第176集：大卫·罗布林谈如何打造一家科技生物公司

欢迎收听《长跑》。这是一档为生物科技探险者准备的播客节目。我是主持人卢克·蒂姆曼。今天的嘉宾是大卫·罗布林，他是总部位于伦敦的Relation Therapeutics公司的首席执行官。

Welcome to The Long Run. This is a podcast for biotech adventurers. I'm your host, Luke Timmerman. Today's guest is David Roblin. David is the CEO of London based Relation Therapeutics.

该公司运用多组学工具在人体组织样本中寻找药物靶点。借助机器学习算法，公司致力于发现异常生物状态与疾病之间的关联。Relation还特别强调另一种意义上的'关系'——即生物学家、工程师、医师等不同领域专家之间需要融合各自专长，以整体提升生物医药研发成功率的文化纽带。截至目前，Relation已从一群对科技和生物技术怀有广泛兴趣的投资者处筹集了超过8000万美元资金。去年底，公司宣布与葛兰素史克达成临床前合作，这家制药巨头将支持Relation开发针对纤维化和骨关节炎的创新药物。

The company uses multiomic tools to look for drug targets in human tissue samples. The company seeks to find the relationships between perturbed biological states and disease with the help of machine learning algorithms. The company also emphasizes relationships in another sense, the cultural sort between biologists, engineers, physicians, and other professionals who need to find ways to meld their disciplines together to improve biopharma r and d success on the whole. Relation has raised more than $80,000,000 to date from a group of investors with a myriad of tech and biotech interests. Late last year, it announced a preclinical partnership with GSK in which the pharma giant is supporting Relation's work to discover novel drugs for fibrosis and osteoarthritis.

大卫在开启这项事业前，曾长期担任医师，随后在拜耳和辉瑞等大型生物制药企业担任研发高管，并在与Relation公司相距不远的弗朗西斯·克里克研究所从事转化研究。我一月下旬在伦敦见到了大卫和他的团队，当时我正以另一个身份——为达蒙·鲁尼恩癌症研究基金会筹款的'提穆兰穿越'活动参与者——出席活动。借此机会，我了解到伦敦生物科技圈如何通过学术界、医院、初创企业、大型科技公司和制药企业的多方协作，共同推动生物医药研发体系的完善。希望您也能有所收获。

David comes to this work after a long career as a physician, then as an R and D executive at large biopharma companies Bayer and Pfizer, and in translational research at the Francis Crick Institute, a short distance down the road from Relation. I met David and his team in London in late January. I was wearing another hat there as part of a fundraising event for the Timuran Traverse for Damon Runyon Cancer Research Foundation. And while I was there, I learned about the London biotech community and how the many players, academia, hospitals, startups, large tech, and pharma companies, are seeking to work together and put the pieces together to improve the biopharma r and d enterprise. I hope you will learn something too.

现在有请大卫·罗布林加入我们的《长跑》节目。大卫·罗布林，欢迎来到《长跑》。

Now please join me and David Roblin on the long run. David Roblin, welcome to the long run.

谢谢卢克。能参与节目我深感荣幸，我一直很享受收听这档播客。

Thanks, Luke. It's absolute pleasure to be on. I enjoy listening, to the podcast.

非常感谢。今天我来到位于伦敦知识区（国王十字附近）的Relation Therapeutics总部。其实在见到你之前，我有幸先接触了你的团队——这种从基层了解企业文化和听取更多成员意见的体验相当不错。或许我该多采用这种方式，以获得不同视角的...

Well, thank you. I'm here today at Relation Therapeutics headquarters in London in the Knowledge Quarter, near Kings Cross. And I had the the pleasure to meet your team before I actually met you, which was a pretty good experience actually to kinda take the the pulse of your culture and and hear what more people had to say about what you're doing. I might need to do this more often actually, to get a different perspective on

自下而上的视角，而非自上而下。

The bottom the bottom up perspective rather than the top down.

是啊，是啊。

Yeah. Yeah.

也许它能告诉我一些关于我公司的事情。

So maybe it'll tell me some things about my company.

我对此表示怀疑。不过，很棒。所以，我来到这里。在我们深入之前，大卫，你能先简单介绍一下你自己吗？这个知识区是什么？这里有什么？在这里建立一家生物科技初创公司是怎样的体验？

I doubt it. Well, great. So it's, I'm to be here. Can you just before we dive in a little bit about you, David, what is this knowledge quarter? What what's here, and what's it like building a tech bio startup company here?

嗯，这里聚集了许多真正专注于知识密集型领域的公司。要知道，这里有全球顶尖的三所大学——帝国理工、国王学院和伦敦大学学院。实际上，如果类比湾区的话，车程几小时内还有剑桥和牛津，相当于门洛帕克的位置。

Well, it's it's condensation of companies that are really focused on intellectually rich areas. And I think, you know, there are three of the world's best universities here, right? Imperial, King's, and UCL. And actually within hours drive, if I was talking about the Bay Area, that would be Menlo Park, I guess. You've got Cambridge and Oxford.

因此伦敦能够成为生物科学乃至整个科学界的核心枢纽，汇聚了极其重要的智力资本，艺术领域也不例外——这里多所学院的艺术专业都非常出色。知识区成立于十多年前，当时大英图书馆已入驻国王十字区，图灵研究所也已落成。

So London can offer a kind of epicenter or a condensation, a collection of really significant intellectual capital for biosciences, but science generally, and actually the arts. A number of those colleges are outstanding at the arts. And the Knowledge Quarter was formed just over a decade ago now. The British Library was already present in King's Cross. The Turing Institute had opened.

还有我特别关注的克里克研究所，这座法国克里克研究所拥有超过百万平方英尺的综合性湿实验室。正如你所知且我们稍后会谈到，我曾是该研究所开发的关键人物，并担任首任首席运营官兼科学转化主任。克里克研究所就像这里的华特迪士尼项目——建好了自然有人来。虽然人才本就聚集于此，但确实需要一个核心来凝聚力量。

And something very close to my heart, the Crick Institute, the France Crick Institute was built over a million square foot of integrated wet laboratories. And as you know and we'll speak about, I was key to the development of the and was the first COO and director of scientific translation. And the Crick, I mean, a sort of Walt Disney thing here, right? Build and they will come. I mean, a sense, the people were here, but it needed a focus of concentration.

随后谷歌旗下公司如Isomorphics等纷纷入驻，像Relation这样的企业也认定知识区是理想的落户地点。这片区域大致以国王十字为中心，覆盖约一平方英里范围。对我们生物医学研究团队而言，另一个优势是通过本地医院能接触到2000万患者数据。

Windforward, Google come. They spin off some of their companies, Isomorphics and others. And you've got this collection of other companies like Relation that decide the knowledge quarter is a really great place to be. And roughly, it's around a mile square around King's Cross. And the other thing from our perspective, a biomedical research group, we've got access to 20,000,000 patients through the hospitals here.

再次强调，这里有许多杰出的医院，或许我们可以讨论如何利用这些医院获取对我们工作至关重要的样本类型。

There's, again, a large number of outstanding hospitals and then maybe we will talk about the way in which we utilize those hospitals to bring in the kind of samples which are important in our work.

我们一定会谈到这点。正如你提到的，这里确实存在生物科学、医学和技术的独特融合，三者之间距离相当接近。

We'll definitely get to that. There's a really interesting confluence of assets here, as you mentioned, both bioscience and medical and, technology, all in a pretty close, proximity to each other.

是的。

Yeah.

好的。让我们稍微回溯一下，聊聊你的经历，你是如何来到这里的？从头开始说吧。大卫，你是哪里人？

Okay. So let's back up a little bit and talk a little about you, how you came to this place and start from the beginning. Where where are you from, David?

我是威尔士人。来自英国的一个公国——威尔士。我在那里出生，算是X世代吧。虽然不想完全透露年龄，但我更接近婴儿潮一代而非千禧一代。

I'm a Welshman. So I came from a a principality of The UK called Wales. I was born there. I'm generation X, I suppose. Without giving my age completely away, I'm closer to the boomers than the millennials.

你知道，威尔士人在英国是个相当自豪的民族。我们有一句我喜欢的诗：'生为威尔士人就是生来享有特权，不是嘴里含着银勺，而是心中有诗，血中有歌。'我可能记错了原句，但大意如此。我们有着丰富的文化遗产，音乐、体育和教育都非常重要。我们那里主要是采矿和炼钢的社区。

And, you know, the Welsh are quite a proud nation within The UK. We have a saying from a poem that I like, which is to be born Welsh is to be born privileged, not with a silver spoon in our mouths, but with poetry in our hearts and songs in our blood. I probably butchered the quote, but you get the idea. So a rich cultural heritage of music, sport, and education was really important, right? Were, I guess, mining communities, steelworking communities.

你家人是做什么的？

What did your family do?

我的家人都是教师。我父亲是有机化学博士，在高中及更高年级任教。母亲则是英语老师，她培养了我对阅读和文学的热爱。

My family were teachers. So my father was a PhD in organic chemistry and taught in high school and beyond. And my mother was an English teacher. So, you know, she taught me about a passion of reading and literature and the like.

好的。所以你们家是文理结合的。

Okay. So you're bridging the arts and the sciences in this household.

家里还充满音乐氛围。我祖父在威尔士合唱界是位举足轻重的人物。你可能听说过威尔士男声合唱团，他在五六十年代鼎盛时期是核心人物。音乐和体育在我的成长中同样重要。

And a lot of music in the household. My my grandfather was a massive an important figure in in the coral community in Wales. You may have heard of the the Welsh male voices. And he was a very significant figure among When it was big, right, in the fifties and sixties and seventies. And certainly music played an important part as well as sport in my upbringing.

这教会了我团队协作的真谛——聚集优秀人才共同成就伟业。你从事什么运动？作为威尔士人，我打橄榄球，直到医学院时发现自己体型不适合专业比赛。

And I think that taught me a lot about teamwork, really, and, and assembling great people around you to to to deliver great things. So what sports did you play? I played I'm a Welshman, so I played rugby. Rugby until I realized I was probably a little bit too small to play it seriously. And that happened when I was in medical school.

当时对阵英格兰大块头球员。但我主攻板球（虽然在美国不流行），水平不错，曾代表年龄组参赛。我属于那种很难被淘汰的打法。

I was playing against some very big English players. But my main sport was cricket. Not an American sport, don't think, but it's And a global I was pretty good. I played age group representative cricket. I would describe myself as someone who was difficult to get out.

得分不算高但极难出局，所以让我开场很实用——通常比赛结束时我还在场，虽然得分不多，但能为另一端的击球手创造进攻机会确保团队胜利。

Didn't score as many runs as I should, but I was very difficult to get out. So it was useful to have me at the top of the innings because I was almost there generally at the end of the innings with not enough runs, but I was still there. So it allowed this person at the other end of the wicket to play shots and and and make sure that the team won.

那么你从中学到了哪些团队协作的经验？有哪些至今仍受用？

So you learned some lessons about teamwork. What was one that you think stayed with you from those years?

我是说，要与比你优秀得多的人为伍。你说你见过我的团队，我对此很放松，嗯，我姑且这么说，虽然显然我希望这不是真的，但我姑且说我喜欢成为房间里最笨的人。我很适应进行对话，因为这样我能感觉到，如果有各部门和具备专业能力的人为我工作，他们当然应该比我更了解他们的领域。我观察到领导者常常为此挣扎，他们觉得自己必须事事精通。所以我很坦然接受这一点。

I mean, surrounding yourself with people who are a lot better than you. So you said you met my team here, and I am very relaxed about having well, I I say loosely, and it's obviously I I hope not true, but I say loosely I like to be the the most stupid person in the room. And I'm comfortable with with having conversations Because then I can feel that if I've got departments and people working for me with functional expertise, they of course should know a lot more than I do about what they do. My observation has been that leaders often struggle with that they feel they need to be on the top of everything. So I'm comfortable understanding that.

我在招聘时也会寻找这样的人，一个让我感到自在，但角色与我截然不同、技能互补的人。第二点是关于使命，对吧？本质上就是组织的目标和意义。你们追求什么？长远目标是什么？

And I kind of look for that as recruiting someone, someone I can be comfortable with, but definitely plays a different role to me and brings different skills. And the second thing is about mission, right? Literally, is the what and the why of your organization? What are you seeking to do? What is the long run?

你们最终要达到什么长远目标？为什么要这么做？因为我认为动机很重要。它代表着意图。比如对我们公司而言，就是患者在等待。

Where are you gonna end up at the long run? And why you're doing it? Because I think why is important. That represents intent. So, you know, for the relation, it is the patient is waiting.

这就是我们做这件事的原因，因为有患者正等待着我们变革性药物的创新成果问世。

That's why we do this, because there are patients who are waiting for innovations in our transformational medicines to come.

嗯。那么你小时候是什么样的学生？

Yeah. Yeah. So what kind of student were you growing up?

我在学业上如饥似渴。学习非常刻苦，会计算学习时长，非常科学化。主攻化学、数学和物理，其实不包括生物——有意思的是，我觉得生物有点松散随意，如果可以用这个词的话。

I I was hungry academically. I worked really hard. I counted the hours I was working and very scientific. So chemistry, maths and physics were my main, not biology actually, interestingly enough. I found biology a bit loose and goosey, if that's a term I can use.

它

It

生物学曾缺乏规则，至今仍有所欠缺。我们可能会回头讨论这一点，因为这是我们处理的重要议题之一，或许能从中看到生物学数据。但我觉得很明显的是，当年生物老师的教学方式偏向松散的观察性方法，而非归因于因果关系和方程式。我特别喜欢数学，因为可以写下方程式；也喜欢化学，因为能清晰看到方程式的起点、终点及其中间过程——也就是所谓的对世界的‘白盒视角’。

had a lack of rules and still has a lack of rules. We may come back to this, think, because this is kind of one of the important things that we deal with and may see data of biology. But I think it was obvious to me that when I was, that the biology teachers taught in a kind of loose, excuse me, observational way rather than a reduction to causation and an equation. And I really liked the maths where you could write the equation down or chemistry where you could definitely see the beginning and the end of an equation and everything that was in between. You know, so called kind of white box views on the world.

我能看到输入、输出以及中间过程，所以被这些学科吸引。实际上我喜欢挑战——这是另一个原因。我不介意尝试新事物，因此涉足音乐、体育，我的职业生涯也横跨了几个不同的小领域。

I can see input, output, and I can see what's in between. So I was kind of drawn to those. And actually I like challenge, think was the other thing. I didn't mind starting something new. So hence music, hence sport, and hence, I think my career has sort of straddled several different mini careers.

所以我喜欢变化。我想可以称自己为‘不安分的人’。

And so I like change. I think I'd call myself restless.

那么，在没有太多规则的情况下，你是怎么最终选定生物化学或进入这个领域的呢？

So how did you end up settling on biochemistry or or coming to this area without a lot of rules?

是的。我先后攻读了生物化学和医学学位。说实话，这和我现在给别人的职业建议一样——当有人问我如何成为首席执行官时，我说：‘其实只需考虑未来两年。’

Yeah. So I I did a degree in biochemistry and then a degree in medicine. Mhmm. And honestly, I think it's the advice I give someone now when someone asked me about the career and they said they wanna be a chief exec. I said, well, actually, just think about the next two years.

思考哪些事既有趣又能提升你作为人和职业人士的价值。我从未规划超过两三年后的事——这可能不是个令人满意的答案，但...

What are the things that you could be doing that are interesting and adding to you as a human being and your career. I don't think I've ever looked much further than two or three years. That may not be a very satisfactory answer, but You

你并没有宏伟计划。

had no grand plan.

没有什么宏伟计划。某种程度上说，我当时就像是在不安分地游荡，寻找挑战。所以，嗯，当时的情况是

No grand plan. It was like, in a sense, I was wandering in a restless fashion, looking for challenge. So Well, what was

是什么让你选择了生物化学或者医学？

it about biochemistry that caught you or or then medicine?

坦白说，就是因为难。学校的导师并没有建议我学医。实际上，我当时考虑的是药理学，还研究了一大堆相关方向。后来有个人——具体是谁我记不清了——对我说：'你聪明得足够学医，为什么不试试呢？'

Well, because it was difficult, frankly. My tutors in school didn't suggest medicine. And, you know, in fact, I was looking to go into pharmacology and there were a whole bunch of things I was looking at. And then someone, some, I actually don't remember who it was, said, you are smart enough to do medicine. Why don't you do medicine?

可能也和我成长环境有关，我所在的社区在我上学时对学生的志向推动不够。所以我申请了医学院，被录取后去了伦敦大学的圣乔治医学院。那是全新的医学院，我喜欢这种新鲜感——作为创校本科生。嗯。

And it may be something to do with the community that I came from didn't push enough when I was in school for the pupils to reach out. So I applied for medicine, got into medicine, went to St. George's in the University of London. It was brand new med medical school, so I liked the feel of something new that I was a founding undergraduate. Mhmm.

我真的很享受学医的过程。我想这种享受源于医学涵盖的多元学科——你能接触到广阔领域，既有基础科学，当然还有临床实践。

And really enjoyed it. And enjoyed it, I think, because there's so many different disciplines within medicine. You get to cover a lot of area, basic science and, obviously, in the clinical practice.

好的。所以你在攻克难题... 对... 学习新知识，然后成为医生，成为临床医师。对。接诊病人。

Okay. So you're tackling hard things and Yep. Learning new things and became a doctor, became became a clinician. Yep. Saw patients.

对。你在那里执业的是哪一科的医学？

Yep. What what type of medicine did you practice there for a while?

在英国，医生需要经历一段轮转不同专科的时期以获得广泛培训。我认为英国培训体系的重要之处在于前五六年培养全科医生，这一模式至今基本未变。我女儿现在是名医生，正处于中级阶段，正体验着各种不同专科。我认为这种经历能拓宽视野，最终对成为专科医生大有裨益。

In the in The UK, you have a period of moving around several specialties to get a broad training. I think UK training is significant in the fact it gets physicians for the first five or six years. And it's still largely the same. My daughter is a doctor of medicine now and she's at mid grade, gets to experience a lot of different speciality. And I think that gives you a breadth of understanding, which helps you as a specialist ultimately.

所以我完成了整套课程：呼吸内科、传染病科、心内科、肾内科，还轮转过重症监护室——如果继续从医的话，我可能会成为重症专家。另外也学过麻醉科，涉猎很广，但我真正享受的是救治危重病人。因为我性子急，喜欢快速见效的工作。在传染病科和重症监护室，你能直观看到药理作用，对吧？

So I did a whole course like respiratory medicine. I did infectious disease. I did cardiology, did nephrology, did a period of intensive care, which if I'd stayed in medicine, I think I'd have been an intensivist and also did some anesthetics, so quite broad, but really enjoyed the critically ill patient. So I enjoyed actually making a difference rather quickly because I'm restless. And in infectious disease, in intensive care, you know, you can see often the action of pharmacology, right?

比如用药物治疗感染诺如病毒的儿童，孩子又吐又泻奄奄一息，两三岁的年纪。你给他挂上静脉点滴，转身写病历的功夫，回来就发现孩子已经能玩耍了。

You give, I mean, pharmacology, like giving a child who's come in with norovirus, vomiting, diarrhea, who appears moribund, you know, a two or three year old child, you put them on an intravenous drip and you literally go and write the notes up and come back and the kids play.

你和病人一对一治疗，采取行动后，他们要么好转，要么没有。

You treat the patient one on one, you do something, and they either get better or they don't.

通常情况确实如此。这让我不断思考自己能做什么——再加上本科时我在科学课程表现突出，这种对科学的浓厚兴趣，某种程度上与临床实践中的人文关怀形成了微妙的平衡，不知道你能否理解。

They don't. And and in those scenarios, that's generally the case. And because that that also, as well as my undergraduate studies, where I kind of excelled at the scientific elements, made me continue to think what what I could do. And it also, my deep interest in science kind of was almost balanced against the humanity of being a clinician in practice, if you understand what I'm saying. Yep.

那五年里我反复思考职业方向，最终在两种考量间摇摆：对科研的热爱与临床实践的抉择。

And I thought a lot about that in my five years, you know, where I should best go. But, and those are the sort of two measures that were playing off, this enjoyment of science and research versus versus clinical practice.

那你为什么离开医学领域？

So why did you leave medicine?

我离开了医学领域。是的，这种不安分的感觉又回来了。当时我已经完成了五年的学习，通过了英国的顾问医师考试（那时他们称之为专科考试），成为了皇家内科医师学会的会员。我本已准备好进入专科培训，最终成为一名顾问医师。这意味着选择，对吧？

I left medicine. Again, yeah, this the restless theme comes back. I I was at the stage having done five years where I'd got my, what were my consultant exams in The UK then, specialist exams, they called, I became a member of the Royal College of Physicians. So I was all ready to progress into specialist training and then ultimately become a consultant. So it meant choice, right?

当时我面临的选择，当然，这个选择将影响未来二三十年的人生。我要么成为心脏病专家，要么成为重症监护或呼吸科专家。这其实让我感到害怕，因为我在想，以我不安分的性格和喜欢不同挑战的个性，真的能坚持二三十年吗？或许我可以成为一名学术型医师，但在八十年代我成长的那个时期，这条道路并不那么明确，对吧？当时这类人相对较少。

And I was looking at then a choice, of course, that was gonna last twenty or thirty years. I was gonna become a cardiologist or I was becoming a intenserist or a respirologist. And that scared me actually, because I thought, could I really do that for twenty or thirty years because I'm restless, because I like different challenges? Maybe I could have become an academic physician, and that, but that wasn't such a strongly tried path back in the eighties when I was coming through, right? There were relatively few of them.

也许是因为我面前没有导师。这触及到一个重要问题——拥有导师的重要性。也许我当时就读的医学院就在附近。现在那所医学院里的导师们，当时也在那里。所以我某种程度上是在寻找一个能让我不安分、能接受挑战的地方。

And maybe I didn't have the mentors in front of me. This touches on the important thing of having mentors present. And maybe I went to a medical school that was near. So those mentors, which are there now in that medical school, were there then. So I I was kind of looking for some place to to be restless, I guess, to be challenged.

当时英国确实还没有真正的生物科技产业。

Well, there wasn't really a biotech industry in Britain at the No.

但制药业很发达，对吧？是的。当时的制药业比现在更强大，公司数量更多。比如辉瑞就在英国，其三分之一的研发工作都在一个叫桑威奇的地方进行。

But there was a big pharma industry. Right? Yep. Which was much stronger than it is now in the sense that there were more companies. So, you know, Pfizer was here and had, like, a third of its research efforts in The UK down in a place called Sandwich.

但你当时意识到这一点了吗？

But were you aware of it at that time?

并没有特别注意到。

Not particularly.

这有点像学术界和工业界之间的那堵墙。

It was sort of like the wall between academia and industry.

非常明显。我这个年龄段的人。对吧？所以当我向我的资深临床导师们提到，我正在考虑进入制药行业时，他们的评价是你太聪明了，不该在干这行时考虑这个。因为当时普遍认为，选择这条路的人都是因为做不了其他事。

Massively. My my age group. Right? So when I spoke to my senior clinician mentors and said, I'm thinking of going into the pharma industry, the comment was you're just far too bright to do that while you're doing it. Because it was generally seen as something someone went to and they couldn't do anything else.

嗯哼。当然，这种看法并不正确。

Uh-huh. Which wasn't true, of course.

他们赚很多钱，但并不怎么聪明。

They make a lot of money, but they're not very smart.

就是那种情况。对。他们有公司配车之类的福利。但其中有一位叫乔治·格里芬教授的传染病专家，他实际上对我说——他是我敬重的前辈之一，因为你知道，人总会在某些人身上看到自己的影子——他说，真的，如果我是你，我会认真考虑这个机会。

That sort of thing. Yeah. They get the company car and and all of that. And but one one of the guys, a guy called professor George Griffin in infectious disease, he he said to me actually, and I he was one of the the folks I looked up to because I've, you know, like, you see you see yourself in in people. And he said, actually, yeah, if I was you, I'd be thinking about this.

我认为那里正在发生真正激动人心的事情。要知道，他当时处于免疫学应用于传染病研究的最前沿。

I think there are really exciting things happening. And because, know, he was in the cutting edge of immunology at the time as it as it's deployed into infectious disease.

所以这是八十年代的事？

So this was the eighties?

是的，那是八十年代末。

Yeah. This was late eighties.

八十年代末。好的。那么你是怎么最终决定跳槽到工业界的？

Late eighties. Okay. So how did you end up making that leap to industry?

嗯，说到底，当时我正处于一个转折点，要么申请专科培训，要么做点别的。嗯。然后我开始申请各种工作，收到了Upjohn（那家老牌美国公司）的offer。后来又被叫到辉瑞在Sandwich的基地面试，说实话被辉瑞在这个东南部宁静小镇的布局震撼到了。

I well, ultimately, it was at I was at that juncture, right, where I had to apply for specialist training or do something else. Mhmm. And I started to, you know, apply for various jobs, got offered a job with Upjohn, that old American company. And then I was called down to Sandwich on Pfizer to have an interview and was blown away actually with what Pfizer put down in the sleepy town in the Southeast.

是啊，你喜欢那里的什么？

Yeah. What did you like about that place?

那里设施非常完备。你知道，那是个研发团队。我能看到商业部门，开发部门，还有一些非常令人印象深刻的实验室。那里有些特别厉害的人物，特别是在化学领域。

It was full thickness. Like, you know, it was an r and d group. I could see commercial units there. I could see development units there, and I could see some very impressive laboratory. And they had some extraordinary characters there, you know, particularly in chemistry that come out of chemistry.

梅菲辛三明治在化学领域尤其强大，有些人的思维方式非常老道。我当时就觉得这会是个很酷的工作地点。他们跟我谈过做传染病研究的事，想让我用氧氟沙星开展鼻腔昏迷肺炎和腹腔脓毒症的临床研究，都是重症患者。这正好契合我对重症监护和传染病的兴趣，所以算是个不错的契合点。

Mephisin sandwiches are very strong in in chemistry in particular, and there were some people who were very old at thinking. And I just thought this would be quite a cool place to come. And they talked to me about doing work in infectious disease, and they wanted me to run clinical studies with a drug culture of ofloxacin in nose coma pneumonia, did abdominal sepsis, so very sick patients. So that really appealed to my interest in intensive care and infectious disease. So it was kind of a nice nexus.

利用了你在患者护理方面的经验。嗯，所以你最初是从研发做起的。

Leveraged your experience with patients. Yeah. So you're starting out in development.

事实上，当时在辉瑞公司，他们鼓励保持临床实践。所以大概有两三年时间，直到辉瑞的工作变得难以兼顾，我一直在当地医院坐诊。

Well, fact, in in Pfizer at the time, they encouraged maintenance of clinical practice. So for, I think, two or three years until it became too difficult with the Pfizer job, I I held clinic in in the local hospital.

哦，真的吗？所以你还在接诊病人？是的，同时还在工作。

Oh, really? So you're still seeing patients Yeah. While working

辉瑞？没错。这可能让你再次——稍微跳脱一下话题——我希望这能让你部分感受到科技生物领域的氛围。它非常以患者为中心。嗯。

Pfizer? Yeah. And that may give you again, jumping forward a little bit, may give you, I hope, part of the feel you've got in this tech bio. It's very patient orientated. Mhmm.

这种理念一直伴随着我。

And that that remains with me.

所以你当时在那里。那是一个充满激励的环境，有很多做不同事情的有趣人士。你不会感到无聊。确实不会。

So you were there. You you it was a stimulating environment. Lots of interesting people doing different things. You weren't gonna get bored. No.

那你当时是否想过，这个地方能让我产生比传统一对一诊疗更广泛的影响？是的。

And were were you thinking like this is a place where I could have broader impact than sort of the classical one on one Yeah.

完全正确。作为一名医生，完全正确。我不想听起来太过自负，但确实有一种感觉，我显然在一对一诊疗中产生了影响。事实上，我当时选择的医学专科放大了这种一对一的效果。

Totally. Physician. Totally. They they you know, I I don't want to sound too ego egotistical about it, but there was a definitely this feeling that I was making a difference one on one, clearly. And in fact, I was choosing I was choosing specialisms in medicine, which which amplified the one on one effect.

对吧？就像我跟你举的那个肾脏诺如病毒的例子。嗯。我之前确实这么想过。比如，你知道，如果我研发出一款药物，那将产生巨大的影响。

Right? Like the example I gave you about the kidney norovirus. Mhmm. And I had definitely thought that. Like, you know, if I if I develop a medicine, that makes a massive difference.

它会被全世界的人们使用。

It's gonna be used by people around the world.

好的。所以你在辉瑞的桑威奇待了相当长一段时间，对吧？

Okay. So you were there at Sandwich for quite a while for Pfizer. Right?

嗯，既是也不是。最初我在那里待了三年，从事传染病研究，研究一种叫阿奇霉素的药物，这种药现在还在使用。

Well, yes and no in the sense that I spent three years there initially working in infectious disease on on a drug called azithromycin, which is still available now.

著名的抗生素。

Famous antibiotic.

是的。我和一个叫迈克尔·邓恩的人合作做了些有趣的工作，参与了一项名为‘巫师试验’的研究，旨在探究非典型微生物是否在冠状动脉疾病中起作用。当时PCR技术刚开始能识别冠状动脉斑块中微生物的微小片段，这些片段已知会引发炎症。当时的想法是，也许阿奇霉素会有效。这其实是那种如果有数据就能通过计算机模拟完成的研究之一，对吧？

Yep. I did some interesting work with a guy called Michael Dunn on what was called the Wizard Trial, which was looking to see whether atypical microorganisms had a role in coronary artery disease. And so also PCR had begun to identify small parts of microorganisms in coronary artery plaque, which are known to be inflammatory. The idea was, well, maybe azithromycin could be effective. It's actually one of the studies you could have done in silico if you had the data, right?

因为很多五六十岁服用非典型微生物药物的男性，你知道，很可能死于心血管疾病。所以本可以通过计算机模拟来解答这个问题，但辉瑞进行的大型研究结果却是阴性的。

Because a lot of people who were getting drugs for atypical organisms who were in their fifties and sixties men were, you know, likely to be dying of cardiovascular disease. So you should have been able to ask the question that in silico way, but Pfizer ran a big study, which was negative.

但当时的工具是PCR技术，这引发了许多关于你们药物及其可能包含成分的有趣问题。

But the tool at the time was PCR, and that threw off a whole lot of interesting questions about your medicine and what it could include.

对吧？就像悉尼·布伦纳，我特别喜欢他的一句话：新技术带来新发现和新理念。首次测量事物的能力会创造出一系列机遇，这就是其中一个例子。

Right? You know, like Sydney Brenner, a quote I love from Sydney Brenner other than is new technologies provide new discoveries and ideas. And it's kind of the ability to measure things for the first time creates a whole set of opportunities. And that's an example of one of them.

我非常喜欢这句话。它基本上是说技术推动科学发展。是的，而不仅仅是提出假设。没错。

I'm a big fan of that quote. It it says that basically technologies drive science. Yeah. That that, rather than just coming up with hypotheses. Yep.

这是部分原因，但技术带来的远不止这些。好的，那么你研究过一些有趣的项目。你曾在辉瑞工作，之后有什么新动向？

That's part of it, but technology throws off many, many more. Okay. So, you worked on some interesting things. You were there, at Pfizer. What was your next move after that?

其实我是辉瑞的回聘人员，所以你会觉得我在辉瑞工作了很长时间。

So then I then I I'm a Pfizer returnee, which is why you think I was a Pfizer for a long time.

好的。

Okay.

后来我被猎头挖去拜耳领导传染病部门。当时我常驻英国瓦皮塔尔和德国勒沃库森，那时拜耳是抗感染领域的巨头，拥有环丙沙星等药物。好的。

So I I I was head hunted to head up infectious disease for Bayer. So I was based out of The UK, Wapital and Leverkusen, they're German sites. And at that time they were a massive anti infective company. They had ciprofloxacin. Okay.

我为他们研发了一种名为莫西沙星的药物，这是一种广谱喹诺酮类抗生素。要知道，这种药至今仍在某些适应症中使用。

And I developed a drug called moxifloxacin for them, which is a broad spectrum quinolone. And, you know, that is still available now and utilized in certain indications.

所以你升职了。到这个时候，你在管理层步步高升。

So you got a promotion. You're moving up in management by this point.

是的，而且乐在其中，相当商业化。我喜欢这样描述我的职业生涯：我的职业轨迹是从病床到实验室，偶尔带着几个点子从实验室回到病床。这让我形成了独特的行业视角，卢克。我参与了环丙沙星的商业化过程，当时我们做了些不错的工作——要知道环丙沙星是款非常成功的药物。

Yeah. And enjoying it Quite commercial. So I like to say when I'm sort of painting my career, that my career went from bed to bench with a few ideas coming from bench to bed. So I and that gave me a certain line of sight, Luke, which I think is relatively unique in the industry. So I was involved with the commercialization of Cipro, and we did some kind of good things then, which was, you know, Cipro a very successful drug.

我记得它当时是重磅炸弹级药物之一。具体数字记不清了，但它为拜耳带来了数十亿美元收入。我们为滴耳剂撰写了产品改进型新药申请，还获得了儿科适应症——因为我当时很担心许多儿童会在缺乏数据支持的情况下被开具环丙沙星。这主要基于数据库回顾性研究，因为很多医生已在囊性纤维化等儿童重症感染中使用该药。通过药理学实验研究，我们最终拿下了这个适应症。

I think it was one of the blockbusters then. I can't remember exactly, but it was billions of dollars of revenue for Bayer. So we wrote product enhancement NDAs for otic drops and actually got a indication for pediatrics because I was kind of concerned that a lot of kids would be prescribed Cipro without much data. And that was mainly a database review that got that NDA because a lot of physicians were using Cipro in CF and cystic fibrosis and other serious infections in kids. So we were able to, with some pharmacology experiments, studies, get the indication.

莫西沙星项目期间，我亲历了拜耳败血症项目的终止阶段。那是段充满变动的有趣时光，回想起来也是重大晋升时期，因为我当时负责管理大量人员和重要产品。我在那里又待了约三年，直到接到辉瑞的电话——他们问我要不要回去工作？

Moxifloxacin, I was around when the sepsis programs were being terminated in Bayer. So it was kind of interesting dynamic time and enjoyed it. Massive promotion at the time, looking back at it, because I was looking after a lot of people and a lot of important products. So I stayed there again about three years. And until I got a phone call from Pfizer, which was, would you come back essentially?

那是90年代初（确切说是90年代中期）非常特殊的邀约：作为药物开发者，是否愿意回归研究部门？当时的情况是，研究部门做了大量优秀工作，但突然提交的候选药物常涉及难开发的适应症，或忽略了安全性等要素。他们从我首次任职期间就认定我具备科学素养，能回去与科研人员沟通研发方向。我们那时开始制定目标产品档案：如何用一页纸简明阐述医学问题？

And this was like early nineties, a very, mid nineties, a very unique ask, which was as a drug developer, would you come back and work with research? And the idea had been that research was doing a lot of good work, but it was suddenly presenting drug candidates in indications that were difficult to develop, or perhaps other elements hadn't been considered like safety. So they'd identified me in my first tenure there as someone who was scientific and could go back and talk to scientists about what they were doing and develop. We started developing target product profiles then, which was, okay, what is the enunciation of the medical problem here? How do you write that down in the short one pager?

这如何转化为研发计划？嗯，就是这类工作。还包括转化医学研究。

And how does that represent an R and D plan? Mhmm. It was that sort of thing. And also translational research.

嗯。

Mhmm.

你会如何...你怎样...你如何从患者角度审视科学数据，反之亦然？

How would you how how do you how do you how do you look through from the patient to the the scientific data and vice versa?

嗯，这份工作可能正是因为有医生经历才得心应手，因为你最终会接触患者，并且对目标产品概况该是什么样子有些概念。是的，就像

Well, this is probably a job where having been a physician was helpful because you you ultimately had contact with the patient and had some idea of what a target product profile really ought to look Yeah. Like

那确实是

It was quite a

一种独特的实际感受。

unique a practical sense.

对。具体来说，我需要一张特殊通行证才能进入产品发现部门。除了常规的研发通行证和辉瑞园区门禁卡外，还得额外申请一张准入证。进去后在水机旁有过几次非常棒的闲聊，真的就是那种典型的茶水间对话时刻。

Yeah. So to to give you a to give you a concrete, I had to have a unique pass to get into product discovery. So I had my development pass, my Pfizer pass for the site, and there was another pass that I had to have to get in. And, know, went in and had some really great conversations around the water cooler, right? So it literally is the water cooler moments.

其中有个例子可以和你分享，卢克。比如有次是和研发主管托尼·伍德的交流，你认识他的。我们共事超过十年了。

And one of one of which, I'll give you an example, if you like Luke. Yeah. One of which was with a guy called Tony Wood, who you know as head of research. He and I worked together for over a decade.

不过他当时在辉瑞。现在他在葛兰素史克担任首席科学官。

He was at Pfizer though. He at GSK now as chief scientific officer.

他在化学领域和辉瑞的职位与我相当。我们当时在讨论几篇论文，大概是97年发表的，我记得是在《新英格兰医学杂志》上，还有一篇是在另一本期刊。这些论文描述了某些患者群体似乎对感染HIV具有抵抗力，即便感染了HIV也不会发展成艾滋病。研究发现这些患者具有特定的基因变异，称为delta 32多态性。

He was sort of at my level in chemistry and Pfizer. And we were having conversations about a couple of papers that had been published in '97, I think they were published, New England Journal, and one was in another journal. But it described observation that there were groups of patients who seemed resistant to getting HIV, and if they had HIV, were resistant to progressing to AIDS. It had been determined through study that those patients had a specific genetic change. It was called a delta thirty two polymorphism.

delta 32变异位于CCR5基因上，这个基因对免疫细胞（尤其是T细胞）很重要，是HIV病毒结合并进入细胞的两个受体之一，最终导致细胞破坏。研究观察到，一些男同性恋群体和性工作者（记得是尼日利亚或乌干达的）受到保护。后来发现这是一种相当常见的多态性——杂合子概率约20%，纯合子只有1-2%。如果是纯合子，CCR5基因就会失去功能，从而避免患上艾滋病。

The delta 32 was on the CCR5 gene, which is a gene that's important in immune cells, T cells in particular, and is one of the two receptors that bind HIV and internalize, and cause ultimately the destruction of the cell. So it had been observed that there were groups of gay men who were protected and sex workers in, I think it was Nigeria or Uganda, don't remember now. And it turned out there was quite a common polymorphism. I think the heterozygous rate is twenty percent and the homozygous rate is just one to two percent. If you're homozygous, you have a lack of function of CCR five and protected from getting AIDS.

病毒无法进入细胞。

The virus doesn't get in the cells.

确实无法进入细胞。因此从感染HIV之初就具有相对抵抗力。这些人群的研究发现被溯源到这个基因变异，随后人们理解了这与CCR5基因的关联。

Doesn't get in the cells. So and you're relatively protected from getting infected at this from the get go with HIV. So there was an observation in the study in in these populations, and that was backtracked into this genetic change. And then it was understood this was linked to the CCR five gene.

于是人们开始思考：能否开发一种疗法来模拟人类遗传中观察到的这种保护效应？

And then people say, Is there a way we can make a therapeutic that that mimics this effect that we're seeing in the human genetics?

没错。要知道在那个年代，还没人成功抑制过肽间相互作用。当时有位兼具医师和药物开发者身份的人在讨论中提出'为什么不试试这个？'。当然，托尼是位非常聪明的化学家，这确实是个挑战。但可以说，这是首个基于人类遗传学认知推动的研究项目。

Yeah. And, like, back at that time, of course, no one had inhibited a peptide peptide interaction. So kind of in conversations, almost like a physician drug developer said, why didn't we just do this? And, you know, obviously, Tony was a a very smart chemist, and this could be quite a challenge. But I'd like to say now it was the first program that was motivated by an understanding of, human genetics.

实际上进展非常迅速。马拉维若克（Mariviroc），细胞进入抑制剂，于2007年获批。所以最终这是一个非常快速的项目，至今仍是重要药物。但你知道，这故事很迷人，当你审视这些现象时也相当有趣。但为什么为什么为什么它在人群中如此高度保守的多态性？

And And actually moved very quickly. Mariviroc, Cells Entry, was approved in 2007. So it was a very rapid program in the end that remained an important drug. But, you know, fascinating So, story and actually also interesting when you look at these things. But why why why is it so such a conserved polymorphism in in population?

嗯，你正在讲述一个对遗传学产生浓厚兴趣的故事。是的。而这正与基因组学时代同步。是的。你提到了克里克研究所。

Well, you you're telling a story of becoming really interested in genetics. Yeah. And this is coinciding with the age of genomics. Yeah. You mentioned the Crick.

我们能快进一下吗？你是怎么最终去到那里的，以及你在那里具体做了什么？

Can we fast forward a little bit how you ended up going there, and and what what did you do there?

当时克里克研究所还是个建筑工地时，我接到了保罗·纳斯和已故的大卫·库克西爵士的电话——后者去年刚去世。

So I was so the the Crick was a building site when I was called by a a Paul nurse and a guy called sir David Cooksey who who actually passed away last year.

保罗·纳斯，诺贝尔奖得主。

Paul Nurse, Nobel Prize winner.

诺贝尔奖得主，前洛克菲勒大学校长，英国皇家学会会长。实际上他刚连任皇家学会会长。我发现保罗有种魅力能说服任何人跟他共事，他说服我可以当几年'建筑工'来完成克里克研究所的建设。

Nobel Prize winner, ex president of the Rockefeller, president of Royal Society. In fact, he's just been reelected as president of the Royal Society. And Paul has a way of seducing anyone to work with him, I've I've found, and convinced me that I could be a builder for a couple of years and finish putting up the crick.

若你喜欢收听本节目，你定会爱上订阅《蒂默曼报告》。在这里你能读到我对生物科技领域最有趣初创企业的深度报道、每周前沿观点专栏以及特邀撰稿人的评论文章。个人用户可选择月付、季付或年付方案，团体订阅可享折扣优惠。请访问timmermanreport.com并点击订阅获取详情。

If you like listening to the long you'll love a subscription to Timmerman Report. This is where you can read my coverage of the most interesting startups in biotech, the weekly front points column, and commentary from a rotating cast of contributing writers. Individual subscriptions are available on a monthly, quarterly, or annual basis. Group subscriptions are available at a discounted rate. Go to timmermanreport.com and click on Subscribe for more.

辉瑞关闭了桑威奇基地。这大概是你离职的时间吗？

Now Pfizer closed Sandwich. Was this around the time when you left?

不是。我是在它关闭前一年离开的。嗯。它并没有完全关闭。它关闭时就是我们对话开头提到的那个综合性研发基地。

No. So I left a year before it was it closed. Mhmm. It didn't completely close. It closed as that integrated R and D site that we described at the top of our conversation.

所以我想那里大概还有一千人左右，但功能上——不想用这个词，可能有点贬义——但就像个酒店。部分是办公室，部分是实验室。所以人数远不如从前，肯定不再是当初那个综合性研发部门了。

So I think there are still about a thousand people there, but in a, don't want to use, maybe it's a pejorative term, but it's like a hotel in function. It's part office, part labs. So there aren't that many people that are certainly not the integrated R and D unit it was.

嗯。嗯。那个嘛，那是个不同的时代，制药公司的设计理念就是要有这种大型郊区园区，更像大学校园那种

Mhmm. Mhmm. And that well, that was a different era of, like, philosophy on design that that pharma companies would have these big suburban campuses that'd be a little more like a college

对，正是如此。

Yeah. Exactly.

那时候大家都在一起工作，而现在更趋向城市化，公司都想更靠近大学和年轻人。

Where where everyone worked together and and and now it's become more urbanized where companies really wanna be closer to the universities and all the young people.

其实我很多辉瑞同事都去了波士顿，辉瑞波士顿分部。

And a lot of my Pfizer con colleagues went to Boston, actually. Pfizer Boston.

对，对，那次调动确实发生了。好的。所以你离开了保罗·纳斯和杰弗里·库克西那里。

Yeah. Yeah. That move was happening. Okay. So you you left to come to the Paul Nurse and Jeffrey Cooksey Yeah.

打电话给你说，嘿。我们在国王十字车站旁边有块空地，我们想做什么来着？

Called you and said, hey. We've got this patch of dirt here next to, Kings Cross Station, and we want to do what?

嗯，我们想创建世界上最大的单一屋顶生物医学研究机构。保罗曾写过一份名为《无边界发现》的战略报告，其实很值得一读。我总跟他说需要补充说明'为什么'这部分，因为那浓缩了他在洛克菲勒大学和全球多家研究机构顾问委员会的经验，比如马克斯·普朗克研究所等地。他提炼出了我认为最精华的部分融入克里克研究所。

Well, we want to create the largest biomedical research institute under one roof. And, you know, Paul had written a strategy called Discovery Without Boundaries, which deserves reading actually. And I always tell him he needs to write the why part because it is a distillation of his experiences in the Rockefeller on the advisory boards from many institutes around the world, right? Max Planck and other places. And he kind of distilled what I think were some of the best elements into the Crick.

决心开展跨学科研究，决心逐步建立庞大的合作网络。这些对新机构来说相当大胆，比如不设终身制——其实有部分终身制，但只有三分之一课题组负责人是终身职位，三分之二不是。在克里克读完博士就不能留所做博士后。

A determination to do interdisciplinary science, a determination to create a massive network over time. And so there were kind of things that were quite brave for a new institute, like no tenor. Well, there is some tenor, but only a third of group leaders have tenor. Two thirds don't. If you do a PhD in the Crick, you can't become a postdoc.

必须离开。做完博士后也不能直接晋升课题组负责人。初级课题组负责人采用6+6年聘期制——保罗认为，新任负责人若只顾追论文，就做不出突破性工作，只是在重复现有成果。

You have to leave. If you do a postdoc, you can't become a group leader. The junior group leaders have are on a six plus six. So six plus six because Paul figured if you're a new group leader and you're chasing publications, you're not doing something that's stretching. You're doing really contemporaneous signs.

你是在别人成果基础上做研究。这并非不好，但他希望人们在二十多岁智力巅峰期真正挑战难题。所以评估标准是：六年后你还在做开创性工作吗？通过考核再续聘六年。

You're building on signs that someone else has done. Not that that's a bad thing, but what he wanted, but people coming in at their intellectual peak in their late 20s and really doing challenging things. So the measurement was, are you doing challenging things after six years? And then you've got another six years.

嗯。

Mhmm.

然后你就离开了，对吧？你培养了这些人，他们拥有非凡的头脑，而你却离开了。

So and then you leave. Right? So you you you train these people up. They have magnificent brains, and you leave.

所以你不会拥有这种终身职位，而人们有时会，你知道的，躺在功劳簿上或者

So you wouldn't have this tenure, and people sometimes get, you know, rest on their laurels or

你会变得僵化。是的，对吧？你可以在某些机构中看到这种现象，对吧？

You would become ossified. Yeah. Right? And that you can see that in some institutions. Right?

他们因为同样的人长期存在而变得僵化。板球运动则充满活力。随着时间的推移，我们已经看到它建立了一个非凡的网络，当这些人离开去做不同的事情时，对吧？博士生去其他地方做博士后，博士后成为团队领导。他们始终保持着联系。

They become ossified with the same characters there. So the cricket's quite dynamic. And over time, and we're already seeing it create an extraordinary network as these people leave to do different things, right? The PhDs go to be postdocs somewhere else, Postdocs become group leaders. They always have a connection.

他们是克里克校友的一部分，这是个便利的地方。再次强调，这是知识核心港口的一部分。无论他们在世界何处工作，都能轻松来到伦敦，前往克里克，进行一系列会议。

They're part of the Crick alumni, and it's a convenient place. Again, that's part of the ports of the knowledge core. It's easy for anyone wherever they're working in the world to come into London, get to the Crick, have a set of meetings.

国王十字火车站交通便利，你可以从那里乘火车去欧洲。但这是现有资源的组合，对吧？比如，那里有

Easy transport at King's Cross Train Station. You can take the train to Europe from there. But it it was a combination of of existing assets. Right? Like, the there were

那里有两所大型的，实际上是三所研究所。米尔山的国家医学研究所，由MRC资助。林肯菲尔德的伦敦研究所，由英国癌症研究中心资助，还有克莱尔霍尔，也是CI UK资助的。它们提供了约三分之二的同事。另一个重要因素，不仅是缺乏终身职位，另一个重要因素是小团队规模。

There were there were two large well, three institutes. The National Institute of Medical Research in Mill Hill, which was an MRC funded. There was the London Research Institute in Lincoln in Field, which was CR UK, Cancer Research UK funded institute, and Clare Hall, which is also CI UK. They provided around two thirds of the colleagues. And the other big thing, so not just this lack of tenure, the other big thing was small group sizes.

克里克研究所的最大团队规模是10到12人。其背后的理念是通过合作开展大规模科学研究。嗯。这也吸引了伦敦的三所顶尖大学参与——国王学院、帝国理工学院和伦敦大学学院，它们都承诺向克里克研究所投入资金。

So the maximum group size in the Crick is 10 or 12 people. And the idea behind that was you have collaborate to do large science. Uh-huh. That also brought in the three big universities in London. So King's, Imperial, and UCL, all committed to the Crick, put money into the Crick.

并且有许多联合聘任职位。我认为这些大学将其视为参与其中、吸引那些原本不会来英国的人才的机会，而克里克确实做到了。我们吸引了一些可能会去其他地方任职的教授。当然，我喜欢说他们来英国工作、生活、恋爱，找到人生伴侣。一旦有了伴侣，人们往往就会留下来。

And there are lots of joint appointments. And I think those universities saw it as an opportunity to be involved, to be able to attract people into The UK that wouldn't naturally come, which I think Crick has done. We've attracted faculty which might have gone somewhere else. And of course, I kinda like to say they they they come and work, live, love in The UK, so they get a life's partner. And if you get your life partner, you tend to stay.

买房子。孩子上学。

Buy a house. Your kids go to school.

是的。然后，你知道，这些教授会同时在国王学院、帝国理工、伦敦大学学院或其他英国大学任职。

Yeah. And then, you know, they're faculty for King's and Imperial and UCL who who take these on and other other other UK universities.

所以这一愿景强调的是速度、活力和雄心，以及合作精神。在学术医学领域，就像在推动前沿和非常基础的科学探索。这是早期的情况。嗯，你知道，那大概是2010年代中期吧？

So this vision was speed, energy, ambition, collaboration Yeah. In in academic medicine, like pushing the frontiers and the Very blue sky science. And this was the early date. Well, you know, this was, what, mid twenty tens?

不，稍微晚些，克里克研究所成立于2014年。

Or No. A bit a bit later, the crick of 2014.

2014年。好的。那时基因组科学确实发展得如火如荼。

2014. Okay. So the gen genomic science was really humming.

当时一切运转正常。对。所以情况就是这样...这里其实有个故事。我深度参与了辉瑞的研发损耗专项小组。

It was humming. Yeah. So that that was so as I yeah. There's kind of a a story here. I've been very involved with the attrition task forces in Pfizer.

嗯。辉瑞和许多大型药企在二月份时都在深入思考研发损耗问题。庞大的产品组合让我们能够分析导致药物研发失败的因素。这是项非常重要的工作——辉瑞当时开展了四个损耗专项小组，投入了巨大精力。

Mhmm. So Pfizer, like quite a number of big pharma companies in the February, was thinking deeply about attrition. Large portfolios allow you to look at what are the factors contributing to attrition, drug attrition. And some really important work. I mean, Pfizer did, I think, four attrition task forces, massive efforts on this.

我们在毒理学方面做了大量工作：哪些类型的化合物会导致毒性？如何更早识别毒性？应该运行哪些安全药理学程序？针对每个项目反复推敲这些问题。还有大量化学合成和制剂开发工作。

And a large amount of work on toxicology, what are the types that were causing toxicology? How could you identify tox earlier? What were the kind of safety pharmacology programs you should be running? All of those kind of questions, refining all of that for each project. A lot of chemistry work, a lot of pharmaceutical development work.

那么如何发现一个真正可合成的化合物呢？我当时就参与了这个环节。

So how do you, you know, discover a chemical that you can actually make? And I I got involved in in that.

为什么这么多药物会失败？为什么我们要花这么长时间才弄明白？是啊...还耗费了这么多时间和资金...

Why are so many of the drugs failing? And and and why why is it taking so long for us to figure this out? Yeah. Yeah. And spend so much time and money and

其中部分原因在当时就很明显。对吧？关键是你需要更深入地思考化学性质，以及化学可能引发的毒理学问题。我们那时确实在临床前研究领域投入了大量精力。你指出的问题非常准确。

and Some of it was quite proximal at the time. Right? It was just that you you need to think more about chemistry and what chemistry might call toxicology. So, I mean, I think, you know, we've, we focused quite a lot in the preclinical domain at that time. The point you're driving at is exactly right.

我脑海中盘旋的问题是：二期临床才是真正的杀手锏，对吧？至今仍是。我之所以关注这个问题，正是因为二期临床的高损耗率——我们对其影响甚微，失败率依然高达80%到90%，具体取决于治疗领域。

What was humming in my head was like, well, big, the big killing for you is actually phase two. Right? And it remains phase two. And actually, why I'm here with the relation is because of phase two attrition. We've really had very little impact on phase two attrition, which remains, you know, as as you know, like eighty or ninety percent depending on which therapy you're in.

在人群中进行随机研究。就是用来验证安全性和有效性的。

The randomized studies in people Yeah. Where you find out if it's safe and effective.

正确。

Correct.

是啊，你懂的吧？但要达到那一步需要投入大量时间和金钱。那你觉得收集更多数据——比如克里克研究所追求的那种——能帮助提供明确方向吗？

Yeah. You you you know? But a lot of time and money is spent getting there. Yeah. And and were you thinking that gathering more of this data like the kind that they were going after at the Crick would would help provide a line of sight?

当时我在辉瑞担任研究部门的药物开发员，作为一名医生，我的诊断结论是：针对人类细胞和组织的研究太少了。常有人反馈说人类细胞组织很难操作，但我就像威尔士梗犬一样刨根问底，发现其实操作本身并不难，难点在于获取样本。我就是从这个角度切入的。

Well, so at that time at that time in Pfizer, I was a drug developer in research, and my diagnosis of tidings as a physician was there's not enough work being done on human cells and tissue. And I would get a lot of feedback, well, that's very difficult to work with human cells and tissues. And then like a Welsh terrier dog, I would be digging into it and think, well, actually it's not that difficult. It's much difficult as getting the samples. And I approached it from that perspective.

实际上2010年离开辉瑞时，我提议开展一项被我称为'研究人性化'的大型计划。这个说法可能不够准确，但本意是尽可能早地利用人类资源。这部分源于我在CCR5基因上的经验，部分源于我观察到我们在非人体组织上做了太多实验——而显然人类才是关键研究对象。我召集了一批聪明人共同探讨。

And actually, as I was leaving Pfizer in 2010, I had proposed that we we run another big effort on what I called humanizing research. I don't think that was the best saying, but what I meant is like, let's use as many human assets as we can as early as possible in the And that was partly genetics based on my experience with CCR5. And it was partly this observation that we were doing a lot of experiments in nonhuman tissues when obviously the organism of importance is the human being. I got a group of smart people together. We thought about it.

老实说后来我离职了。况且当时基因组测序等技术要么不成熟，要么成本不够低难以系统化。但从2010年起，这个想法就像蜜蜂一样在我脑子里嗡嗡作响——在克里克研究所，你说得对，确实该建立这种联系。

To be perfectly honest, I left. And in any case, I don't think genome sequencing transcript there there was a lot of technology that wasn't present or wasn't present at sufficiently low cost to do it systematically. Mhmm. So I had a a, like, a a, like, a bee buzzing in my head from 2010 on this this is something we could do. In the Crick and you're right to do the link.

在克里克研究所的经历让我突然接触到了常规化的测序世界，比如转录组学。我们开始看到扰动性单细胞模块（如CRISPR）的应用。作为首席医疗官，我发现科学家们对数据存储的需求与日俱增——不断有人问我能否再增加一个PB的存储空间，这逐渐演变成了组织管理上的挑战。

What happened to me me in the Crick is I was suddenly exposed to this world of sequencing, you know, being done routinely, transcriptomics. We were beginning to see perturbational single cell modules, so CRISPR. And as a CMO there, what I saw was the scientists asking for more and more data storage. So I was being asked, can we have another petabyte of storage? And that was becoming a organizational challenge.

但与此同时，假设生成的过程仍然相当自上而下。通常是首席研究员深信某个特定通路或靶点很重要——往往就是他们博士期间研究的课题。他们持续推动对该通路的理解，并运用这些方法——你肯定知道，比如悉尼·布伦纳的例子，但

But in parallel, what was happening was the hypothesis generation was still rather top down. So it would be, the PI would have a deep conviction that this particular pathway or target was important, generally the one they got their PhD on. And they were sort of continuing to drive the understanding of that pathway and be using this you know, obviously, you like Sydney Brenner, but

关于假设驱动科学这点，我深有同感。确实。

I to that point about hypothesis driven science. Yeah.

没错。我特别喜欢悉尼·布伦纳2002年接受诺贝尔奖时说的那句话：'我们淹没在数据的海洋里，却依然渴望知识'——他原话用的是'hungry'（饥饿）而非'thirsty'（渴求）。这正是当时的写照：测序技术、转录组学和扰动方法的普及带来了海量数据，但我们仍以自上而下的假设驱动方式在使用这些数据。

Yeah. I I love that Sydney Brenner quote, right, 2002 when when he was accepting his Nobel, his we are drowning in a sea of data, but remain hungry for knowledge. I should've said thirsty, but he said hungry for knowledge. And, you know, that was exactly it, right? We had all of this sea of data emerging because of the widespread access to sequencing, transcriptomics, perturbation of methods, yet we were using the data in in in a very top down hypothesis driven way.

确实。而且制药界有整套标准流程，比如癌症研究会用小鼠模型——对，小鼠、大鼠、狗

Yeah. Yeah. And there's all these standard procedures in pharma. Like, you know, for studying cancer, there's mouse models and Yeah. Mice, rats, dogs Yep.

还有非人灵长类。它们终究不是人类。在获得人体组织数据或进入临床试验前，就已耗费大量时间和资金。结果多年后发现方案根本行不通。

Nonhuman primates. They're not human beings. Yep. You spend a lot of time and money before you even get to some of this human tissue data or or, you know, into clinical trials. And then you find out years later that it didn't work.

不过现在让我们聊聊关联性。稍微快进一下——你提到的这些技术正在集中爆发。

But okay. So let let's talk now about relation. Yeah. Fast forwarding a little bit. All these technologies, you've referenced a few of them, have been bursting on the scene.

对。

Yep.

这提供了一个机会，让我们能够从零开始重新思考如何利用现有工具发现新药。请描述你的思考过程以及这家公司是如何成立的。

Presents an opportunity to rethink from a clean sheet of paper how you'd want to discover new drugs with all the tools that you have. Describe your thought process and how this company came to be.

当时我们注意到数据量非常庞大，但很多都是杂乱无章的数据。也许我们稍后应该再讨论需要什么样的高质量数据。令人沮丧的是，我们的思维方式仍然是自上而下、假设驱动的。作为一名药物研发者，我深知我们在二期临床试验中仍有80%到90%的失败率。显然，我们对人类生物学的理解以及正确靶点的选择还不够到位。可以说，这些主题贯穿了我的职业生涯，最终汇聚在了一起。

So there was an observation that data was plentiful. A lot of it was messy data. Perhaps we should come back to that to talk about what good quality data is required, a frustration that our methods of thinking was still top down, hypothesis driven, and a knowledge as a drug developer, we're still failing 80 to 90% of the time in phase two. So clearly we're not very good at understanding human biology and picking the right target. So it was those kind of themes through my career were kinda coming together, I guess.

后来我遇到了一个叫查理·罗伯茨的人。查理也是英国医生出身，后来成为英国的风险投资人，并创立了Freedom公司（你们应该知道）。但查理确实很特别，因为他花了十年时间研究机器学习并自学编程，与机器学习社区建立了深厚联系——要知道那时候机器学习主要还不在生物科学领域，而是在社交媒体和金融科技等领域。所以他这种认知在当时相当独特。

And I met a chap called Charlie Roberts. So Charlie was another UK physician, became a venture capitalist in The UK, and then founded a company called Freedom, you will know. But Charlie's quite unique actually, because he'd spent a decade thinking about machine learning and teaching himself to code, getting deep connections with the machine learning community, which of course wasn't largely in biosciences then. It was social media and fintech and others. So he had this appreciation, which was quite unique.

我和他之间，你知道的，人脉很重要。我们偶遇过好几次，彼此欣赏，但始终没找到合作的火花。

And he and I, you know, it's like networks are important. We bumped into each other several times, liked each other, but never really had a a spark of something to do.

但他也是医生对吧？

But he's a physician as well. Right?

他确实是医生。

He's a physician as well.

所以你们可以用相同的患者护理语言交流

So you could speak the same language of patient care

也许吧。对。确实如此。没错，我非常以患者为中心，对吧？我们可以始终把患者放在首位。

and Maybe. Yeah. There's yeah. Yeah, I'm very patient driven, right? We can think patients at the start and the end of the day.

他，你知道的，他，我当时正在讨论克里克研究所和我们看到的数据集，并对那些仍在反复推敲的假设感到有些沮丧。然后他说，听着，有种东西叫机器学习。对我来说，那时机器学习还是个新鲜概念。

He, you know, he, I and I was talking about the Crick and the data sets we were seeing and to an extent of frustration about how hypotheses were still being thought through. And he said, well, look, there's this thing called machine learning. And for me then it was this thing called machine learning.

然后

And

你们有这些大型语言模型，有这些主动学习循环。如果能整合起来不是很棒吗？那大约是在2019年底、2020年初，我们关系的真正开端。

you've got these large language models, you've got these active learning cycles. And wouldn't that be cool just to to pull together? And that was the the beginning of relation really in in like late twenty nineteen, 2020.

存在这些庞大的数据集——基因组学、转录组学、蛋白质组学、代谢组学。是的。而且你知道，有些人正在讨论如何将这些数据与电子健康记录连接或整合

That there's these large data sets, genomics, transcriptomics, proteomics, metabolites. Yeah. And, you know, some people are talking about how do we connect this or integrate it with electronic health records and

是的。

Yep.

下游表型，疾病的临床表现。这个话题被讨论了很多次，但人们始终未能真正实现

The downstream phenotypes, the clinical manifestations of disease. It's been talked about a lot. People haven't really been able to

但这技术突然变得具有预见性了，对吧？关键就在这儿，对吧？

But the technology was suddenly prescient. Right? And that's the thing. Right? The

计算能力也在提升，是的。还有机器学习。如果能把这些拼图都整合起来

computing was improving too Yeah. And and the the machine learning. So if you could put all these pieces together

既精通计算又懂生物学的机器学习专家凤毛麟角，而理解机器学习的生物学家也同样稀少。嗯。所以这些非凡强大的新技术虽然正在兴起，却需要彼此互补。正因如此，我从不把公司单纯定义为机器学习公司或单细胞公司，我认为真正的价值在于这种融合。

Very few machine learners who are computational over biologists and very few biologists who understand machine learning. Mhmm. So you have these, you know, extraordinary, powerful new technologies emerging that need each other. So, you know, I think that's why I never talk about relationship as a ML company or a single cell company. I think it's really this integration.

但查理和我还意识到必须与理解研发流程的人合作。因为——容我打个比方，我阅读了大量文献，特别是现在关于机器学习和生物学的。很多研究在回答有趣的学术问题，但并非实用性问题。所以我们Relation在思考时非常谨慎...可能我的团队会跟你开玩笑，但我常说希望自己是房间里最笨的人。

But what Charlie and I also decided was it had to be integrated with people who understood the R and D process, Because, you know, there are My best way of describing this, I read a lot of literature, particularly from machine learning and biology now. A lot of it is answering interesting academic questions. They're not practical questions. So we think here in relation very carefully I mean, maybe my team joke with you, but I've always asked the I said I'd like to be the stupidest person in the room. Yeah.

但我总会问'那又怎样'。我不怕问这个问题——如果我们真做成了，是否能让我比一两年前更坚定、更快速地做出决策？因为正如我们谈到的，这些都不是静态的，高密度生物实验的成本每天都在下降。

But I asked the so what question. I'm not afraid of asking so what? You know, if we do do this, if we are successful, does it allow me to make a decision with more conviction and higher speed that I might have, you know, a year or two ago? Because you have got, what we've also touched upon here is these things are not static, right? You've got the cost of doing high density biology is coming down day by day.

新技术层出不穷，其实在湿实验室就能做很多事，直接动手实验就行。

There are new techniques. So you there's a lot you could do just with a wet lab. Just go and do the experiment.

对吧？那我们具体聊聊Relation公司吧，比如在你加入前公司已经运作了一段时间

Right? Well, let's talk about, like, relations specifically. Like, company had been going for a little bit before you got

到了。或者不。我是被要求监督原始种子投资之间的相互关系的。所以我成为了首位独立董事。当时我基本是和查理、杰克·泰勒·金以及本杰明·斯维纳一起工作。好的。

got here. Or No. I was the so I'd been asked to oversee the original seed investment interrelation. So I was the first independent director. So I was kinda working with Charlie, Jake Taylor King, and Benjamin Swirner Okay.

关于我们在这个界面能做些什么。

On what could we do in this interface.

好的。那么最初的步骤是什么？或者说，最初的愿景和大构想是什么？

Okay. And so what were the first steps to well, what's the vision, the big idea for for starters?

这个大构想是为了更好地理解生物学，对吧？通过综合使用这些工具，因为第二阶段的损耗率非常惊人且没有改善。我们现在有了能做出更好选择的工具。所以问题在于使用哪些工具以及如何组合？公司真正起步是在2020年。当然，新冠疫情爆发了。

Well, so the the the big idea was to understand biology better, right, by using these tools together because phase two attrition is appalling and it's not improving. And we now got the tools to be able to choose better. So the the question was what tools and in what combination? The company got going really in 2020. Of course, COVID hit.

我们主要在研究组合算法。由于新冠疫情，这迅速推动我们开发了一个机器学习系统，它能推荐对新冠可能有效的非常规或意想不到的药物组合。实际上，我们确实与Mila的约书亚·本吉奥合作，并在比尔及梅琳达·盖茨基金会的资助下进行了相关研究。抱歉。

We were working significantly on combinatorial algorithms. And of course, because of COVID that drove very quickly into, you develop a machine learning system that could recommend unusual or unexpected combinations of drugs that might have an impact on COVID. And in fact, we did work with Joshua Bengio at Mila and sponsored by the Bill and Melinda Gates Foundation to look at that. Sorry.

好的。所以你们最初是获得了查理、方舟投资和盖茨基金会的一些资金支持？

Okay. So you got started with a little bit of money from Charlie, Ark Invest, Gates Foundation?

初始资金来自Juvenescence。好的。我们用比尔及梅琳达·盖茨基金会的拨款作为补充。这让我们得以开展关于药物再利用的组合方法研究以应对新冠。

The the initial funding came from Juvenescence. Okay. And we supplemented that with the Bill and Melinda Gates Foundation grant. And that got us working in combinatorial methods around drug repurposing for COVID.

那么，这些联合创始人是谁？他们来自哪里？

Now, who were these co founders? Where did they come from?

是的，查理我已经介绍过了。本杰明·斯威纳是我们中最有条理的思考者。他实际上是神经科学专业出身，虽然不喜欢被称为神经科学家，但乐意被称作风险投资人——转型后的风险投资人。他曾在DN Capital工作过一段时间。

Yeah. So Charlie, I've described already. Benjamin Swearner is the ordered thinker amongst us. He's neuroscientist actually by training, although he doesn't like being called a neuroscientist, but he likes being called a venture capitalist, a reformed venture capitalist. So he'd worked with DN Capital for a period.

他负责确保我们以商业思维运作，当然其他一切也都井然有序。杰克·泰勒·金是个特别的人物，他是位非常优秀的数学家，但后来涉足遗传学领域，尤其对骨骼遗传疾病和单细胞研究感兴趣。他在苏黎世联邦理工学院做过博士后，是早期将数学应用于单细胞和机器学习的探索者之一。这四位就是初创阶段的成员。

So he, you know, he ensures we think in a business like manner and, you know, obviously everything else runs to ship shape. And Jake Taylor King is one of these unusual characters who is a mathematician, very good mathematician, but got involved with genetics, in particular, actually interested in bone, the genetics of bone disease and single cell work. So he did postdoc at the ETH Zurich. So he was an early exploit of mathematics in single cell and machine learning. So these were sort of four characters involved at the start.

各有所长。

With some complementary skills.

确实各有所长，而且他们都有种真正渴望做些不同事情的冲劲。

Some complementary skills, yeah, and, you know, a real sort of hunger to do something different.

嗯。你们最初需要多少启动资金？

Mhmm. And how much money did you need to get started?

我们从Juvenescence获得了100万出头，后来又从比尔及梅琳达·盖茨基金会获得了160万。

We got a million, just over a million from from Juvenescence, and then a million 6 from Bill and Melinda Gates Foundation.

那么这让你在获取工具方面能够做些什么呢？

And what did that enable you to do in terms of getting the tools in Yeah,

嗯，这让我们得以引进公司其他重要人才，比如强大的机器学习专家林赛·爱德华兹担任首席技术官。他在澳大利亚的学术同行Kings那里就一直在思考这个问题。之后他在GSK和AZ工作了十年，将数据科学带入机器学习时代——我这么形容。虽然过程中积累了不少教训，但他给我留下深刻印象的是他也理解药物研发，是那种罕见地跨越多个领域的复合型人才。

well, it allowed us to get some other important characters in the company, strong machine learners, Lindsay Edwards was recruited our chief technology officer. And he'd been thinking about this problem through his peers, an academic in Australia, Kings. And And then he spent a decade with GSK and then AZ, moving data sciences into the machine learning era, is the way I would call it. But lots of scars about how to do it. But he impressed me as someone who also understood drug discovery, one of, again, one of these quite unusual characters that has crossed crossed the domains.

对你们所有人来说最激动人心的是什么？是能够从零开始、不必摒弃旧有观念、完全从头构建的想法吗？

What was so exciting for all of you? Was it this idea of, like, starting fresh and not having to unlearn a whole lot of things and kinda build it from the ground up?

我认为像这样将两项新技术部署到成熟的研发体系中，确实需要这种自由度。这也是我们与大型药企合作的原因之一——比如我们与GSK的合作。GSK其实具备我们所有的能力，他们有自己的机器学习团队和单细胞技术。

I I I think an enterprise such as this with two new technologies being deployed into a very established system of R and D requires that kind of freedom. And I think it's one of the reasons that we do big pharma collaborations, right? Because like we've done this this deal with GSK. GSK have got all of the capabilities in a sense to do what we do. They've got a machine learning group, and they've got single cell capabilities.

而我们独特之处在于能将这些资源整合起来。我常提到跨学科协作——不需要对其他领域有很深了解，但要能沟通。所以我们设计了开放的实验室空间（干湿实验室结合）来促进这种融合。这正是我们的魔法所在。当然我们也愿意讨论具体技术。

But what we can do that's uniquely is to pull these together in an integrated way. You hear me talk a lot about interdisciplinarity with people who have not a deep knowledge necessarily of other domains, but can speak the language. This is why we have a very open laboratory, wet and dry lab space in order to bring these folks together. To create, and I think that's part of the magic here. We'd like to talk about the technologies.

我认为关键在于人员配置，以及我们如何促成团队间的对话——确保计算科学家和生物学家能充分理解彼此的问题。毕竟最终要靠他们产出的结合点来解决问题。

I think the people and the way we connect the people and the conversations we ensure are happening so that the questions are very well understood by the computational people and the biological people because, ultimately, it's the interface of what they produce that has to work.

确实。你们现在有70名员工，如你所说兼具湿实验室和干实验室专长。能聊聊你们的工作流程平台吗？比如...

For sure. And you've got 70 people here now, and, like you say, a mix of wet lab and dry lab, expertise. Can you talk a little bit about the the platform, like, you do, the workflow, what it

看起来如何？因此，这项工作从与盖茨基金会的合作中脱颖而出，转变为一种专注于通过人类遗传学和单细胞研究来确立生物学信念的硬核方向。我们进行了大量沙盒测试，吸收了许多可能有趣的算法，并在药物研发过程中尝试应用。目前，我们有三个最成熟的核心平台，其中包括最优秀的基因映射大型语言模型变体。

looks like? So so it emerged out of that work with the Gates Foundation into this sort of hardcore focus on getting conviction in biology using human genetics and human single cell work. We do a lot of sand pitting, so we take in a lot of the algorithms that may be interesting and try them in the R and D drug discovery process. Our core, most mature platforms at the moment are three. We have the best variant of gene mapper large language model.

我们有一个名为Roslyn的大型语言模型。Roslyn能够沿DNA链上下滑动，本质上是对非编码DNA进行特征提取，并预测受调控的基因，我们可以将这些基因称为疾病风险基因。

So we've got a large language model called Roslyn. And what Roslyn can do is slide up and down DNA and basically featurize non coding DNA and project the genes that are being regulated. So we can call them disease risk genes.

你正在观察暗基因组。

You're looking at the dark genome.

实际上是暗垃圾，我训练了Luke，就是那些垃圾

The dark the junk actually, I trained Luke, the junk

我记得。我知道。他们称之为垃圾DNA，但它并非垃圾。是的。

I remember. I know. They called it junk DNA, and it's not junk. Yeah.

结果证明。

It turns out.

它参与调控哪些基因处于激活状态

It's involved in regulating which genes are on

或者说关闭。我们97%的DNA实际上并非垃圾。嗯。当然，生物学是非常高效的。

or off. 97% of our DNA turns out not to be junk. Mhmm. Of course, biology is very efficient.

但是这些是基因组中的大片区域，你们有机器学习算法可以分析并发现其中的模式吗？

And but this is a these are vast stretches of the genome, and and you've got the machine learning algorithm that can look at that and find patterns?

是的，发现模式。通过GWAS研究我们知道，人类基因组超过90%的变异发生在这些区域，对吧？我们有22,000个基因。在任何一个细胞中，有8,000个基因正在被转录并翻译成蛋白质。

Yes. Find patterns. And, of course, what we know through GWAS studies is well over 90% of the variation of the human genome is in those parts, right? So we've got 22,000 genes. In any one of our cells, 8,000 genes are being transcribed and translated into proteins.

对于更广泛的听众来说，这就像是我们的毛发细胞与心脏细胞不同，与肝脏细胞不同，与肺部细胞或免疫细胞也不同。它们拥有相同的DNA。正是这些非编码DNA确保了这些细胞具有使生物体运作所需的独特功能，同时也驱动生物体产生疾病的独特功能。因此，我们首次利用这些工具研究非编码DNA的结构和特征，以理解基因表达。

And it's one of, you know, for the more general audience, it's one of the, you know, sort of, in a sense, like all of our hair cells is different from your heart cells, different from your liver cell, is different from a cell in your lung or an immune cell. It has the same DNA. And it's this non coding DNA which ensures these cells have the unique function that is required to make an organism work. Also, the unique function which drives an organism to disease. So we're looking, because of these tools, the first time at the structure and the feature with the non coding DNA in order to understand the gene expression.

所以我粗略地称之为CHACCPT4 DNA。

So I loosely call it CHACCPT4 DNA.

嗯。你之前提到获取人类样本。你们收集和分析什么样的样本？

Uh-huh. And you mentioned getting the human samples earlier. What kind of samples do you collect and interrogate?

我们拥有已存在的独特生物样本库，如英国生物银行。50万个体，全基因组测序，大量表型数据。当然，它缺少的是对病变组织的获取和相关疾病单细胞的访问。所以我们首先利用这些专有数据集，包括ENCODE、参考基因组、Phantom five等其他大型数据集。其中一些你也知道是美国的数据集。

So we have unique biorepositives that exist already, like the UK Biobank. 500,000 individuals, whole genome sequence, to whole bunch of phenotypes. What that doesn't have, of course, is access to the disease tissues and access to the single cells of relevance in disease. So what we do is we, first of all, use these proprietary datasets, including others like ENCODE, Reference Genome, Phantom five, many of the large datasets. You will know some of them are US as well.

我们首先构建了一个DNA基础模型，然后利用我们自己的临床研究数据进行微调训练。我们的首个适应症是骨质疏松症，因此我们专注于骨骼疾病领域。

And we create a foundation model of DNA, and then we fine train it with our own clinical study data. So our first indication is osteoporosis. So we're interested in bone disease.

你们是如何选定骨质疏松症的？

How did you select osteoporosis?

因为平台适配性。骨质疏松症是多基因性最强的疾病之一，可能仅次于肥胖症。通过全基因组关联研究(GWAS)已检测到超过1100种DNA变异，因此我们知道有大量遗传信息需要解码。

Because a platform fit. Osteoporosis is one of the most polygenic diseases. Probably obesity is the other. There's over eleven hundred DNA variants that are being measured with GWAS, genome wide association studies. So we knew there was a lot of genetics to decode.

我们还认识到存在巨大的医疗需求。要知道，任何人超过20岁后都会出现骨量流失。女性绝经后骨质流失速度极快。以我这个年龄段的男性为例，患骨质疏松性骨折的风险是三分之一，女性则高达二分之一。然而现有药物通常只在骨折高风险或已发生骨折时使用。

We also know there's a massive medical need. We, you know, anyone who's beyond the age of about 20 is demineralizing. In women post menopausal, they demineralize very quickly. And if you live to be my age, I've got a, as a male, I've got a one in full risk of a osteoporotic fracture, women one in two risk. Yet the drugs that are used are only used generally when people are at very high risk of fracture or have indeed fractured.

嗯。这意味着早期阶段缺乏有效干预手段。即便是现有药物也不适合长期使用。因此存在重大医疗需求。我之前和你讨论过靶向预防的重要性。

Mhmm. So there's nothing that could be used earlier. And even those drugs are not used in a long term way. So there was a big medical need. So I talked to you about target prophysema, why they're important.

因此我们判定这是重大医疗需求。转化路径明确——可以通过骨折预防效果进行评估。我们有个称为骨密度的终点指标，健康志愿者也能进行检测。

So we determined it's a big medical need. The translational pathway is clear. So you can measure fracture prevention. We have this endpoint called bone marrow density. That can be measured in a healthy volunteer.

这类似于超声波扫描。实际上英国生物银行已对大多数参与者进行了跟骨骨密度测量。我们还能通过血液检测关键效应细胞活性，包括成骨细胞的骨形成活性及其分泌物，以及破骨细胞的骨吸收活性。从药物开发角度看，这是相当理想的条件。

It's like an ultrasound scan. And in fact, the bone mineral density has been measured in the UK Biobank so that the heal bone mineral density has been measured in the majority of those participants. And you can measure key effector cell activity in the blood. So you can measure bone building cell activity in the blood of the osteoblast, some of the secretions of the osteopath, and the same for the resorbative cells, the osteoclasts. So from a as a drug developer, that's quite good.

对吧？我能看出有些东西从一开始就运作良好。我们主要做单细胞研究。所以我掌握了遗传学数据，可以预测哪些基因可能很重要。

Right? I can see that something is working right the way back. What we do is a lot of single cell work. So I've got the genetics. I can predict the genetics that might be important.

我们通过用于DNA分析的大型语言模型（类似ChatGPT）已经发现，约有200到300个基因以不同方式表达。重要的是，其中半数基因并非令人兴奋的类型，而是与现有药物的代谢通路相关。但作为药物开发者，这对我意义重大——它表明机器能不带偏见地分析数据集，基于我们临床研究数据精细训练后，不仅能准确预测关键基因，还能发现大量全新基因。

We already knew by using our large language model for our chat GPT for DNA that there were around two to 300 genes that were being expressed in a in a different manner. What was important, half of those genes were not genes that would excite you. They were genes that were involved with, colonial pathways for existing drugs. But in a sense, that was important to me as a drug developer because it said that the machine, in an unbiased way, could look at this dataset with its fine training based on on the data we'd got from our clinical study and predict the right genes out, but also predict a whole bunch of genes that were very novel.

哦，所以它...它指引你们找到了已知的靶点？那些我们确认有效的靶点？

Oh, so you could you you you it pointed you to targets that we know. We know. Are valid targets

没错。

Yeah.

针对骨质疏松症的。

For osteoporosis.

那天我简直要开香槟庆祝了。

Days. I was celebrating.

但还有这些前所未见的新发现。是啊，或许值得一试。

But then there's these other things that nobody's really seen before. Yep. Maybe that's worth a try.

是的。准确来说，这代表着搜索空间增加了143%。所以我们让搜索空间翻了一倍多。接下来我们在实验室进行的第二部分工作，就是利用ChatGPT处理DNA的大型语言模型，以及我们称之为'实验室闭环'的主动学习循环强化学习系统。

Yeah. So and that represented an increase of 143% to be precise of the search space. So we over doubled the search space. So what we do then in the lab, then the second part, so we've got the chat GPT for DNA, the large language model. We've got what we call lab in the loop, which is an active learning loop, reinforcement learning.

我们采用计算机模拟方法来预测单细胞受扰动后的影响。大量运用CRISPR技术，敲除这些新基因后观察其对成骨细胞或破骨细胞的表型影响。不仅限于转录组学，药物开发者青睐此检测的另一原因是——我们还能测量骨沉积过程。这些单细胞会分泌羟基磷灰石矿物质，形成骨骼，这一动态过程是可量化的。

So we have in silico methods to predict the impact of perturbations on single cells. And we use a lot of CRISPR, so we take those novel genes, CRISPR them out, and look for a phenotypic impact on the osteoblast or the osteoclast. Not just transcriptomics, so the other reason why a drug developer likes this assay, we can also measure the process of bone deposition. So these cells, these single cells, they lay down bone. They extrude hydroxyapatite in mineral, and you can measure that as a dynamic process.

因此你可以敲除基因，观察是否影响矿化过程，然后将数据反馈给机器，从而优化你的推荐方案。

So you can knock up the genes. You can look to see whether you impact mineralization, and then you could feed back to the machine and you could refine your recommendations.

你们最初是从样本中收集这些数据的吧？比如附近医院提供的骨骼样本。对，就是手术置换的髋关节和膝关节。你们能获取那些组织

And you're gathering this data originally from samples, like a bone that comes from nearby hospitals. Yeah. Hips and knees that are being surgically replaced. You you can get that tissue

没错。

Yes.

在这边进行分解并做单细胞分析。所以

Over here and then break it down and do a single cell analysis. So

你问为什么选择骨质疏松症——遗传学基础、医疗需求、转化路径这三要素都具备。第二是我们具备单细胞操作能力。骨质疏松症曾一度是研究禁区，但作为初创公司，我们得以整合全球从事骨组织脱矿提取单细胞的实验室资源——当时全世界仅有六家。

you asked why osteoporosis, the the genetics were there, the medical need was there, the transitional pathway was there, the sort of three of the big pillars. The second was this ability to work with single cells. That was for a period of the red light for osteoporosis and indication. But again, because we were a startup, we were able to look at the kind of people around the world who were demineralizing bone in order to extract the single cells. And there were like six labs worldwide.

我们与其中大约四位交流过，目前仍与其中两位保持合作，并引进了他们的方法加以改进。因此我们是唯一能剥离骨组织中破骨细胞、成骨细胞和骨细胞，同时保留细胞原位单细胞图像的工业实验室。关键在于——你也清楚——若从患者组织中取出单个细胞，其危险在于它不再代表患者组织中的原始单细胞状态，而是会发生改变。

We spoke to about four of them. We still work with two and have imported their methods, improved them. So we're the only industrial lab that can strip osteoclasts and osteoblasts and osteocytes off the bone and retain a single cell picture of the cell in situ. Because the problem is, right, and you'll be aware of this, that if you take a single cell out of a tissue out of a patient, the danger is it no longer represents the single cell in the tissue in the patient. It changes.

你会失去背景信息。

You lose the context.

确实会失去背景信息。所以我们一直非常谨慎地保持这一点，这其中涉及大量专有技术知识。当然，你也需要工具来验证是否成功保留了背景信息，对吧？这是个相当重要的流程。

You lose the context, yeah. Yeah. So we've been very careful to maintain that, and there's a lot of proprietary knowledge about how you maintain it. But of course, you've also got the tools to measure that you've retained it, right? So, you know, that's quite an important process.

操作起来相当具有挑战性。但一旦完成，我们就建立了第二道技术护城河——关于如何在实验室循环环境中操控单细胞，以观察基因单一位点扰动对该细胞表型的影响。

It's quite challenging to do. But once we've done it, we've now got this, the second proprietary moat really about how can you manipulate single cells in the lab in the loop context in order to see the effect of genetic single single genetic perturbations on that cell's phenotype.

说到公司名称的关联性，这里面其实蕴含了多重含义。

And coming to the company name relation, this is where those relationship there's multiple meanings in there.

是的。

Yeah.

我十分确信，无论是内部——比如生物学、化学和计算团队之间的协作方式，彼此尊重各自提出的问题和贡献；还是外部——比如与医院合作者的关系，让他们愿意及时提供足够数量的样本供你们开展研究。

I I'm pretty sure, both with internally, like how you get along between the biology and the chemistry and and, computational people Yeah. With respect for what each of you, questions that you bring, to this task. But then there's, like, relationships with collaborators like those at the hospital so that they're willing to give you their samples, enough of them and in a timely manner, so that you can do the research.

实际上远不止如此。所以我一直在做的是建立联系，我想传达的还有与领域内其他专家的关系。我参与过的许多生物科技公司和大药企，其科学顾问委员会往往是每年受邀进行一次对话的人。这里的情况相当困难，也相当不同。

And actually more than that. So I've been And there's a relation So what I like to convey as well is relationship with other experts in the field. So a lot of the biotechs and big pharma I've been involved with, the scientific advisory boards are often people who are invited in once a year for a conversation. It's quite difficult here. Quite different here.

我们大多数科学顾问都持有公司股权，并深度参与公司事务，如果不是每周，至少也是每月。所以他们参与得非常深入。这让我再次感受到，成为房间里最笨的人其实很重要。事实上，大多数聪明人并不为我工作，对吧？这是事实。

Most of our scientific advisors they are, have equity and they engage in the company, if it's not weekly, it's monthly. So they're very deeply involved. And this is something I've felt that, again, has been important about being the most stupid person in the room. Actually, most of the smart people don't work for me, right? That's a fact.

这其实是比尔·乔伊的引述，当时他在微系统公司时，持有与比尔·盖茨不同的理念。当盖茨在招聘人才时，乔伊选择合作。所以必须找到真正高效的合作方式。这其中既有艺术也有科学。我们自认为在这方面处于绝对前沿。

And it was a quote from Bill Joy actually, he said microsystems, When he had a different philosophy to Bill Gates at the time. When Bill Gates was hiring talent, Bill Joy was collaborating. So you gotta find a way to collaborate really effectively. And there is an art and a science to doing that. And we like to think that we are absolutely cutting edge in that.

我们与全球各领域的专家合作，让他们成为项目的一部分。他们常进行与我们项目相关的科研工作。这使我们能够跨越多种适应症，因为我们会转向其他适应症。这一点至关重要，同时我们称之为'关系中的关系'的另一个原因，当然是指我们一直在讨论的多组学关联。

We work with people around the world who are expert in their field, make them parts of our projects. Often they perform some science related to our projects. And that allows us also to cross multiple indications because we'll move on to other indications. But that's been quite critical, as well as the other reason we called relation to relation is, of course, the multiomic relationships that we've been talking about.

观察不同类型信息之间的关系及其如何指导药物研发...确实如此。你提到新增适应症，平台已启动运行，人员也已完成培训并投入操作。

Seeing the relationships between those different types of information and how that guides drug Yeah, for sure. So you mentioned additional indications. You you got this platform up and running. Yep. You you've got people now trained and and operating it.

你们已经处理了约300份骨骼样本...是的，都经过这里

You've had something like, I think, 300 bone samples Yep. Come through here

是的

Yep.

用于分析。

For analysis.

是的，正如我所说，这是地球上最大的骨骼图谱，相当具有挑战性。

Yep. Largest bone atlas on earth for the reasons I described. It's quite challenging.

你们刚与GSK建立了合作关系，我认为这除了骨质疏松症外还涉及其他项目，包括纤维化和骨关节炎。

You just did a partnership with GSK, which I believe pertains to other programs besides the osteoporosis. It's fibrosis and osteoarthritis.

因此骨质疏松症仍属于专有领域。

So osteoporosis remains proprietary.

这完全由我们所有

That's wholly owned by We

选择合作是因为我们认为这个平台在其他适应症上也会非常有用。当然我们希望开发一个具有广泛适用性的平台。GSK是思考这些技术融合的领先企业之一。我们与GSK原本就有良好关系，是他们主动来洽谈纤维化项目的。

to collaborate because we thought that we could the platform would be very useful in other indications. And of course we want to develop a platform that has got broad applicability. GSK are one of the leading companies in thinking about the merger of these technologies. There were good relationships with GSK in any case. They came to us to talk about fibrosis.

我们在纤维化领域的独特视角是利用皮肤纤维化，特别是硬皮病这种病症。其皮肤表现称为硬斑病。硬皮病还伴随系统性病变，尤其是肺部疾病。因此我们能够创建匹配样本——从系统性硬化症患者获取的遗传学和单细胞数据，当然我们已掌握临床表型，并已开始招募患者参与这项名为'皮肤组学'的临床研究。

Our unique angle on fibrosis is using cutaneous of skin fibrosis, specifically in a condition called scleroderma. Then there's a skin manifestation called morphia. Scleroderma also has systemic elements, particularly of lung disease. So we've got an ability to create this matched matched samples of genetics, single cell from patients who've got systemic sclerosis, and of course, we understand the clinical phenotype, and we've already started recruiting into that clinical study. It's called dermatomics.

那么这是一个小分子吗？不，它又是生物制剂？首先我们要理解疾病的生物学机制，对吧？

And And is this a small molecule? No. It's again, it's Biologic? Let's understand the biology of the disease first. Right?

所以我们理解生物学，理解通路，选定靶点。这发生得相当晚，对吧？所以我喜欢说我们的流程是从疾病的DNA到疾病中失调的基因。然后我们将这些基因置于基因调控通路的背景下。

So we understand the biology, understand the pathways, elect the target. That occurs rather late. Right? So I like to say that our process goes from the DNA of disease to the genes that are dysregulated in disease. Then we put the genes in the context of the gene regulatory pathways.

了解通路后，我们就能考虑在通路的哪个环节用药。我们在骨质疏松症研究上正处于这个阶段，与GSK合作的纤维化研究则刚刚起步。我们还开始了与GSK的骨关节炎合作，这建立在我们骨骼图谱的基础上。很多我们接收的髋关节样本患者同时患有终末期骨关节炎和骨质疏松症。

Knowing the pathways we could think then about where we drug the pathway. We're just at that stage in osteoporosis, right at the start in the fibrosis collaboration with GSK. And we've also now started an osteoarthritis collaboration with GSK, which builds on our bone atlas. So a lot of the patients, a lot of the hips that we've been getting in, have got end stage osteoarthritis as well as osteoporosis.

所以

So

我们正在研究这些样本，并开始通过膝关节样本研究骨关节炎的早期阶段。实际上这周我们刚收到第一批膝关节样本。

we're studying those and we're moving into earlier stages of osteoarthritis with knee samples. So we just had our first knee samples in this week, actually.

另一个需求巨大但生物学基础认知薄弱的领域。

Another big area of need and not great underlying understanding of the biology.

没错。确实如此。有些适应症更明显适合与大药企合作，因为他们在这些领域有成熟的能力。比如GSK研究骨关节炎多年，他们对从运动员早期膝关节损伤到终末期骨关节炎的发展过程有着非常精深的见解。

No. Correct. And so we there are certain indications that lend much more clearly to big pharma collaborations because they have established capabilities in those areas. So GSK have been working osteoarthritis for a long time. They've got some very sophisticated views on how osteoarthritis develops from early knee insults, for example, in athletes right the way through to end stage.

坦白说，当他们向我展示时，我完全被震撼了。我意识到，作为一家小公司，我们根本做不到这一点。我们没有他们十年或更长时间积累的深刻见解。这需要长期投入和真正的热情，而问题本身也极其庞大。

And frankly, they bowled me over when they presented that to me. And I thought, actually, that's that's something we could never do as a small company. We don't have that insight of having worked ten years or however long they've worked on this. It's a long time and a real passion to do something. It's a massive problem.

对吧？

Right?

但或许你们可以在遗传学领域用ChatGPT找到折中方案。是的。然后说，算法是以无偏见的方式发现这些的。没错。它既验证了你们已知的部分，也可能带来一些新思路。

But maybe you can meet them in the middle with your chat GPT on the genetics Yep. And say, the the the algorithm has found these in an unbiased way. Yep. And it confirms some of what you know and maybe some new ideas over here.

某种意义上，这就像个漏斗。对吧？从22000个基因筛选到几百个，然后你们开始在硅基实验和RettLab里研究组合。

In a sense, like, it's a funnel. Right? And the funnel from 22,000 genes to a few 100, and then you start working combinations both in Silicon and and RettLab.

嗯。嗯。好的。听起来你们正在生物学领域进行非常艰巨的工作，以更好地理解这些靶点。是的。并提高成功概率。

Mhmm. Mhmm. Okay. So it sounds like you're doing really hard work on the biology to better understand those targets Yep. And improve those odds of success.

那我们回到之前讨论过的一个话题。然后让生物学来指导治疗方式的选择——绝对应该

So come back to something we talked about earlier. And then let the biology guide what the choice of modality Absolutely might

没错。早期有三个至关重要的决定：一是专注于生物学（我们讨论过），二是保持治疗方式不可知论。事实上我曾提出，治疗方式公司正经历一次寒武纪大爆发。

correct. There were two really important decisions early on, about three, I guess. One was to focus on biology, which we've talked about. The second was to be modality agnostic. And, in fact, I made this argument that was there was a Cambrian explosion of modality companies.

对吧？所以如果你有意愿的话——实际上，我在一家名为Nexera的公司担任董事，那是一家GPCR领域的公司。如果你想从事这类工作，就该找该领域的专家。这又回到了协作的重要性。因此我们始终保持技术中立。

Right? So if you want and in fact, I'm on the board of a company called Nexera, right, A GPCR company. If you want to do this kind of work, find somebody who's an expert in the area. It's again, that collaborative piece. So we've remained agnostic.

正如你所说，让生物学特性来决定我们采用哪种治疗模式。因为在辉瑞桑威治研发中心时我就深有体会，我们在化学领域非常出色，可以说是全球顶尖的化学研究团队。

And as you say, allow the biology to dictate which modality we use, because I feel it, like in Pfizer in Sandwich, we were outstanding at chemistry. Right? Arguably the best chemistry group in the world.

研发小分子药物。

Making small molecule drugs.

非常明显。是的。即便有时使用抗体或其他生物制剂可能更合适。由于第三方机构已能开发抗体和小分子药物，我们决定退后一步，专注于实现重大突破的领域——即致病生物学的机理研究。

Very evident. Yep. Even when it might be better to do an antibody or a biologic, other form of biological. So we've, because of the development of third parties who can create antibodies, create smaller molecules, I think we've decided to stand back off that and focus on where we think the big jump will be, which is in the understanding of causative biology.

嗯。新靶点发现。

Mhmm. New target discovery.

没错。

Yep.

而AI在这方面至今尚未取得真正成功。

Which AI has not really been successful at yet.

不。但所有迹象都表明这是解锁多组学数据的关键。

No. But everything points to it being the key that unlocks multi omic data.

你已经做了大约两年半了。是的。你觉得到什么程度你才会坐起来说，哇，真的有效了。这是来自人类临床试验的数据还是更早的数据？

You've been doing this for two and a half years or so. Yeah. What what point do you think what what's it gonna take for you to sit up and say, wow. It's actually working. Is this data from human clinical trials or something sooner

比那更早？我是个药物研发者。对吧？经历过的。情况总是这样。

than that? I'm a drug developer. Right? Living through. That's always gonna be the case.

是的。但我选择骨质疏松症是因为正如我所说，沿途有很多可以展示的实验，对吧？有很好的类器官模型。实际上，骨质疏松症是少数几种体内动物模型效果非常好的领域之一。你给老鼠做卵巢切除，它们就会得骨质疏松。

Yeah. But I picked osteoporosis because there are very show me experiments along the way, as I've described, right? There are good organoids. Actually, the osteoporosis is one of the few areas where the in vivo animal models work really well. You refractomize a mouse or a rat, they get osteoporosis.

目前所有的药物都有效，对吧？所以沿着这条路，我将能够展示我们的新疗法正在发挥作用。在某种程度上，人们会展示它们的能力和改进。我挺喜欢这一点。对我来说，当大型语言模型重新发现已知生物学时，那是个重要时刻，就是我们之前讨论过的。

All the current drugs work, right? So along this pathway, I would be able to show that our new modalities are working. And to an extent people show their horsepower and their improvement. I kinda like that. There was a massive moment for me when the large language model rediscovered known biology, you know, what we've talked about earlier.

我认为那是个重大趋势。

I thought that was huge trend.

这并非完全离谱。

It's not completely off base.

完全偏离了重点。那是一种全新的工具，之前没人以那种应用方式使用过，这很酷。我在其他适应症上运行过这个模型，结果类似。并非完全五五开，已知生物学与新生物学的比例在三七开或七三开之间波动。

Completely off base. That was a a new tool no one had used in that applied way, and that was cool. And I've run the model on other indications, and it looks similar. Not quite fiftyfifty. It varies thirtyseventyseventythirty in terms of the split of known biology to new biology.

因此存在一种一致性：当你面对一种多基因疾病，具有尚未被充分理解的大型遗传特征时，大型语言模型将会发挥作用。

So there's a consistency where you've got a disease that's polygenic, got large genetic signatures that have not been that have not been well understood. The large language model will help.

嗯，嗯。大卫，最后我想问你的是，你知道，这里有很多内容。许多人对AI药物发现的潜力感到非常兴奋，但也有不少怀疑。是的。

Mhmm. Mhmm. Last thing I wanna ask you, David, you know, there's a lot to come here. A lot of people are very excited about the potential for AI drug discovery, and there's also quite a bit of skepticism. Yep.

你认为这个领域会走向何方？这会成为药物发现的新常态吗？技术、生物、工程优先，技术工具驱动的心态？这会成为未来的标准做法吗？

Where do you think this field is going to be? Is this going to be just the new normal, the way drugs get discovered, tech, bio, kinda engineering first, technology tool driven Yeah. Mindset? Is this is this gonna be just the way things are done

几年后？我认为最终会是这种情况。我的意思是，我希望Relation成为生物技术领域的领军企业，打造一系列变革性药物组合。从某种意义上说，这相当于将生物技术融入制药业。

in a few years? Think ultimately that will be the case. I mean, I think Relation I mean, I want Relation to be a sector leading biotechnology company. So, you know, creating a portfolio of transformational medicines. So in a sense that merges as biotech into pharma.

对吧？我认为从零开始构建，以及我们被赋予的空间来实现这种程度的整合非常关键，也很重要。在流程固化的大公司里，这种模式很难建立。所以这一点至关重要。不过我对当前流行的‘AI发现药物’这个概念有些担忧？

Right? I think the blank sheet of paper and the space that we've been given to create this level of integration is really key and been important. And I think it would be difficult to establish in a larger company, which has got defined processes. So that's been very important. I'm somewhat concerned about this AI discovered drug piece that goes around?

因为我认为它非常模块化，明白吗？我描述的很多内容其实是非常传统的湿实验室生物学。所以关键在于整合。

Because I don't think any, I think it's very modular, right? I mean, a lot of what you've heard me describe is very conventional wet biology. So it's integration.

在

At

你如何界定一种药物是由AI发现的还是通过传统湿实验室发现的？所以，我...嗯，所有

what point do you call something an AI discovered drug versus a wet laboratory, a conventionally discovered drug? So I'm I Well, all

你收集的各类数据是否用于指导湿实验室生物学进行发现？

the types of data that you're gathering is to wet lab biology guiding that discovery?

是的。所以我有点反对那些认为AI将带来完全计算机模拟药物的人。我只是...最近有些人公开对此大做文章，我认为应该采取更综合的方法。

Yeah. So I kind of rail against people who think that AI, we're gonna have in silico drugs. I just I just and then there have been some some some people who have been quite noisy about that recently publicly, I think is a much more integrated approach.

你视其为众多工具之一。

You see it as one tool of many.

它是众多工具之一。而且我认为这不会改变。在化学领域，我们可以更接近真正的计算机模拟方法，但仍需验证。因为化学的规则通常已被充分理解，对吧？

It's one tool of many. And I don't think that's gonna change. I think in chemistry, we can get closer to a true in silico approach, but it needs to be tested. And it's because the rules of chemistry are generally well understood. Right?

物理学的规则通常已被充分理解，它们被简化为数学方程。生物学远未达到这种程度。生物学中的关系大多尚未确定。中心法则确实存在：DNA、RNA、蛋白质。

The rules of physics are generally well understood. They're being reduced to mathematical equations. Biology is nowhere near that. The relationships in biology are pretty much undetermined. The central dogma is there: DNA, RNA, proteins.

这一切如何紧密结合在一起，说实话目前仍相当模糊。但我认为现在我们拥有了工具，两项新技术将推动巨大进步，能成为像Relation这样的公司一员，此刻实在令人振奋。

How all of that screws together is, you know, really still rather dark. And but I think now we have the tools, the two net two new technologies to make massive progress, and it's, you know, it's such an exciting moment to be part of the company like Relation.

这确实是个激动人心的时代，几年后再回来看这个故事如何发展将会非常有趣。David Robland，非常感谢你参加《长远之计》节目。谢谢Luke。感谢收听《长远之计》，Timmerman Report出品，Headstepper Media的Pedro Rosado负责音效编辑。

It really is a fascinating time, and it's, it'll be it'll be really interesting to come back in a few years and see how this story plays out. David Robland, thank you so much for joining me on the long run. Thanks, Luke. Thanks for listening to The Long Run, a production of Timmerman Report. Pedro Rosado of Headstepper Media was the sound editor.

音乐由D.A. Wallach提供。下期节目再见。

Music is from D. A. Wallach. See you next episode.