瘤在 RBM 中的分佈非常廣泛，其他腫瘤也一樣。

How do you enrich how do you enrich the biomarker for the patient population? How do you ensure that the patient populations you're going into have certain over-expression or under-expression that is predictive of greater signal to noise? I can speak more about that, but that's one of the things that we're doing, leveraging clinical genomic data like Tempus, but then also a lot of predictive algorithms that we are developing from our cell line work that we're doing internally as well as preclinally. So much more to say here. The last thing I want to say, sometimes enrollment and recruitment gets forgotten.

您如何豐富患者群體的生物標記？您如何確保您所針對的患者群體具有一定的過度表達或表達不足，從而預測更高的信噪比？我可以就此多說一些，但這是我們正在做的事情之一，利用 Tempus 等臨床基因組數據，同時也利用我們在內部和臨床前進行的細胞系工作中開發的許多預測演算法。這裡還有很多話要說。我最後想說的是，有時候招生和招募會被遺忘。

And also, sometimes real-world data to contextualize in the open-label study also gets forgotten. Both of those are really important, not just for regulatory purposes, but also for internal go, no-go decision-making, how much conviction do we have in the signal [indiscernible]. So, a lot of those approaches is really scaling up in the last year.

而且，有時開放標籤研究中用於背景化的真實世界資料也會被遺忘。這兩者都非常重要，不僅是為了監管目的，也是為了內部的決策，我們對訊號有多大的信心[音頻不清晰]。因此，許多方法在去年確實得到了擴大。

Perfect. Next up, we have Melissa who asks what criteria were used to prioritize certain programs over others? And how does this focus advance or align with Recursion's long-term strategic objectives, please?

完美的。接下來，梅麗莎詢問我們使用什麼標準來確定某些項目優先於其他項目？請問，這項重點如何推進或與 Recursion 的長期策略目標保持一致？

Yes. I mean in terms of the criteria that we use is very much what is best-in-class in the industry. So first and foremost, it always starts with what is the potential value of the drug, patient population, unmet need, scientific data starting from preclinical, clinical data, competitive differentiation, et cetera. Also in terms of the development plan, is there a feasible development plan? And then you look at risk, which is the other side of the coin.

是的。我的意思是，就我們採用的標準而言，它是業內最好的。因此，首先，它總是從藥物的潛在價值、患者群體、未滿足的需求、從臨床前開始的科學數據、臨床數據、競爭差異化等開始。另外在發展規劃方面，有沒有可行的發展方案？然後你再看看風險，這是硬幣的另一面。

So for each program, look, if I can coin it in one sentence, taking all of those components, and we have done our computational approach bottom up, so we're not being objective or we're being objective in terms of the decision-making, it comes down to, do you believe this can be a differentiated medicine? And is it serving an unmet need that will exist by the time you are going to be in the market. That's the most important. That bar has to be very, very high for us.

因此，對於每個程序，看，如果我可以用一句話來概括，那就是把所有這些組成部分都考慮進去，並且我們已經自下而上地完成了計算方法，所以我們不是客觀的，或者我們在決策方面是客觀的，歸根結底，你是否相信這可以成為一種差異化的藥物？它是否能滿足您進入市場時尚未滿足的需求？這才是最重要的。對我們來說，這個標準必須非常非常高。

Perfect. And the final question, Brendan from Cowen asks, do you expect any meaningful impact to your internal partnership strategy in light of the FDA's updated animal testing guidance? And I'll take that one to end us. Look, Recursion was built for an evolving regulatory framework like the one we're seeing from the FDA, and we'll continue to monitor for additional updates as the FDA explores all the ways that AI and other tools can be used. But from our discovery platform to our predictive AME platform to even our in vivo platform, we are generating large-scale data sets.

完美的。最後一個問題是，Cowen 的 Brendan 問，鑑於 FDA 更新的動物測試指南，您是否預計會對您的內部合作策略產生任何有意義的影響？我會用這個來結束我們。你看，Recursion 是為不斷發展的監管框架而構建的，就像我們從 FDA 看到的那樣，隨著 FDA 探索 AI 和其他工具的所有使用方式，我們將繼續監控更多更新。但是從我們的發現平台到我們的預測 AME 平台甚至我們的體內平台，我們正在產生大規模資料集。

We're building foundation models that are allowing us to move from a test at scale in the lab sort of regime for preclinical studies to a predict and validate regime for preclinical studies. And so I think Recursion is not just positioned to take advantage of these regulatory updates, but actually positioned to lead in this space going forward. So with that, we appreciate everybody's deep attention. Thanks to everyone for joining, and we look forward to seeing you all out on the Street. Thank you so much, everybody.

我們正在建立基礎模型，使我們能夠從實驗室大規模測試的臨床前研究方案轉變為預測和驗證臨床前研究方案。因此，我認為 Recursion 不僅能夠利用這些監管更新，而且實際上能夠引領這一領域的發展。因此，我們感謝大家的深切關注。感謝大家的加入，我們期待在街上見到你們。非常感謝大家。

Bye-bye.

再見。