Patrick Short  0:03  
Welcome, everyone to the genetics Podcast. I'm really excited to be here today with Dr. Daphne Koller, who's the CEO and founder of NC tro, a company that uses high throughput biology and machine learning, better drug discovery. Prior to Instituto, Daphne was the chief computing officer at Calico. And she co founded Coursera, which, in my opinion, having taken a number of these courses, as university student helped start really a revolution and massively open online learning. We're not going to talk a lot about this today, we may touch on it a little bit. But this was about Daphne was a professor at Stanford in computer science. So definitely, first of all, welcome to the podcast. And thank you so much for taking the time.

Daphne Koeller  0:38  
Thank you, Patrick. Glad to be here.

Patrick Short  0:40  
I want to start somewhere maybe a little bit unexpected. When I was doing my research for the podcasts, I actually came across your thesis from your PhD, which was fairly enigmatically titled, from knowledge to beliefs. And I read a little bit of it, I definitely didn't read the whole thing. But from what I could gather was, it was really like a fundamental dive into how decisions are made under uncertainty. And your people can't see because it's audio only. But you're, you're rolling your eyes and saying, I'd love to start there, if you don't mind to learn a little bit more about that. What drove that as your initial research interest, and also the the thread from there to what you're doing today. 

Daphne Koeller  1:16  
So I started my career, actually, if you go all the way back to my undergraduate days as a mathematician, and initially was very much enamoured by the conceptual beauty of mathematical frameworks for things that are challenging to model in the world, and, and so my work followed the thread of let's create these really elegant frameworks. And my thesis falls into that category. And at this point, if I look back, I mean, there's some really deep mathematical theorems there, and some really interesting insights. But a turning point for me actually came at the end of my PhD, when I went to Berkeley to do my postdoc, and my postdoc mentor, Stuart Russell took me to lunch, and asked me, so if you were to take your PhD thesis, now, were to give you three exceptionally talented undergraduates to help you with the implementation, which part of it would you actually implement? And I was dumbstruck? Because not only had no one ever asked me that question, the answer to that, as I reflected on it was none. Because it was just it was conceptual, and there wasn't really good for anything. And I think that put me on a journey, to really start to think much more deeply about how I make sure that the work that I do has an impact in the world. And that journey took me on a progression of going from the more theoretical to the more and more applied at every step, starting from work, going from that thesis to working on machine learning, which at least in principle, was more applied then to apply machine learning, not just methods development, but actually applying them to vision, robotics, natural language, then moving from there to something that I felt was even more applied, which is, how does one take machine learning techniques and apply them to biomedical data, which is a transition that took place around the late 90s, early 2000s, as data from patients and from biological data were starting to come available. And then from there, I get, I became more and more applied in how I applied machine learning to biomedical data. And ultimately, that actually was what precipitated my departure from Stanford in going out to found Coursera because it was something that was not an expected part of my career journey. But it was an incredible opportunity to have a lot of impact in a very short amount of time. And I couldn't pass that up. And so in some sense, there was this journey that just kept going.

Patrick Short  3:53  
And maybe you could take me back to where the founding of NC tro started, what was the Where were you? What were you thinking, and what was the kind of spark or the catalyst that got you wanting to start this, which I can see from your T shirt, it's your the four year anniversary coming up or just past it.

Daphne Koeller  4:09  
So actually, if I could go a little bit back to the reason to leave Coursera and go to Calico, because that was, in some ways, the beginning of that journey. That was because I actually been at Coursera for just about five years from the time that we started incubating it at Stanford as a as a company, as opposed to as a internal Stanford project. And so I looked at where Coursera was at the time, and it's a wonderful company, and I'm incredibly proud of what we have accomplished with building that. But it was primarily a company where the impact was derived from having great content provided by the great universities, or at this point, a broader set of organisations, and there wasn't a very significant role for technology and certainly not for science. And so that was to me an opportunity to ask myself work with I personally have the greatest amount of incremental impact. And I felt that Coursera without me would continue on a great trajectory with great people. But what I saw around me and this was in 2016, four years after the machine learning revolution, and seeing that machine learning has been transforming the world around us, but not really having that much of an impact in the life sciences. And when I asked myself why that is it, because there's not a lot of people who speak both languages. By and large machine learning, people don't know very much about biology at all, and most biologists are certainly not trained in machine learning. And so and that's a chasm that makes it very hard to have a meaningful deep application of machine learning methods to life science data with a handful of people who can really make that connection. And so I felt that the impact that I personally could have would be much larger, if I went and build something that probably would not get built. If I if some if I or someone like me, didn't do that, versus just continue on the trajectory that we were on with Coursera, which would likely continue even if I left. And so that was the point when I didn't quite know what I exactly how that impact would translate. I had been out of both machine learning and science for about five years. So I didn't even have an idea of what I wanted to do. And I was fortunate enough to connect with Art Levinson, who was the CEO of Calico, previously the CEO of Genentech. And so that was an opportunity to work with a truly great leader. And I was really inspired by that. And so I went and spent 18 months at Calico and, and realise that ultimately, I was not a biologist, I was still a computer scientist and machine learning scientists. And so what I wanted to do is to build a platform that would help solve biological problems in a very different way and specifically transformed the way hopefully in which we design and develop medicines. And it didn't make sense to build that within a company that focused on a specific biology of the biology of ageing. And so that led me to depart Calico after about 18 months and and yeah form in Sidra, which is four years old this week. 

Patrick Short  7:05  
Amazing. Congratulations, I bet time has felt like it's flown by Yes, how and maybe it's a good segue into how your model is different from so many others. And one of the things that also has always struck me about your company is the variety of diseases and hard to tackle diseases that you focused on from cancer to metabolic diseases like Nash, that don't have any treatment and have been a graveyard for new therapies for many years, to ALS and Frontotemporal dementia, in which I think there's only one therapy that was approved decades ago, that adds a couple months of of life at best. So, you know, barely really scratching the surface there. Maybe you could talk a little bit about your model, and how you how you've arrived at these incredibly challenging diseases that you're focused on.

Daphne Koeller  7:51  
So I think the idea behind the platform that we're building is that our understanding today of human disease is incredibly out of date. The taxonomy that we use to classify diseases is based on very coarse grained clinical manifestations, oftentimes filtered through the subjective lens of a patient and the clinician. And we categorise diseases based on this coarse grained similarity of these clinical symptoms that has very little to do in many cases with the underlying biology of the disease. This is a realisation that has become very clear as we look at oncology, which is the area in which we've had enough molecular data from enough patient samples to realise that breast cancer is not a thing, it is multiple things. And at this point, we have different subtypes of breast cancer, each of which has a very different treatment. And it's what allowed us to move away from chemotherapy, which is truly the lowest common denominator of trying to address cancer and generally ineffective except in the very earliest stages and has horrible consequences to something that is much more of a targeted therapy that treats the underlying causes of the disease, with different treatments for different patients, we've not really had that level of clarity in the broader set of diseases. And that is partially because the data that we've been able to acquire from patients up until quite recently has been quite limited. In cancer by the sort of nature of the disease, you end up biopsy, seeing samples and extracting, and extracting samples from humans and at the same time, cancer cells because they're so robust. They're also very easy to culture in a lab and so one can do experiments on those cancer cells and intervene and and assess different treatments. It's a lot harder to do either of those in the context of most other diseases. That situation, however, has changed in the last few years with a growing amount of both high content data that are objectively that are objective and acquired from it. human patients, whether it's in some cases, less invasive biopsies and other cases, it's, you know, other types of measurements, like MRIs of different kinds, PET scans, and so on. And so we finally have the ability to look at the biology of actual patients and potentially identify patterns there. And on the other side, there's been an incredible set of developments in terms of life science tools, that enable us to interrogate the biology of these diseases in the lab. And that includes the development of cell based systems that are much more similar to human natural cells using this concept of induced pluripotent stem cells, which allow us to take any cell from from any one of us and transform it to this stem cell status, and then create neurons that have our genetics or, or liver cells that have our genetics. And so we can actually start to look at what those cells look like with different genetic burdens of disease. And we can further modify those cells using technologies like CRISPR, in order to introduce disease causing variants and see what that does. And finally, importantly, you can also measure those, those systems in again, very high content, very detailed ways using microscopy using single cell transcriptomics and various other approaches that really allow us to gain a very deep and holistic view of what disease looks like at the cellular level. So with that flood of data, what we have available to us is now an opportunity to really redefine the taxonomy of disease, and then for the appropriate subtypes. Because we know what disease looks like at the biological level, understand what are some things that we can do to protect against or revert disease. And so I think this opens the door to exploring a range of different diseases that maybe share some, the some some characteristics like the existence of genetic drivers. So it's not something that is derived from, you know, lifetime of wear and tear. And it's really hard to model in a lab. So you need to have certain characteristics in order for the technology stack that I'm talking about to be applicable. But with that set of tools, you can interrogate a whole range of diseases. So what we have elected to do it in Seto is really industrialised that process. So create a technology stack that allows us to unlock not every disease, it's not going to be universal, but but unlock enough of them that we can kind of have a much more reliable and less sort of artisanal and stochastic process for uncovering new treatments against those diseases. So we're still trying to figure out which are the right diseases for this type of intervention, the first four years of the company were in many ways in exploration, to figure out which tools are the ones that are most valuable, most robust, most applicable and where they are applicable. Not sure if Nash, in fact is the right place to start, we started there at the beginning of our journey, I don't know that that is the right place where we would want to first deploy this technology. But but we're now getting to a much greater level of clarity on where we can really deploy this and and hopefully start to really reap the fruits of the platform that we've built.

Patrick Short  13:29  
What have you learned about which factors of the both technology stack and disease characteristics that make it a really good fit for your platform?

Daphne Koeller  13:37  
So one thing that we've learned is, first, we've learned that causality is really important, which I think was maybe it's maybe fairly obvious in retrospect, although you'd be surprised just how many drug programmes have been launched based on purely observational correlates with have no causal threads anywhere there. So this is obvious, but

Patrick Short  14:00  
as an example, maybe controversially,

Daphne Koeller  14:03  
it's contrary there is some genetics that that has supported those but but those exist in a very, very small segments of patients and has been extrapolated much more broadly than I think is justified. And maybe that has led to some of the sorry, not some, the incredible number of failures that have occurred in all comers, patient trials. By the way, one important lesson learned here is the importance of selecting the right patient population for whatever intervention and I think a lot of the mistakes that have a lot of the failures that have come about across multiple drug. multiple trials for complex diseases have been that we have in fact, tried to deploy interventions against a patient population that is much broader than it should be, and effectively trying to cure what are most Apple diseases was one drug, which is the thing that really very rarely works. But sorry, I digress. So we learned that causality is critical. We learned about the incredible amount of information that exists in high content data in subtle patterns that people just can't see, that has turned out to be something that we've repeatedly shown across multiple data modalities, both cellular and human, that there's way more information in histopathology, than even an experienced pathologist, for example, can discern same is true for, for cellular images, and so on. So that's not that's one thing that we've learned, we've learned about as one considers the tools and protocols that we use in the lab that if you want to apply machine learning to high content data, the reproducibility and error parameters of the of the tools that you're applying are absolutely critical. Because the more advanced the machine learning that you apply, the easier it is for the machine learning to get confused about things that are sort of artefacts and batch effects and latch onto them and make incredibly accurate predictions that have absolutely nothing to do with the underlying biology. And so that has actually driven the choice in many cases in our lab of methods that are, that are really where the selection is very much driven by how reproducible they are, how robust they are, how much can we protect against artefacts and batch effects, that also has driven a very big focus on automation, because one of the biggest sources of batch effects is when you have different humans doing the same task, or even the same human doing the same human task on two different days where you when they woke up on the wrong side of the bed, or whatever. And then similarly on the exist on the large on the human cohort data, which is the other big data source that we use in our discovery work, really focusing on datasets that are large and well curated and well collected. And in that respect, the work that we did with the UK Biobank and many others, I mean, to me, the UK Biobank is an incredible resource that has driven so much insight across so many different aspects of human biology, human disease, and even just just epidemiology, because it is so well collected and well curated. So the availability of such resources, I think, is absolutely critical to scientific progress.

Patrick Short  17:36  
What is it that's different about what's happening in the UK? You've worked with UK Biobank? I think genomics England as well, what's the UK doing right that other places can learn. And I'd also be interested in the converse. So there are other countries that are well ahead of the game that the UK or the US or other places can learn from.

Daphne Koeller  17:53  
So I think the UK is probably the most advanced of anyone with maybe a couple of exceptions. I mean, Finland, Iceland also have some incredible resources in this respect, some other countries are starting to deploy resources towards that. And I think it was just a combination of having the National Health Service on the one hand, which forces allows the system to track people throughout their life time. And so if people agreed to participate in research studies, there is a wealth of data available on them that can be brought in with their consent, of course, the other is just a number of people within the UK data ecosystem had remarkable foresight about the value of creating these very well curated very carefully collected large population cohorts. And, and, and equally, the willingness to create that as a resource, not to a small, tiny subset of researchers, but really to anyone who is able to do good science and is willing to comply, of course, with the constraints of the of the cohort, which is the insights should be introduced back into the cohorts of the benefit everybody and that, and, of course, the protection of patient or individual privacy. And that I think was just an incredible amount of foresight. And I really hats off to the to the people who put that in place. When I look at similar efforts elsewhere, I find that they often lag behind on one or the other of these dimensions and the way that I think is unfortunate. First, they oftentimes don't make the same investment in really collecting very detailed, dense phenotypes. A lot of the efforts that I've seen are just, you know, Tronic health records, maybe with some amount of genetics, but not the very detailed phenotyping that exists in the UK Biobank. Many of them don't do the longitudinal follow up that the NHS provides, and many of them don't provide the almost universal access to qualified researchers that the UK Biobank has, has done. And I think that's a real shame. Because there's one of the things that has really happened in the last I would say 20 to 30 years in sciences is democratisation of science, is the fact that people all over the world by having access to computers and technology, and so on, can come up with new ideas and new forms of analyses. You don't have to be at a top academic Centre in order to have a really great idea of how to extract value from these resources. If you shut those people out, then I think it just limits our ability to provide benefit to patients.

Patrick Short  20:45  
Why did you personally decide to start in C tro as a standalone company versus trying to go to one of the giants and innovate from within? I'm I, I'm really curious to hear your personal motivation for doing it that way. And the pros and cons of trying to rebuild one of these behemoths into something new versus creating a platform of its own.

Daphne Koeller  21:10  
I agree, it was a dilemma and not an easy decision. There's a lot that if you go to one of the big pharmaceutical companies, for example, there's a lot there, certainly they have a lot of money. They also have, in some cases, a lot of relevant people already in place. And they claim although sometimes with much greater confidence than is deserved to have a lot of data. And while it's true that they do have a lot of data, the data is oftentimes fragmented, not harmonised, not harmonised, a bowl because it was collected by different research programmes with very different approaches. And coming back to my earlier comments about batch effects, artefacts, it's oftentimes difficult if and perhaps impossible to actually create a single unified data resource from from that that is actually valuable. So why did I not choose to do that? I think because fundamentally, I think that if you're looking to do something that is completely different, it's difficult to do it from within a company that has a very significant investments and doing things in the old way. And and groups of people for whom it's, in some sense, their careers at stake their livelihoods at stake to completely change the paradigm. And you are basically fighting that cultural headwind, all the time. And I think that one of the biggest things to really struggle against is what I call cultural inertia. It's just the unwillingness of an organisation to transform the way that it does things. And specifically, what I think is really challenging in a lot of Big Pharma is to take an organisation that has been a lead, where the research has really been led almost entirely by life scientists and and actually give the data scientists and machine learning scientists an equal seat at the table versus having them be kind of a secondary service organisation. Because people don't think of data scientists like that. It's always it's always been the way that you know, the life scientists are calling the shots. And I think that the parody, the equality between those two groups of people is absolutely central to this new paradigm of how we are going to do things in a very different way. And maybe to take an analogy from, you know, from the broader world out there. If you look at the tech enabled giants that have emerged in other parts of the of the economy, in whether it's in, whether it's in media, whether it's in commerce, whether it's in transportation, the really successful companies have almost never emerged as transforming the incumbents. You know, Amazon did not emerge from Walmart. Google did not emerge from the Yellow Pages. And so I think it's it's useful sometimes to kind of just take a blank sheet of paper and say, How should I build this in the right way?

Patrick Short  24:20  
Yeah, I couldn't agree more. What are the other building blocks look like? You're four years in and writing, rewriting the paradigm? I imagine it's going to take some time. What does it look like here? If you're to project it out from starting with that blank sheet of paper to where we are in 510 15 years? How much of that is is kind of crystal clear to you and how much of it is to be determined maybe depending how different technology shifts align and how quickly they come on board.

Daphne Koeller  24:47  
I mean, we are drug discovery is in general, a slow business. It's, you know, the, I think, with the exception of the drugs that were created during the pandemic, which was, frankly a world record that blew everyone's mind, the fastest that you expect a drug to go from idea to approval is five years, the expectation is more like 10. So it's a journey is quite long and has many, many steps that need to go through. And one of the things that as we go through our journey, which is currently focused on early biology discovery, eventually, we're going to need to turn those discoveries into molecules and then for those molecules into the clinic. So as we go through those different steps, it'll be an opportunity for us to assess what has been done so far and, and make a decision on whether that is a perfectly fine way to do this, or whether there are opportunities for improvement. And I think there will always be opportunities for improvement, one needs to be judicious in selecting the ones that are truly going to make a big impact versus Yeah, you can make things 5% better. But is this really the place where you want to invest your energy. So that I think is is, is kind of the journey that we're looking forward into the future. And I don't really know today, whether when it comes time to the sink, take something that's you know, several years out, putting things into the clinic, how, where the technology at that point will be and how much incremental benefit, we would be able to bring over what is the, you know, standard practice. So I mean, just maybe to take a very an example that is very important, as we think about the arc. And and as I mentioned earlier, the importance of really tailoring the treatment to the right patient population versus going the default of let's just give it to pretty much everybody and hope that it works. That requires a certain rigour when you go into clinical trials, and the willingness to deploy technologies and, and data science tools for patient selection, and perhaps also for tracking patient progression and in a way that is more quantitative and rigorous than the oftentimes very squishy and subjective approaches that are used today, where I mean, neuroscience, of course, is a, I think, the easiest to think of example, where you often do that based on very subjective questionnaires that are administered to a patient or caregiver in a very artificial environment that is very stressful for them, where oftentimes, they're trying to look good for the doctor. And so that biases their answer, which is why we see such a huge placebo effect in many neuroscience indications. So I think there's, when we get to that stage, there will, I can already see huge opportunities to make things better. But maybe by that point, someone else will already have made progress towards that. And we'll just be able to piggyback on top of what's already been done. So that's why I don't know today, whether we will have to be the ones innovating in that space or not.

Patrick Short  28:06  
Yeah, I was gonna ask a related question, have you reference drugs, taking five years at best, often closer to 10? I think the average number of drugs approved per year by the FDA is about 40. And there's there's 100 times that many that are going through trials and, and failing at one point or another. I'm curious, what are the what are the big inefficiencies that you see what struck me is you've taking a very engineering approach to this, and I'm sure you see them everywhere, is it? Where do you see the biggest speed gains? Where do you see the biggest success gains? And what is the what is the fastest and most successful? You think that it's possible to push this? Could you get drug approvals down if you design the system from scratch down to down to what what percentage success and what kind of timeline? So I

Daphne Koeller  28:54  
think it's important to be sober and realistic about what is and is not feasible here. So first of all, I would say that the biggest area that I would like to see improvement is first and foremost, and the probability of success because the failures are, are what caused the costs to go up astronomically. And also really critically, the failures are what limits the number of attempts that we have to try and bring drugs forward and therefore leaving a lot of a lot of patients with significant unmet needs. So to me, the probability of success is really the thing that you'd like to optimise first and foremost, and I think that the understanding or the ability to interrogate because understanding is something that people do and not machines, machines don't understand. She's make predictions. The ability to make more reliable predictions about the effects of causal intervention in a human is at the core of being able to make drugs that are more likely to succeed because most drugs fail, because we have just been wrong about what intervention is likely to have a clinical impact in humans. So that to me is absolutely the foremost focus of what it's going to take to this better. Now, in terms of whether it's possible to shift the timelines, I think the answer is a bit nuanced on that. One can certainly shave off time in some of the preclinical work, certainly, potentially engineering some of the discovery approaches, and that's already happening. We're doing it on the biology side, there's various, I think, exciting efforts that have demonstrated shortening of the timelines on the chemistry side going from a target to a molecule. So I think those are great. When it comes time to clinical development, you need to understand what are the pieces that you get to play with, if the disease is progressing at a certain rate, you need a certain amount of time to see the impact of clinical efficacy. And you can speed up human biology, sometimes you might be able to shorten things somewhat, by having finer grained measurements of you know, in terms of an efficacy biomarker, you can see that there's progress that's being made by having a finer grained instrument to measure progress. I think that is a way to shorten things a bit, you might be able to shorten the trial by having a better selected patient populations that you see a larger effect size, that is also part of the, you know, the ability to appropriately select the right group of patients, you have a larger efficacy signal, maybe it's also faster to recruit into the trial. But you have to sort of be sober of if you know, cognitive decline is a multi year process, it's going to require multiple years to see the effects of your treatments. And so we I really dislike hyperbolic promises that are just not aligned with reality.

Patrick Short  31:54  
Yeah, what was remarkable to me was how quickly as you mentioned earlier, vaccines were developed and approved. And there are some, I think there's some really interesting lessons to learn that, but there are also potentially some red herrings as well, because things were done there, that probably wouldn't apply to all 7000 Odd diseases we need to treat, but something to learn for sure.

Daphne Koeller  32:16  
And I would say, as you look at the evolution of the pandemic, some of those things that worked really well, at the beginning, when there were lots and lots of patients all over the world, as it became as the pandemic evolved, and people developed natural immunity because of the they had been exposed to the disease, and, and so on and so forth. Some of those trials also became harder to execute and more expensive, and so on, because you didn't have as many cases in either the case or the control group. And so that already became harder there as well. So no, I don't think one could fully extrapolate from the pandemic. But I will say there were certain elements that I found really inspiring about some of the clinical trials that were conducted during the pandemic that I think we could all learn from, like the recovery trial that was done in the UK. And I know I keep coming back to the UK as an example, because they've just done so many things. Well, the recovery trial, which was an umbrella trial across multiple different candidate drugs against SARS, cov, two, that measured that was structured in a way that allowed many more hospitals that are not sort of the the academic centres that are so limited, and are the ones that typically participate in trials, a lot of much broader participation than just those. I think that was one of the very few really impactful trials on everything except the vaccine trials, which were, which were obviously, you know, the largest and, and brought us but there were so many failed trials on hydroxychloroquine. And by that I say even when I say failed, I don't mean because the drug didn't work, it didn't. But because of the trials were underpowered, Ill constructed. And we're not able to provide that clarity on whether the drug worked or didn't. And the recovery trial was designed in the way that allowed those questions to be answered quickly, effectively, and at a relatively low cost and in a way that allowed a much broader participation by both hospitals and patients. And so I really hope that someone's watching and saying that that form of trial design that allows multiple drugs to be tested, especially in when you have diseases with significant unmet need, that we should be constructing the trials much more broadly.

Patrick Short  34:39  
Yeah, I couldn't agree more. I wanted to come back to something that you mentioned at the very start, which was that you growing up as a mathematician really found something beautiful and in trying to understand something at a very fundamental level. I also come from math background, one of the things that initially bugged me but I think I've come to peace with a little bit is how messy biology is And there were some very elegant formulas in your thesis that just don't apply in the machine learning models and the neural networks are a, as messy as it gets. I'm wondering if you could talk a little bit about that aspect as a mathematician computer science and wrapping your head around what can effectively be a black box, in some cases that's needed to understand some of the complexity or maybe you don't see it that way. But I'm really curious to hear your thoughts on that. Yeah. So

Daphne Koeller  35:25  
first of all, I have to say, I'm shocked that you actually read my thesis, I meant to say that first chapter, I

Patrick Short  35:31  
didn't read the whole thing.

Daphne Koeller  35:32  
Thank you. I'm glad to hear that. I mean, I thought that was long buried. So I mean, life is messy. And I think one of, to me, one of the growth, I mean, to me a growth trajectory that I've had as a scientist is the realisation that life is messy. And I think all of us need to make a decision of whether we would like to go towards conceptual, elegance and life as we think it should be, versus life as it actually is. And I would say that, that, in general, the willingness to look life on reality straight in the face, and say, I'm gonna take it as it is, and not as I would like it to be as one of the things that we all need to do, as we grow up. To kind of maybe take that to another place that I found to be really important in, especially in the last three years as a leader trying to lead a company is the realisation that we did not want to have the pandemic, and it sucked. And and, and there were a lot of people that I saw around me who tried to basically ignore reality and say, yeah, no, this can't be happening. So I'm going to pretend it's not. And one needs to sort of be willing to look reality in the face, and then adjust one's actions to that rather than to what you wish reality were. So to me, if you want to have an impact in the real world, you need to take the real world as it is. And that was the decision I made when I decided to go into something that is more impact driven is you have to take it with all of its warts. And that is both true on the science side, this also true on other aspects of what it takes to get a drug through which is what was what would it take to get it through regulatory? What would it take to get it through commercial approval, to get payers to pay for it, because if they don't pay for it, the patients don't get it. So all of those pieces are not elegant. They're not fun, but they're necessary. If you want to have an impact.

Patrick Short  37:37  
I'd love to close actually on something a little different, which is your perspective on company building. And I know that company culture is something that's really important to you, I'd love to hear your philosophy on this and how four years in you've gone about building things and etc.

Daphne Koeller  37:52  
So I touched I think on some of those issues, when I answered your question about Big Pharma and why didn't go and build and Sutro within an existing organisation. It's because I think that fundamentally, to build a certainly the kind of company that we're building, which is something that has never really been built before, which is an organisation where you absolutely need to have people from different cultures and different types of experience, who normally don't talk to each other. And normally think about the world in very different ways. Because the way an engineer thinks about the world is much more, it's not linear. But but so if you're looking for patterns, you're looking for the general principles that underlie a lot of messy data that you see. Whereas if your life scientist, oftentimes, you're actually looking for the exception to the rule, you're looking for that thing that doesn't fit because the thing that doesn't fit often is a new insight or a new discovery. And so you think about the world in completely different ways. And how do you bring the best of both of that type of thinking into a single environment? How do you get these people to collaborate, to communicate, to speak a common language. And so building a culture that captures that and actually enables and enables that and makes it the norm is really hard? And it requires having that be sort of at the base root of everything that we do you hire to that you congratulate that when it happens, you criticise when it doesn't. And you also set up organisational structures that make that the norm rather than the exception. So for example, in our NC tro, first of all, one of our foundational values is that we engage with each other openly, constructively and with respect. Each of those words matter. The engagement is that communication. The openness is the openness to asking naive questions and to accepting ideas that run counter to how you would normally think about things constructively means you have to do it in a way that is From a place not of being the smartest person in the room, but really making the answer better, and the deep respect that needs to happen that needs to exist for everyone who's at the company, because we all bring something really important to the mix. And so that is a thread that runs through everything. And we design projects around that ability to come together and people from different disciplines who are working as a single team. And so I think that, to me, the standard approach that exists in most other organisations and by the way, i not i don't mean just in industry, but also in academia is the siloed kind of divide and conquer each group working on his own thing. And occasionally they get together to, to transmit information, I don't even say collaborate, that's just not going to work for the kind of company that, that I think we need to build for this type of new way of doing things.

Patrick Short  40:52  
Just as a quick follow up on that, how do you organise, especially as you grow to the size that you are to prevent that siloing? What do you do specifically to make sure groups that don't normally talk to each other actually are collaborating and not just transmitting information.

Daphne Koeller  41:08  
So we've organised things that and see throw in ways that small companies normally don't do and have maintained that as we grew, which is that while people do report functionally into someone who understands their work, and and is able to provide them with the kind of professional guidance that they need. So data scientists report to data scientists and you know, chemists report chemists and biologists, biologists, all of the work is done in cross functional project teams, with a project team lead, who often by the way, as a junior person isn't even necessarily a functional manager. It also gives great growth opportunities to people to be able to lead a project at a relatively early stage in their career. And the projects include people from all the disciplines within the company and the teams come together to construct their work plans. And oftentimes, it's not by saying, Okay, here's the problem that you need to solve when you all figure out the solution together, but rather, figure this is kind of the direction that we'd like to go figure out what's the right question to ask. And we find that when you put people in the room with these very different perspectives, they don't only come up with better solutions, they come up with better questions that we would never have thought to ask. And some of the best work that we've done has emerged from some of those cross functional collaborations.

Patrick Short  42:25  
Last question, I promise, how do you find those people who are really comfortable with that ambiguity, and can take the really broad, we want to go generally that direction? How do we get there even we don't know which direction we want to go? Where should we go? How do you find those people in a finite window of interview time?

Daphne Koeller  42:42  
It's I agree, it's hard. And let me layer another type of difficulty, which is you have to hire people who want to work with people who are very different to them and are comfortable with looking stupid, a good fraction of the time because they have, they have no idea what that term that someone else used even means. And so that's another level of discomfort that a lot of people who are very successful in their discipline are oftentimes unable to come to terms with, which is I don't always know the answer. And I sometimes asking really naive questions. So I think that is, admittedly a challenge. I think that there's a self selection, because there's people who just like to be pushed out of their comfort zone. And they naturally gravitate to startups in general, and to our startups, specifically, because of because of the ability to really reach out and, and learn something new. And it's an incredible growth opportunity for the people who appreciate that. So it's definitely an interesting hiring experience. And we're quite selective, but we've been able to collect incredible group of individuals. And one of the things that I'm really proud of, is that when we have people come to visit, whether it's our investors or candidates who are looking for a role, and they meet the team, almost universally and irrespective of whether they end up joining us or not, for whatever reason, they always say that this is one of the most incredible group of individuals that they've ever seen at a company and some really very privileged, I've been able to recruit such an incredible team.

Patrick Short  44:26  
Great. I'm really glad that we covered this because we talk a lot about precision medicine breakthroughs and new companies but not often enough about the teams that that put them together. So thank you, I really appreciate you taking the time. Thank you. Thank you everyone for listening. And as always, please share the episode with a friend if you liked the episode and of course, leave us a review on your favourite podcast player. Thanks again and see you next time.

Transcribed by https://otter.ai