Jonathan Wolf: Some people were having these crashes for certain meals and when they had those crashes, it was creating hunger and that hunger was actually leading those people to eat 200 calories more. And 200 calories more is a lot. If you were doing that every day, you could put on, I don't know, 10 kilos or something over the year. And that had nothing to do with the calories at that point, it was the way that it was changing your behaviour in the future.
Dr Rupy: Welcome to the Doctor's Kitchen podcast. The show about food, lifestyle, medicine and how to improve your health today. I'm Dr Rupy, your host. I'm a medical doctor, I study nutrition and I'm a firm believer in the power of food and lifestyle as medicine. Join me and my expert guests where we discuss the multiple determinants of what allows you to lead your best life. Today I'm talking with tech entrepreneur and co-founder of the nutrition app Zoe, Jonathan Wolf. We dive into his story of starting in physics, moving into the world of tech during the tech boom and ultimately applying his knowledge of how machine learning works to the complex world of nutrition science. Now, if you're a podcast fan, you'll know just how convoluted nutritional science can be, the pitfalls of dieting as well as the lack of personalization. I'm a firm believer personally in the utility of more investigations to determine our health status and ultimately help us achieve more consistency in our healthy habits. And you'll hear about how Zoe is trying to do this. Today, we talk about Jonathan's background in startups, the citizen science projects, why nutrition science is so confusing and misguided, the vision for Zoe, the future research studies, and also we round up our conversation with what actions Jonathan has taken personally to improve his health. Remember, you can download the Doctor's Kitchen app for free to get access to all of our recipes with specific suggestions tailored to your health needs. We are adding new recipes every single week and Android users, bear with us, we are working on a version for you guys too. Also, whilst you're on the doctorskitchen.com website where you'll also get the show notes for today's episode, check out my newsletter Eat, Listen, Read. Every week I give you something to eat, something to listen to and something to read to help you live a healthier, happier week. And lots of you have indicated how much you're enjoying those as well. We're growing the newsletter bit by bit and I'm taking on board all of your feedback weekly as well. For now, on to my podcast with Jonathan Wolf. So Jonathan, thanks so much for coming on the show. I want to deep dive into a bit about you because your background is super interesting. It's a bit of a turn from the norm in terms of what we usually talk about on the podcast. We usually talk to researchers or doctors or people from the medical profession, but I really wanted to chat to you about the background behind why you started Zoe, what you were doing before, where the motivations came from, and really dive into the success thus far of the company and where it's going. So why don't you give myself and the audience a bit of background into who you are and what you've done before Zoe.
Jonathan Wolf: Fantastic. And I guess maybe it all leads off up, obviously to Zoe at the end and I guess a lot of listeners here may have heard of us either for personalized nutrition or many people know us because of the the Zoe COVID app as it's referred to. And you're right, I think if you'd been talking to me 10 years ago, you'd be really surprised to hear that I was entirely obsessed by nutrition and working with all these research scientists. So I originally studied physics and I think I had this idea when I was leaving school that I really wanted to understand the way the world works. And it's like, well, you should go and do physics, right? That's like the fundamental way to understand this. I started studying physics as an undergraduate and I realized that there was no way I was going to be a physicist, that basically you ended up becoming this world expert in this tiny, tiny thing that seven other people in the world understood and you sat in a lab all the time or maybe in an office with a pencil. I was like, I'm too much of a people person who's going to drive me mad. And it's it's a very, very tough life. Now I see this working with lots of research scientists. So I think I have an even better view of this. Like you're you're fighting really hard to get funding for years and years. It's a huge amount of your time is spent raising the money rather than actually getting to do anything. And it's very hard to take really radical views because actually people, the funding is all sort of defined, so it's got to be very, very incremental. And you know, I was that might be true in physics, it's definitely true in nutritional science or anything like that. So I basically decided, you know what, that was sort of interesting, but I wasn't going to to do that. And so when I left, basically I fell into tech, like many people did. This was at the end of the 90s, so it was the start of the sort of the dot com boom for for listeners who are old enough to to remember this. And so I think if you had that interest and you know, I did programming and things like this as a kid, that seemed like a natural place to to end up. And I had some interesting experiences with some businesses that grew really fast and exploded in the early, you know, sort of 2001, 2002. I then ended up going to Yahoo back at a time, which again, if listeners are old enough to think that it sounded like a company that you would like to go to, which is definitely not true now, which tells you how fast sort of tech moves, right? Compared to, you know, maybe, you know, other industries where no one's going to say, oh, you were mad to go and join, you know, Nestle or or whatever. And basically what I realized after a lot of time was I was just trying to find the right job that was really I could be passionate about. And I kept doing these different things that I didn't really love. And I ended up leaving Yahoo back in about 2009, so about 13 years ago, and joining a very early stage startup that was using machine learning, actually, you know, on online advertising. So, creating ads. And over the next sort of sort of seven years, we took that from, you know, no revenue and no people to 2,000 people and $2 billion of revenue using very large scale machine learning in order to create individually personalized ads to try and sell, you know, shoes and holidays and all the rest of it. And it was a lot of fun. It was a lot of fun partly because it was technically very challenging. And I think many people listening to this are not necessarily aware of just sort of how much technology is going on on the back end if you are scrolling through your Instagram feed, for example, like how much technology is being applied to try and figure out, you know, what is the next piece of content that is going to be most relevant to you, or indeed, you know, if you see, you know, an ad pop up, how are they trying to figure out that that is the ad that is going to be most likely to work. So we were doing that. It was fantastic. We were sort of building a company, which is which is very exciting because it's a bit like creating a a family in a way, all these people who are working together and you're trying to make that a great environment. On the other hand, we were not curing cancer, right? What we were doing is we were helping people who sell shoes to sell slightly more shoes, you know, people who sell flights to sell slightly more flights. And I don't think there's anything bad about that, but I think it clearly wasn't like really making the world better. And at this point, my my son was about six or seven and I just started to think that he was becoming old enough to understand sort of what I did. And I thought basically, you know, the biggest thing that affects you really, I think, as for me at any rate is like you see your parents and what do they value and what do they do. And I felt like actually I wanted to do something that had more purpose. I was lucky enough out of this success to have a lot of freedom. So I decided to take a sabbatical and basically try and figure out what I wanted to do when I grew up. Okay, I was already 40, but you know, I felt like I've got to sort of figure this out. And I think I had, you know, two real things that I felt. The first was that I wanted it to have some some more purpose and meaning than what I had been doing previously. And the second one is I'd seen this amazing power of machine learning to be able to do things that are just not possible with people. So this idea that if you can collect enough data, you could suddenly let these machines automatically figure out something that is personalized to you in a way that just hasn't ever been possible before. And I think, you know, maybe the best analogy now for people listening is, you know, anyone who has tried something like Google Translate or has, you know, spoken to Siri or, you know, Alexa or any of the rest of it and literally it understands the words you speak, right? And it then gives you back an answer. Like, if you think back even 10 years ago, that sounds like magic. And I think that is the best example of machine learning when it's really good is it does something that just feels like magic, but it does it in a really stupid way. So like we learn things, you know, I have a I have a daughter as well, she's now three, you see this how she's learning. She, you know, she's really figuring out how all of this stuff works. But the way the machine learning works isn't like that at all. You just give it this vast amount of data and a really, really dumb machine basically can look at all of this data incredibly fast. So instead of having to be really smart like a human being, it's just imagine you've got this perfect memory and you can instantaneously look at millions of things. Well, in that case, I can actually understand your voice, I could turn it into text without ever really understanding what a voice is or what text is. And so I think that was the the second part was how could we use this amazing tool, which was mainly being used for internet advertising, right? Which felt like it's sort of like you've discovered the elixir of life, but you're only using it to, you know, help dogs be more successful in the Crufts dog show or something. That's not a great analogy, but you see what I mean? It's like this is amazing technology and it mainly wasn't being used, it felt for the things that were really important. And so I left and I'm sorry, this has become a long story about my background. I ended up meeting my co-founder, Tim Spector, and I have another co-founder, George, who I've worked with for many years and is is wonderful and also has, I'd met at Yahoo, has an internet background. But Tim, and I think he's been on your your show in the past, right, Rupy?
Dr Rupy: Yeah, yeah, that's he's been on a couple of times.
Jonathan Wolf: So, I was lucky enough to hear Tim speak. He'd just published The Diet Myth, which a few of your listeners might be familiar with, which I think is a fantastic book. And George had read the book, I'd gone to the talk because I'd been very interested in nutrition for a long time, which, you know, we may end up talking about. And I just thought that's amazing. He was talking about this twin study that he'd been doing for 25 years, the fact that even twins that are identical clones were different, and it was because of these microbes inside them, and that everybody had these completely different microbes. And then he sort of got stuck because his problem was, well, so everybody should be eating something different because they've all got different microbes. We can see this and their responses are different. You know, I can see this with these twins, but I have no idea, Rupy, what to tell you to eat that is different from what Jonathan should eat. And basically the light bulb went off in my head and I'm like, that is literally what machine learning is for. If you can collect enough data, then suddenly you could do something where you could do a test, you could take a set of your results, you could compare it with thousands of other people doing these tests and ultimately millions of other people, and therefore actually get advice that is personalized to you and where that advice can get better and better over time because as more and more people participate, they can actually share their results with the rest of the the community and that is how, you know, this results get better and better because just as we see with, you know, the way that, you know, voice recognition has got better and better, it's because the number of samples and the number of people using it is getting more and more. And so I thought, that's amazing and that sounds like it's something I'm personally sort of passionate and excited about. And you know, I think it was clear to me even then that what we eat is so central to our health. And I think now, you know, six years later, actually, I think I underestimated how important nutrition is to our health because I think at that point, I thought about it as as many people do very much to do with, you know, do you put on weight and you know, do you get maybe sort of cardiovascular disease, whereas I think now, you know, talking to so many researchers, there are such strong links to things like mental health and depression and we saw some really amazing links between diet and your immune system and in fact how sick you got with COVID during COVID, which were a number of papers. So, so actually I'm I'm even more convinced that what we eat is is just sort of central to our to our overall health. And that is the that is sort of the journey from Jonathan thinking he wanted to understand the world to trying to help create Zoe with with Tim and and George.
Dr Rupy: That's great. I mean, like, if you ever need to answer that question again, you can just point people to the first 10 minutes of this podcast and just say, listen to this and then you'll have a complete overview of why a physicist went in, went on to become a tech person who went on to to found or co-found Zoe. So, no, really, I'm personally fascinated in the trajectory of people's careers and where they get to to working today and and where that may direct them going forward as well. And I think your background as a physicist is super impressive. We actually had Professor Jim Al-Khalili on the podcast a couple of months back explaining to me the concepts behind quantum entanglement. And if you ever want to listen to someone being basically confused for about 60 minutes, listen to that because that's me for about an hour trying to trying to get to grips with it. I mean, it is confusing and even he admits that, you know, you are meant to be confused by it. But um, but yeah, that's so that was interesting. And then also, just to anchor the listener because I think this is going to be particularly important as we talk a bit more about the application of the investigations and the underlying technology, the difference between machine learning, artificial intelligence, you know, the word algorithm gets used quite loosely, I would say, by by both people in the in the medical world and the scientific world as well as people selling different types of products in various industries. So maybe we could just anchor the listener into what we mean by machine learning versus the sort of concepts of of artificial intelligence. And then we can talk about, you know, I'm fascinated in the relationship between between yourself, Tim and and George and and how that came together as well at the very early stages.
Jonathan Wolf: No, I'd love to talk about that. So firstly, I've used the word machine learning because it is less fancy and ambitious sounding than artificial intelligence. Actually, I think you should think about them as the as the same thing. It's basically saying this computer, you know, does something that makes it seem like it's intelligent. It's just that I think artificial intelligence tends to conjure up this idea that it's like a human being, right? So it's sort of like generally intelligent and can, you know, understand things. And the reality is we've made no progress so far in terms of creating like a general intelligence. So, you know, for those of you who are worried that, you know, Skynet is going to take over and blow up the world, that still seems to be a long way away, I'm pleased to say. So in general, what we're doing is these sort of building these pieces of software that are incredibly good at a very specific task. And where that approach, the reason why you sort of talk about the machine learning is because you're feeding them all of this data. So they're learning from all of that data and then applying it. So if I, you know, if I think about it, make it specific about Zoe, so the whole idea about Zoe is that you can do this test kit at home and you want to understand for yourself, you know, what are your results, how do you respond to food, and then you want to get individual guidance to understand like here's a program just for me and how do I respond to all the foods that I normally eat and how might I change it. And so what the approach is to say, okay, how can we compare you with, you know, now tens of thousands of people and and their results, and so then the machine can take all of that in and sort of use all of the other information to compare you against other people. And so that is why I tend to use that word machine learning, but you're right, actually, I think often on the website, we might use artificial intelligence because I think that is a more generally used word, but I think one that, you know, I think conjures up this idea of the of the robot overlords that are going to to take over. And there are companies working on that as we speak, but so far we don't seem to be any closer and although there are these amazing breakthroughs actually in sort of the machine learning and you, you know, you may hear of some of them because, you know, DeepMind, for example, based in London has been really important. They continue to be in fact in sort of this more narrow approach rather than really anything that they're all still really stupid compared to my three-year-old girl, you know, that's what's interesting.
Dr Rupy: Yeah. Yeah, absolutely. It's one of the one of my favourite podcasts actually is the one with Hannah Fry at DeepMind and she's she's going into the offices. I think she's been working there for the last couple of years. Hannah Fry is a professor of mathematics and the way she describes everything is just very, very simple and it's very understandable as well, even for something as complex as artificial intelligence, general intelligence. And I do like the way you made the parallel between how we've currently used that form of intelligence to optimize for, you know, sales and the commerciality of it. And I think that's always going to be the starting point for these things. But I've always wrestled with the sort of reasons as to why we haven't applied that to medicine. And actually during COVID, I was helping out in ICU for a bit and I was reflecting on when I was a junior doctor, whatever, 10 years now, and my jobs as a junior doctor were to be filling out the blood work every single morning, right? And putting the HB here and the CRP and then mapping the trends. And luckily we have computers that do that automatically, but there's no interpretation of that data. And when I was on an ICU and I was we're helping out with the the consultants there and looking at all the the numbers as we were privy to it during MDTs, I remember thinking, depending on the day in which your ICU consultant is working and the trajectory of your condition at that at that severe point and the numbers that are being displayed on that day, you can have a very different sort of management plan. And there isn't like an underlying machine that looks and ingests all that data and gives you some indication of what might be the best practices compared comparing this center to perhaps another center in the middle of Italy or another center in the middle of the US. That hasn't the leap hasn't been made there in acute medicine, let alone preventive medicine, which is, you know, Zoe. So I wonder if you had any comments on that or whether you see an application of what you guys have created in in other forms of medicine.
Jonathan Wolf: So firstly, I have no experience of intensive care. I'm not a doctor, so let me let me sort of start with that that disclaimer. I think that, um, firstly, I would say this, but I think there is an enormous opportunity for application of data and these sorts of tools into health. And you know, at the end of the day, there's basically nothing more important than your health, right? I think we all know this because as soon as we're unwell or one of our loved ones is unwell, like the whole world gets turned upside down and Rupy is a doctor, you see this all the time, right? You're, um, and my wife is a doctor, she's a dermatologist. And so, um, clearly there's sort of nothing more important. So then I think the question is, well, hang on a minute, you know, everybody's picked up their phone and it does all of these amazing things. And let's be honest, like you might be addicted to Instagram or Tik Tok, but it's not really as important as the thing you've just talked about about somebody being in intensive care. So why is it that there's so much more technology working on your phone than than in this? And so, uh, there are some reasons. And, um, I think I got some visibility of this, um, through the whole COVID experience as well as, um, obviously what we've done through through nutrition because I think one of the things that was interesting is in COVID, suddenly lots of things happened in six months that previously might have taken 10 or 20 years. Um, and so famously, of course, vaccines normally take, you know, at least a decade and we had this amazing thing where we actually had vaccines being delivered to millions of people within 12 months, right? So that was that's a great example. But, um, you know, I think our own app is another example of that where suddenly we we launched this, um, uh, actually maybe I'll take a second to give to give the context for that to to make any sense. So we built this, um, uh, this app in March, uh, the very beginning of the pandemic, um, that was designed to try and understand what was going on with COVID. And basically, we'd sort of sent everybody home from the office a few weeks before the official lockdown because I think it seemed pretty obvious to us and I think many people in the country that it was mad that people were still being allowed to go and mingle. And so we sent everybody home because, you know, who knows how that was going to end up, you know, taking COVID to somebody who was, um, you know, old or frail or whatever. Um, and Tim, Tim Spector, uh, called me up as, uh, as I was at home, as I was going to be sitting, in fact, in in the room that I'm sitting right now and have spent most of the the last, uh, couple of years in. Um, and said, look, I've had this really interesting idea because we don't understand anything about COVID. So what about if we, um, took the app that we've built for all these clinical studies for these, uh, for nutrition, and we've been running, you know, uh, what had become the world's largest nutrition science study and built all of these tools and apps to be able to do all of this remotely. He said, what about if we took that tool, uh, and we repurposed it for COVID, and I gave it to all my twins, and they can just track whether or not they feel sick, and then I can link it back to sort of 25 years worth of history, um, of these twins, and we can start to try and understand like why are people getting sick? Because again, Rupy, if you think back to that point now, it's like, oh, it's really obvious, it's mainly how old you are, and it's also maybe about, you know, whether you have, you know, diabetes or you're overweight. But at that point, nobody had any idea what was going on. And you remember all these stories, right, about doctors dropping dead from COVID because they've been exposed and at that point, you know, Northern Italy, those hospitals had collapsed and the death rate was really high. And we were sitting here thinking like the NHS is all going to be overwhelmed. You know, my wife was being told as a dermatologist, she was about to go and work in intensive care. And Rupy's smiling because you all know that when the dermatologists are being called up into intensive care, things are quite bad. You say fair, Rupy?
Dr Rupy: I would say so, yeah. Yeah. My wife's a dermatologist. She tells me that's an acceptable joke. Even dermatologists accept that because they specialize, right? So they they did their general training, but then for 20 years, they only look at skin. So this is obviously a long way from from intensive care. So, um, you know, things were pretty bad. And so his idea was, well, could we try and understand why, you know, maybe these people who have particular genetic risk or particular blood pressure or whatever. And, um, Tim comes up with basically a brilliant idea twice a day. So most of my job is to tell Tim that that's a brilliant idea, but we can't do that because we're already working on these three really great ideas and it's going to take us, you know, another year to make those happen. So I think he was really surprised when I said, Tim, that's a brilliant idea. Um, and if we're going to do that, though, let's build the app so it can work for millions of people because I think the joy of technology, this is, you know, this is its strength, right? Is if you can build this to work for a thousand people, if you have a good engineering team, it's really easy to get it to work for a million people or 10 million people. And I said, and if we could do that, then we could apply, um, you know, large scale data science to actually understand what's the level of infection, we can build machine learning models to predict because there was no testing, if you remember back to that point, you know, the first wave, nobody knew if they had COVID, right? Because we decided it, you know, it wasn't really that important to, uh, you know, to test or understand what was going on. And we could try and predict basically at this point, we were worried like where which hospitals will fail first and can you help the NHS to plan so that maybe, you know, the rates are three times higher in London than in Birmingham, and in Birmingham, therefore, it's going to be in three weeks time, you'll start to see this, um, this huge, um, flow of people and that you could try and organize. Um, and so we decided to build this app and, um, our engineering team was amazing. In fact, our VP engineering, Julien, basically didn't sleep for three days and pretty much single-handedly built the first version of, um, of this app. It was a wild time. I remember, and it was scary. I remember feeling like it was a bit like being at war. And if you were a doctor, right, it was really clear what you could do, or if you were, uh, in for a key worker and you were doing something really important, like making sure that we still got food and all these sorts of things. But if you're just some sort of company co-founder, it's like you sort of feel useless. And I remember my my grandparents saying, well, in a war, like everybody contributes and has to do something, um, towards this. Uh, and so for us, I think it felt like we we should basically stop what we've been doing and and do this. And we we launched the app and amazingly, literally in the first 24 hours, more than a million people downloaded the app. Um, and, uh, you know, six months later, it was four and a half million people, uh, both here in the UK and then ultimately in the US and Sweden as well, recording every day, just very simply, their health, and then if they weren't well, sort of their symptoms. Um, and this is a very long way of of coming back to your original question about how, you know, you can apply this stuff to health and why it's hard. Once we had these millions of people participating, suddenly, we were able to get a really accurate picture of COVID across the country. And not just on a total picture, but we could actually understand in a local region. So, you know, I'm sitting here in, uh, in a part of London, what's the level of COVID in London? And we could do that even before we had, um, tests because actually, we could take the symptoms, we could then compare those with the smaller set of people who ultimately did tests, uh, and we could understand and build a prediction models for whether or not you had COVID. And that was actually one of our first nature medicine papers in COVID was about building a, uh, a prediction model based upon these symptoms. But why am I telling you all of that that story? It's because once we had these amazing members, and they're still actually almost a million people who are still participating in this. It's now become the the Zoe Health study and and maybe we'll talk about that that later. Um, if you have these amazing participants who are willing, you know, to to do this every day, basically to try and improve, uh, everybody's health, help fight this, um, this pandemic, then suddenly that gives you the raw data to get some really smart data scientists working with some really smart scientists to to look at that, and then you can do amazing things. And so I think the point about like why is it that you're sitting in the intensive care room and you're not getting this support, but then you go on to Tik Tok and you get like the perfect thing just for you, is that in, you know, in Tik Tok's case, there are billions and billions of pieces of data that are being produced every day showing what people like, uh, what they click on, what they don't. And so you sort of get all of these relationships. Um, in health care, in general, all of this data is completely locked up in little separate areas. None of it is linked and available to, you know, some sort of data scientist to try and work on it. And if you haven't got that information, then you can't actually build anything interesting. So in order to get the information you're describing, you need to have lots and lots and lots of examples of different patients going through this with all the information around them about, um, you know, their health and their background health issues and what drugs they've been on. And like because it's very complicated, isn't it? What you're doing.
Dr Rupy: Yeah, exactly.
Jonathan Wolf: And so what we were able to do sort of with COVID is because there was so much data, actually we could then apply these tools and start to understand, um, and in fact, um, we built some prediction models that could predict basically at day five based upon your symptoms, you know, really quite accurately whether you were going to get better on your own at home or whether you're going to end up being hospitalized. You know, just as another example, just based upon the track of your first five day symptoms. And I think that's actually, you know, it's a super exciting idea about the future of medicine because I think, um, you know, you could definitely imagine something like that as a as a GP in the future, you know, if somebody starts to feel sick, I actually believe if you were logging your symptoms, um, daily, it's quite plausible that, you know, that pathway of symptoms for many other things could really basically drive this sort of red alert signal as you described to your local GP to say, you know what, actually that pattern looks bad. I really think that, you know, you should call this person and there's a good chance that you're going to say, you know what, you should go down to A&E because that doesn't look bad, versus somebody else, right? Where actually you know, when you look at that, um, pathway, you can feel really confident. And in some senses, this is what GPs do themselves, right? They have their own learning over time, but of course, they can't monitor like all of these different patients. And so I think the opportunity is not to think about any of these tools replacing doctors, but actually to think about them as these amazing tools that, um, could support them and allow, uh, allow them hopefully to be focused on a lot more of the high value, um, decisions and just have, you know, because the one thing that machines are good at, right, is they they never sleep, right? They're there all the time, they can look at all of this data. So I believe that ultimately, you will absolutely see the sort of, um, support tools that you're describing in intensive care and everywhere else. Um, but it does require access to data, it does require a willingness to innovate. Um, and I do think, you know, during COVID, there was a huge willingness to innovate, but in general, you know, I would say, Rupy, you know, the healthcare profession is pretty cautious, right? About, um, uh, sort of new technologies because, uh, they deal with life and death situations. And I think that, um, that that is also probably, you know, means that these things may not spread as fast as they might do in other other sectors. I don't know if you think that's that's fair.
Dr Rupy: Yeah, yeah, I think it's definitely a hindrance to adoption, um, within the NHS. There's definitely like the new generation that are more bullish on the idea of utilizing tools like predictive models. And I think you've demonstrated the the use case for that and how that could be applied to different areas that aren't optimized like, and I think it's going to come out of necessity, let alone if if it's not from, um, the the sort of desire to have more technology into the into the healthcare system, it's definitely going to come out of necessity because if you look at GP waiting times, if you look at A&E waiting times, we're going to have to think more laterally about how we reduce the footfall going through healthcare centers and more medicine in the home place. Um, and I think, yeah, the symptom tracker is a is a really good example.
Jonathan Wolf: I think the thing I would add to that is wouldn't it be wonderful if we focused more on preventative medicine and less on fixing people once they're they're really sick. And of course, I would say that sitting here being obsessed by by nutrition and lifestyle intervention. Um, I would say that, uh, you know, as a host of a podcast which is basically all about the idea of, you know, how can you get this actionable advice from the scientists. But but I really believe this, you know, there is something really crazy, like the example would be imagine that you never ever serviced your car, it got, you could hear there's all this stuff going round, you know, wrong with it. It's like you're starting to hear squeaks and like the brakes don't work very well and you can see the tires are really run down. You're like, I'm absolutely not going to do anything until I crash. And when I have a crash, then I'm going to go and do this incredibly expensive like reworking of the whole car. And that is basically the health care that we have, right? It's like everybody's walking around and it's not like, I mean, Rupy, I'm not a doctor, but I think it's fair to say that most of the time, it's not like they instantaneously go from completely healthy, I would have had no idea this was going to happen, and then suddenly they get, you know, a really serious disease, whether it's, uh, you know, a stroke or a heart attack or diabetes, right? I mean, there are this is a long and steady, um, transition for most people. Is that is that fair?
Dr Rupy: I would say so. I mean, if you think about the data flows from multiple aspects of our life, our health is sort of like right at the bottom. If you ask people anything about their their bank balance, how much fuel is in their car, how many steps they've done that day, you know, we have like a lot of information and we're we're comfortable dealing with that. But I think the sort of uh propensity of looking at ourselves and our own bodies in the same way is not it's not really the same. And I think, you know, in a in a decade, maybe even less, it's going to think, we're going to think that we were crazy for for not doing this stuff, for not, you know, investigating more regularly. And I'm bullish on that idea, but I think it has to be sort of um, it has to be balanced with the health anxiety that obviously comes with this. I mean, I'm privy to a lot of this stuff when we talk about food and nutrition, you know, it has to be tailored in the right way. Um, but we can get into that in a bit. I just wanted to double uh double click on the um the fact that you spun this app up so quickly, really speaks to the quality of the engineering team because as someone who has created an app that launched early in January, and we literally spent like years and it's not it's not a particularly uh smart app in in the same way that um, you know, we're not using machine learning or anything like that. We have a very simple algorithm in the background. But that that was pretty amazing. So how how quickly did you create this symptom tracker after Tim had the idea?
Jonathan Wolf: So I think it was five days in total from deciding we were going to do it to to launching it. Um, now, just to be clear, we didn't start this thing completely from scratch, right? So we had already completed, um, the first two Zoe predict studies. These were both, you know, for nutritional science, huge studies, a thousand people in both of those studies, which, you know, for our listeners, that might sound like, oh, a thousand people doesn't sound very big, but, you know, it's important to realize that most nutrition studies are 20 to 50 people. So these were huge studies, um, that we'd already built and we'd done something very novel, which is, you know, one of the things I think about that the reason why I think at Zoe, you know, we've been lucky enough to really, um, you know, I think crack some problems that we haven't been able to crack before is, um, scientists use this word multidisciplinary, which I think in normal life people just mean people with lots of different backgrounds and lots of different understanding. And in general, in science, things are very, very, um, narrow, a bit like doctors, right? So doctors, you know, you're a GP or you do acute medicine or you're a dermatologist, so you do skin. And you after you you never do any of these other things, right? So and you hang out mainly with all the other people who do this as well. So you create this great expertise in an area, which is amazing. And so, you know, you could know, you know, exactly what this is. And so, you know, if you have cancer, you want to go and see someone who's like not just even a general cancer specialist, right? But probably going to specialize in this particular thing. However, if the problem itself is much broader, the challenge is you need to bring in maybe 10 different people who are specialists in different areas. And normally the way that science works, and I think it, you know, it's a bit a bit similar in in medicine, I think as well, Rupy, right? Is it's not used to being so collaborative across all of these different specialties. It's much more designed to get sort of deeper and deeper in these verticals. And so I think what we had figured out and it took took us a long time and it was, you know, really because Tim was able to help us and guide us and, you know, because he's so credible, you know, he's he's one of the top scientists in the world, right? Um, we were able to get to these different people. And so some of them might be an expert on, um, microbiome sequencing, which is how you actually take, you know, the, uh, you get a a stool sample, right? A sample of your poo, um, and you smash it up to get out the DNA and you sequence it. And this is amazing complexity to, um, uh, to get this. This is one of the hardest sort of, um, data problems there is. And then you need to go and talk to somebody who's used to carrying out like large scale human trials. These people would never normally talk to each other at all. And then you need to go and speak to somebody who really understands, um, uh, maybe blood sugar and diabetes. And then you want to go and speak to somebody who really understands, you know, uh, lipids and cholesterol, all these different things. So you're sort of pulling all of these together. And and a big part of what we realized was, um, part of why these studies are so hard is that everybody still uses technology from, you know, 50 years ago. So, you know, there's, I think you were describing pieces of, uh, paper, I think a bit earlier, Rupy, weren't you in some part of the hospital experience? Is that right?
Dr Rupy: That's correct, yeah. We still, uh, I think one of the biggest users of fax machines in the NHS in Europe. Um, so yeah, it just speaks to, uh, the
Jonathan Wolf: And most of your listeners will say, I haven't seen a fax machine for 15 years, right? So that's, um, that's crazy. And do you still use pencil and paper for anything or is that now finally gone?
Dr Rupy: Well, we do definitely ingest a lot of the notes from previous decades into the system, but they're usually just like, um, PDFs that can be read by the computer. We, I mean, uh, we do do pen and paper on ward rounds and stuff, but then we it's usually typed in straight into the computer afterwards. So we are we are adopting some, uh, things. It's nowhere near as fast as it should be because really we shouldn't really be typing anything. It should be dictation with the with the level of, uh, uh, um, with the amount of information that can come out of an MDT or a ward round, it should really be ingested straight into a computer rather than, you know, manually typed as well. And just to to to, um, go back to the MDT, I think within certain elements of healthcare, we do do that sort of collaborative element quite well. Um, for example, with, uh, urology, you know, you might have MDTs where you have the radiologist looking through the images in front of the nephrologist, the palliative care team, uh, some of the specialist nurses, the flow nurses, as well as the the urologist as well. We do that in various elements within the healthcare system, but certainly that needs to be accelerated and I think there are different ways in which to to do that too. Um, so sorry, you were describing, uh, the way in which all these different specialists would collaborate on this on this very, very tough problem. And what how, I mean, how often does that happen? Is that is that happening regularly? Is that, you know, was that in the at the inception of the idea? Like what what's talk talk us through that?
Jonathan Wolf: I think it's, um, it's all the time. So I think the the central realization was, um, you know, if the objective is to say, you know, Rupy, how can you do this test? So we're going to send you this test. This is what happens now. So this exists. This is a, uh, a commercial product that's now available, you know, in the UK, um, and the US. And we'll send you this this test kit at home. We're going to do a series of of tests. Um, those tests, uh, map to the original clinical studies that we've, um, been doing these series of studies and sort of growing on. And then how do we take the results from those tests and use them to go and give you a set of results, which you get, uh, a few weeks after, uh, after doing these tests. And then how do we then create this program? And that program is now supported by coaches because we realized that actually it's it's no good just to give you like this, you're you're nodding because you're so obvious to you as a doctor, but you can't just tell people this is what you need to do. You actually need to really help them. And so in the end, a huge part of the the, um, uh, the whole project has been focused much more on the program and how we help you to understand how to take your existing diet and change it sort of with support to this diet that is, uh, that is tailored to you. Um, and what it turns out is understanding all of this involves all of these different experts. And I think one of the reasons that we've been able to do this is that we realized you can't just go and talk to someone who's a nutritional scientist, you need to get all of these different expertise. And so there's people sort of that come much more from the sort of data science background that I've come from and we and we've talked about, which is something which has historically not overlapped, um, you know, with nutrition at all and, you know, and hasn't overlapped with a lot of of medicine. But then you've got, um, people who are really, you know, you have to talk to people who are experts in many different areas. So, um, food is incredibly complicated, how food then interacts with us. So there are people, as you know, who who's entire academic career is about understanding, you know, blood sugar responses after meals and how does that tie into particular things. So we we have, um, a set of core scientists who are deeply involved in, uh, in Zoe, and then many more who are on the scientific advisory board, uh, who we go and engage with and who our scientists go and engage with, um, you know, regularly, but not all the time to try and understand how do we go and improve, uh, this advice. And then the other the other bit is we built all of this technology to support this. And so, you know, one of the the first things we realized was if we wanted to do these clinical trials, uh, it's too expensive to bring everybody into the hospital. And so in our first study, you know, a thousand people, mainly twins, everybody came into St. Thomas's Hospital in London for a first day, they were, um, cannulated, which basically means, you know, as you know, Rupy, uh, you, uh, you you basically stick a tap in somebody's arm so you can take out blood whenever you want. Um, and this allowed us to measure, uh, you know, basically on about 10 intervals after you ate these standardized meals, exactly what was going on inside you. So not just your blood sugar and your blood fats, but things like your inflammatory markers, all sorts of, um, uh, things that were going on in there. Then we sent you, um, home for 10 days and we we gave you a really structured, um, uh, set of meals to eat and collected a lot more data. So we needed to build, uh, a way to collect all of this in very high quality way because historically, as soon as you left the hospital, basically you couldn't collect any data. And this is why, um, it's one of the reasons it's been so hard to do nutritional science studies because people eat food all the time, they're at home, how do you collect this data correctly? And basically new technology makes this possible in a way that wasn't possible before. And the phone is the most important thing because it suddenly means you can build these apps, you can allow people to record, you know, exactly what they're eating, you can get them to take photos of what they're eating so you can confirm and correct issues. You can also use that to connect to other devices. So there's now, um, a blood sugar sensor that you can put on someone's arm for two weeks and that can record their, uh, blood sugar levels every five minutes and send it back to the phone and then, you know, that can then send it back, uh, to us. You can also do things like do an at home, um, blood test after eating a standardized meal and you can use the phone to give users sufficiently accurate instructions that it's it's very close to being as good as in the clinic so they can understand you can eat a standardized breakfast, you can eat a standardized lunch, you can then do a blood test in in this particular case, it's two hours after that lunch, uh, and that allows us to understand what is happening to, um, your your blood fats or your cholesterol, your triglycerides after eating. And that's really important because understanding where we are at fasting, which is how historically people have tended to think, you know, I think if anyone on the, uh, you know, listening to this might well have been asked to do fasting bloods at some point, right, by their by their doctor. That only tells you sort of what happens in this very stable state when you first wake up, but that's not actually our real life state. Our real life state is we're almost always, you know, in this after meal point. And what we discovered, this was in the first predict study and and ended up in the first nature medicine paper is there's this huge variation in how people respond to food. So once you stop measuring this just in the fasting state and start to say what happens if you understand what happens after food, you start to discover lots of people who might seem to be, um, you know, managing really well in terms of their their sugars and and their fats, but once you start to measure them after eating, you you're sort of getting, um, you know, our scientists believe a a sort of early indication of metabolic problems. And so what's amazing is because of these new technologies, we could do this remotely. So we built all of these capabilities, that's allowed us then to roll this into the, um, the Zoe, uh, product and take those same capabilities that we've built for these clinical trials and actually make that available as a, uh, as a consumer product and so continue to scale now, um, you know, we're more than 30,000 people who who've done this and I know it sounds wildly ambitious, but, you know, I I believe we could get to millions of people ultimately. And once you do that, you can start to understand many more areas of personalization because we're so complicated. You know, and I hadn't appreciate if I'd realized how complicated we were at the beginning, I don't know I would have done this. If I'd realized how much we don't understand as scientists, I don't know if I would have done this either. It turns out that, you know, again, it turns out we don't know as much about the human body as maybe I had been led to believe. And I feel that doctors are partly to blame for this because I think when you go and see them, they, you know, they make you feel like they understand what's going on. And don't get me wrong, they know lots of stuff, but actually, there's an enormous amount we don't understand, right, Rupy?
Dr Rupy: Oh, that's a huge, yeah, there's a there's a huge sort of, uh, void in our knowledge, um, particularly as it pertains to how we individually, uh, respond to not just foods, but also drugs, uh, you know, how successful we we are our our sort of, uh, our pharmaceuticals are going to be as well and how how individuals can, um, can can respond to them. So there there's definitely like huge voids there and then there's definitely no shying away from that. Um, I want to talk about the process of the of the Zoe program because I I Tim sent me, uh, I was a beta tester for you guys. So he he sent me a pack and I really enjoyed it. It was really good to sort of, uh, put the CGM on and we talked about CGM.
Jonathan Wolf: You can tell me, how did you find it?
Dr Rupy: Yeah, no, I loved it. It was great. I think, uh, the process, uh, you had to sort of commit to it, you know, put the CGM on, making sure that you're you're logging your meals and stuff. Uh, the the muffin, as someone who eats pretty well most days, the muffin was, uh, you know, I had to force that down because I'm usually not like snacking on on cookies and and all that kind of stuff during the day. So for me, that was probably the hardest bit, but for most people, I don't think you're going to have an issue. Um, and, uh, yeah, it was it was it was very interesting to see.
Jonathan Wolf: And just to be clear, you don't have to eat lots of muffins. This is a standardized scientific muffin for those listening, uh, that we do, which which allows us to match the exact macronutrients actually of a standard, uh, UK, uh, breakfast and and lunch, and which we've now given to, as I said, like 30,000 people. And so because of that, it allows us to understand, um, the impact of, uh, both carbohydrates and fats on your system. But you're right, interesting, Rupy, some people absolutely hate it, but you'll be shocked to hear that there are a whole set of people who come back and say, that's delicious. Can I have more of these muffins? So it just goes to show that there is individual variation not just in terms of our body's responses to food, but definitely to, uh, to taste as as well.
Dr Rupy: Yeah, yeah, no, absolutely, absolutely. Um, so let's, uh, I you've you've mentioned the predict studies. I think just give us an overview of of what the outcomes were. We don't have to go into too much detail about, you know, methodology and all the rest of it. But like, what are the outcomes? What are the sort of main takeaways from those preliminary studies? And then I know we're sort of jumping around a bit, but I do want to go into the origin story of how you, Tim, George sort of came together, uh, created the company, raised, all that kind of stuff. Like sort of diving into the entrepreneurial aspect of it because I'm personally interested in that element and I and I think a lot of our listeners are as well.
Jonathan Wolf: Perfect. Well, let let me, I mean, I'll I'll I'll I'll take that in turns. So, um, we have an ongoing program of, um, publishing papers. And I think one of the things about Zoe that I think is really important is, um, we are trying to make the information that we discover as widely available as possible. And so doing this test for yourself costs money, okay? So this is, um, I I can't pretend, right? Like this is not yet something that is, uh, you know, available to every single person, um, in the country because we're using, um, still, you know, these tests for your, uh, for your gut, for your blood sugar, all that. So there's cost this costs money. And so one of the things that we really believe in is how can we try and disseminate a lot of what we, uh, discover and make that available for free. Um, and so one part of that is the podcast. And the idea was that we were just getting this chance to speak to all of these brilliant scientists who were doing really cutting edge research and that often what people understood, um, you know, in terms of what what the general public understood about, uh, nutrition was potentially like 20 years behind what these experts were thinking. And so these experts are all saying, yeah, everybody understands now, like this particular link between food and depression. But actually, like nobody really understands it. Or everybody's like, well, of course, nobody believes in low fat. And then you're like, no, but there's loads of people walking around the streets who have been who are still, you know, influenced by this big public guidelines that came out in maybe the 1980s. Um, and so there was this really big gap between speaking to, um, sort of these leading researchers and just what would happen if you go and speak to, you know, your sister or your mother or or anybody. And so we're like, you know what, wouldn't it be great to just allow people to hear directly from these research scientists, like in their topics where they're experts, but in a way that was designed for a regular, um, audience. So the idea is you don't need to be a scientist, you can just understand it. And, um, also, how about can you make it actionable? So rather than just very theoretical, um, can you, um, uh, actually understand like specific things that you can do. And that's, um, that's worked really great. But the other thing that we're, um, uh, doing is publishing lots of scientific papers. Uh, and I think in total, we've published more than, uh, 50 scientific papers. And, um, and that is about sort of sharing back with the academic community what we're discovering. And it's one of the reasons that all of these great, uh, scientists will will come and work with us. We've already published a, um, a whole series of papers that came out of the, um, the initial studies. And there's a lot of them, so I won't talk about all of them, but I think there's there's maybe three that I I think I'm, um, particularly proud of that came, um, from from nutrition. Um, and so the the first two came out of that first study, the first Zoe predict one study, these a thousand people, mainly twins, um, and the first paper basically showed there was this extraordinary variation in responses of healthy people. And I think generally, uh, this was a big surprise. This was why, uh, I think Nature Medicine published it because historically, um, the idea is, you know, when you're healthy, everything's sort of, you know, everything is sort of like in a relatively narrow band, and it's then as you start to get sick, right, that you start to see that your blood sugar gets bad and that's what tells you that you might have diabetes or that your lipids are really bad. And what we saw, which is really extraordinary, was sort of a tenfold variation in almost everything we measured. So we saw it for blood sugar, we saw it for these these blood fats like triglycerides, we saw it in inflammatory markers, really just anything we measured, we saw this sort of tenfold variation after people were eating food. And so the point was that eating food, which is a completely normal thing we need to do in order to live, creates this sort of cascade of, um, different metabolic, uh, effects. And our body, you know, it's sort of a shock, right? When we when we eat because suddenly we're getting this big jolt of these different, um, things, whether it's blood sugar or or blood fat or whatever, but our body is built to deal with that. What was really surprising was the variation. So you see some people, it's really amazing. You see some people who would eat these muffins, this whole meal across two meals, which gives you about 50 grams of fat, and literally you cannot see any change in their blood fat. Like they just they just deal with it effortlessly. And then you see other people who sort of eight hours, um, after the initial meal, you see these really elevated levels of of blood fat still. Um, and the, um, you know, the scientists who are experts there explain like that is a really significant risk factor. That's what starts to, um, create, uh, this sort of inflammation and other things that ultimately lead to, um, sort of furring up of your arteries and things like this. But what's amazing is, you know, you can't tell on the outside, right? These people, they don't look different. It's not like you can just tell, oh, it's obvious that, you know, Bob is going to be, uh, in trouble and Jill's going to be fine. And interestingly, if you just look at their fasting bloods, in many cases, they also look the same. So I think that was the first paper that's really amazing is is this idea that like this idea of personalization is true. And we weren't really sure to be honest until then. So it could have turned out, you know what, personalization is really important if you're, um, if you're already quite sick, right? So let's say if you have diabetes or or something like that, but earlier than this, you know, everybody is sort of the same. We don't see that at all. We saw this amazing variation and that was the first nature medicine paper. Um, the second one, which I think was was even cooler, was specifically about the microbiome. Um, and that's always been the center of our research is understanding how these microbes are linked to everything that happens. And we haven't sort of talked about that much today, but, you know, this is really central to to all of our our research. And, um, what we found for the for the first time was a specific set of microbes, specific set of bacteria that were linked both to, um, poorer health on many different measures, but also to particular foods. And you could actually see that these, um, microbes were generally linked to sort of foods that we know from other studies are worse for you if they were bad, sort of bad microbes as you put it. And these good microbes, you saw that were linked to better health, were also generally linked to these, uh, foods that were healthier. And so for the first time, we were able to go in the microbiome from just generically saying, you know what, having lots of different bacteria is good for you and not having many is bad, to actually say, you know what, that actually might not be as true as people have said for people who are not really sick. It seems like it's down more to the particular sorts of microbes that you have inside you. And so that was only possible because of this study we've done. Nobody's ever did that before because you had to have this large number of people, you needed to do all of these tests about your blood sugar and your blood fat, and you needed to get this microbiome. Um, and we will be publishing more on that because, you know, we did that study with a thousand people. As I said, we're now at 30,000 people. So as you can imagine, we've now discovered a lot more, uh, you know, and there are some new microbes, some of that information is changing and and so part of Zoe is that this thing is always progressing. I think we're never going to claim, hey, you've got the final answer. It's getting better. And then the third paper, um, was a total surprise. We weren't expecting, we hadn't even thought about being able to study this when we we collected it. And it was looking at the impact of blood sugar on hunger. And so what we discovered, and this was published in Nature Metabolism, was that if you measure what happens to your, so when when you eat something, your blood, if you eat something with sort of any carbohydrates in, your blood sugar goes up. And then your body pushes in insulin and it sort of pushes the the blood sugar back down. And I think many of your listeners will sort of sort of have have heard of that. Um, and so when you wear this sensor, we can monitor that continuously for two weeks. And we gave people lots and lots of standardized breakfast, um, to eat through this period. And what we discovered was that, um, the rise of your blood sugar, uh, doesn't really seem to have much impact on how hungry you are. However, if you start to measure whether or not your blood sugar dips, and not in the first two hours, but actually after that, so sort of between two and three hours, it turns out that people who are seeing heavy sort of crashes, if you like, in their blood sugar, were getting really hungry compared to other people. And those people who were getting hungry were consuming more than 200 calories more than other people, um, over the next 24 hours. And so what you were seeing, and this is really interesting because I think it really just it's a huge attack on this whole calories in calories out idea, is that we were giving people completely standardized meals with exactly the same calories. Some people were having these crashes for certain meals, and when they had those crashes, it was creating hunger, and that hunger was actually leading those people to eat 200 calories more. And 200 calories more is a lot. If you were doing that every day, you could put on, I don't know, 10 kilos or something over the year. And that had nothing to do with the calories at that point. It was the way that it was changing your behaviour in the future. And so I think that was really exciting. Nobody had ever sort of been able to publish anything looking at hunger on a, you know, the previous those papers, those studies have all been in sort of 20 to 50 people because they've had to be in labs. And and this is another area where we we published with a with a few scientists who are sort of world leaders around this topic. But I think what it shows you is that food quality is so much more important than calories. And also your own personal response is hugely important because some people, you know, I think the sort of foods that would tend to trigger this, as you'd imagine, particularly badly would be things like sugary drinks, right? Because they create this spike and then, um, they tend, uh, to create this, uh, crash because they've they've sort of gone. Some people don't have spikes, some people don't have crashes with this. So actually they're able to have that and, you know, it's it's not creating that crash, it wasn't creating that hunger. Other people, they're having that crash, it's making them hungry, and actually, you know, the very fact that they drank this Coca-Cola or whatever is actually going to make them eat more later, right? That's how crazy this is. So basically, we, you know, we've had all this advertising for 50 years, like you need energy, you know, you better go and have something, uh, you know, better go and have something with, you know, easily available, um, sugar in it. Actually, that's going to make you hungrier than potentially not even eating it, right? So it's that's extraordinary. And I think really helps to explain why the Zoe advice basically says, we don't believe in calorie counting at all. The program does not involve any, um, uh, calories at all. There's no constraint in terms of, uh, volume because our our belief is that all the all the scientific evidence that we we see points to understanding the quality of food for you. And if you can eat the right quality of food, actually then basically, um, your your body takes care of sort of how much you you need to eat. And and we see really good impacts on, um, on energy and, uh, I think some promising signs on weight as well.
Dr Rupy: It's incredible, isn't it? That these insights have really been derived from, uh, that little piece of hardware that we now have access to because otherwise, you're right, we would have been using the same advice and we wouldn't have that sort of, um, uh, the progression in, uh, in what we're seeing and what the advice is. And I think intuitively, particularly people who listen to this podcast will understand why the velocity of increase and decrease in sugar levels will have those undesirable effects. But now you can actually see it mapped out. And I think, um, uh, continuous glucose monitors are going to be revolutionary for for a lot of people and and give sort of a lot of validation to how they felt, uh, that the the calorie and calorie out, uh, advice was, um, was not working for them. Um, on the you mentioned the micro, I know we haven't really spoken about the microbiota, uh, up to this point, but, uh, I mean, that that is pretty very central to to to the program and stuff. I I want to ask a couple of things. So what what product do you use and and what do you map as in what kind of sequencing do you use for the product today? And where do you think the blind spots are that would improve the resolution of the the population of microbes that we have in order to direct further insights for them going forward? Because there is sort of like, um, a community of people who feel that microbiota testing isn't at a level where it should be before we can start giving insights, but obviously, that's not an opinion that that you or or or um, uh, Tim hold.
Jonathan Wolf: Um, so, so firstly, the testing we do, we've had to build up the capability ourselves. So, um, we do what is called shotgun sequencing. And, um, what that means is that we're doing the same testing that is now sort of the sort of cutting edge testing that you do in academic studies. So it's the same that we did, uh, in these predict studies. Now, until a few years ago, you know, when we first started actually, Zoe, this was going to cost $1,000 a a sample. So that was obviously pretty prohibitive. The price has been coming down and, um, you know, part of that is, uh, you know, ourselves figuring out how to scale up and and batch this. And so, uh, what that means is that we are able to collect data on all of the individual species of bacteria that are in your gut. And there are historically, these tests have used something called, um, 16S. And what that means is it gives you a, it's a bit like distinguishing between, um, maybe the best example is imagine you went to a zoo with 16S, it'd be like, hey, you've got a bunch of mammals and you've got a bunch of lizards and you've got a bunch of birds and this is how many they are. And that's obviously a lot better than not knowing anything about what's going on. But it just gives you a really small amount of data. So it's like about 10 kilobytes. So if you think about that on your on your computer, right, that's a tiny amount of of information. Um, with a shotgun sequencing, you're getting about two and a half gigabytes of data, okay? So it's vastly more. Um, and that is actually giving you the DNA of all of the individual microbes that are in your gut. And then we do, uh, a huge piece of analysis and we work with, uh, a professor actually in Trento who builds the software that's actually used by all the academics around the world to do this analysis to then figure out, so it's sort of crazy. It's like the it's the world's craziest jigsaw. The basically the way this works is that you smash up all the DNA that was in this stool sample, you put it into this sequencing machine that reads lots of tiny little snippets of it. And so my analogy I sometimes use is like, imagine that you took a photo of the Mona Lisa, okay? Then you put it in a photocopier like a thousand times, then you put it through the shredder, so it comes into tiny little shredded pieces, and then you threw away like 99% of the pieces. So you're just left with 1% now, and then you just put them on the floor and said, could you now like recreate the Mona Lisa? That is basically how shotgun sequencing works. So it is brilliant stuff. It's amazing and it really does work. Um, and part of what's amazing is about half the microbes that you find, no one has actually like discovered them yet. So they don't have a name, they've never been grown outside of the gut. And so there's this sort of dark matter of microbes that we and others who are looking into microbiome are finding. And that's obviously incredibly exciting and I think points to your what you're saying, Rupy, which is our understanding of the microbiome is still very early. There's an enormous amount left to use. So we use, um, this process because we think that allows us to actually identify whether or not you have the specific microbes that we, um, we published in that nature medicine paper. Um, at this point, uh, I think that, uh, firstly, the first thing to to say is that Zoe's advice does not rest upon the microbiome alone. So basically, we are doing, um, these tests where we're measuring your blood sugar, we're doing this blood test, we're getting your microbiome, we're also getting a whole set of health data. We combine all of that together to give advice. So I think if someone was saying like, would do you believe that you could give advice today based upon the microbiome alone? I think, um, the answer is no. I certainly don't think you could give, um, uh, sort of the quality of advice that we're doing. On the flip side, do I believe that the microbiome today can give you real advice that can help to, um, change guidance about what you should eat? Yes. And I think that, um, uh, you know, that nature medicine paper is is one of those breakthroughs and we are now currently running a randomized control trial in the US of the Zoe advice. So I think the real real test, um, will start to come in, uh, probably at the beginning of next year to see whether or not that that delivers results. I guess all I'd say is we're willing to take the risk. We're running a full randomized control trial. Um, you know, you're obliged to, um, as you know, there's a lot of restrictions about how you have to publish this and and what you need to report on. So, uh, you know, I'm I'm hoping that it will turn out to deliver. And the interim, um, clinical study results we've had from previous studies, as well as just lots and lots of members sending back to us, you know, I've had lots of people say to us that Zoe has changed my life. That's amazing. I've definitely never had that in my in my career before. Uh, so I think we are we're pretty bullish about it. Uh, but to come specifically, I guess to this question about the microbiome, like it is still early, okay? The analogy that I've heard, uh, you know, a number of of experts give to me is basically it's like 20 years ago, uh, research scientists suddenly discovered you had the liver. And until then, for like the last 2,000 years, nobody knew there was a liver inside. Now, Rupy, you're a doctor, like the liver is quite important, right?
Dr Rupy: Yes, it is very important.
Jonathan Wolf: And that's the analogy. So basically, nobody and I think the view is like the microbiome should be viewed as like another organ that is probably as important as, you know, as the liver or the pancreas or or whatever. And we never understood, it's not because we didn't know it was there, right? So scientists have known for hundreds of years that there were these trillions of bacteria, they just thought they didn't really matter. They were just sort of there and, um, uh, and, uh, you know, my journey actually started, I had these food intolerances that I I developed after having glandular fever in my early 20s and I ended up seeing a bunch of doctors because I was worried and they were worried that, you know, maybe I had cancer. And so, you know, they they did all of this. And at this point, I remember, they were like, you know, fiber is roughage that's going to help you to like, you know, poo regularly. And there are bacteria, so you should stop eating any of these foods that have any fiber in because actually that might be causing you these problems. I went on the most terrible diet really as as a byproduct of this because and this is what you classically get with people with food intolerances, they're advised, you know, they've historically pushed away from a lot of the foods that are feeding the microbes because often they're having problems to do with how they're, um, they're tolerating this. But at that point, this was 25 years ago, I guess, like, you know, the microbiome was not something that was viewed as having any clinical significance. And today, I think that, um, it's increasingly clear that the microbes and in particular the way that by feeding them, the chemicals they create and those chemicals then, you know, crossing from the gut into the into the bloodstream, really are having an effect on, I think increasingly seems like almost every part of our, uh, of our body. And it's very early. So you have to think like if we've only discovered it 20 years ago, it is of course much less well understood than almost any other system. And it's very complicated because it's not just like there's one bacteria. Um, and so, um, it, uh, is clear that we've got a long way to go in understanding it. And part of, I think, the excitement of of Zoe is the only way you can understand this is to collect because it's vastly personalized, right? So you and I may we'll have a microbiome that is probably 25% the same. Our DNA is 99% the same, right? So it's incredibly different. So to start to understand that, you ultimately need to get hundreds of thousands of, uh, of samples of this microbiome. You'd like to see those over time to understand how they change and there's almost no data on microbiome over over time. And then you need to link it to people's health and their other responses. And so that is a big part of the Zoe like ongoing scientific program is to be able to understand this and therefore to be able to to feed back to to all of our members, okay, as we learn more, this is how you should be adjusting more what you eat because I think we all what we do all know now for sure is that almost everybody in the Western world has a really shrunken microbiome compared to the microbiome of our ancestors or compared to the microbiome of people who are still living a sort of hunter gatherer, sort of pre-modern food, um, pre, uh, you know, running water, sanitation, um, antibiotics. Um, and by the way, you know, running water, sanitation, antibiotics are the best inventions ever and mean that we all live for way longer, but they have had this unintended, um, byproduct on our microbiome. And again, I actually had a fascinating podcast with somebody who analyzes the microbiome from ancient poo samples. So this is now a scientific discipline. So you can get these dried poo samples from humans from thousands of years ago and we can now extract the DNA. So it's like Jurassic Park, we can extract the DNA and actually understand what the microbiome of these people were who were, for example, living in Europe 2,000 years ago or living in Mexico 3,000 years ago. And what they've found, which is really interesting is those microbiomes look quite similar to the microbiomes of people like the Hadza tribe who are living in East Africa and who are still living sort of a traditional hunter gatherer lifestyle. They look quite similar to those and they look radically different from my microbiome and I'm pretty willing to bet, uh, radically different from your microbiome, Rupy, because they have, you know, probably twice as many species. And interestingly, a whole series of specific, we're now this is quite new, a whole series of specific microbes that seem to just be sort of disappearing in our Western world, which you see everywhere in these other, um, societies. And I think we're starting to suspect that some of these are have a lot of health benefits, but unfortunately, you know, the way that they've been passed on, for example, has been wiped out by, um, you know, the sort of sanitary practices that we have now. Um, and also I think because, again, antibiotics are are amazing, but we do use them a lot and we use them a lot, you know, in children and in the first, um, decade of our our life. And, uh, you know, I'm not a doctor, but if you talk to, uh, to Tim and many other people, there, you know, I think there's definitely, you know, a concern about whether we've understood sort of some of these trade-offs, right? About, um, because clearly you save people's lives, that's brilliant, but actually, I think we now understand there can be some negatives. And I guess even more importantly, we use them immensely in agriculture, right? So they're used all the time in our animals. And I think you've really got to question as we understand, you know, the the impact of this, like you can find traces of antibiotics in the water because of the extent in which they're used. And, uh, you know, that's really not saving anybody's life. And so I think I have got, you know, increasingly concerned about that.
Dr Rupy: Yeah, yeah, absolutely. I think, um, uh, there's definitely an increasing recognition from the medical profession that we overuse pharmaceuticals, let alone antibiotics. We've done a whole podcast series actually on the rise of antibiotic resistance. Dame Sally Davies was on, the chief scientific officer, former chief scientific officer talking about antibiotic resistance, as well as a few other people sort of like ringing their sound bells. And I think, you know, the the current pandemic is sort of a warning of how bad things can get. Um, and it's interesting you said, uh, you use that analogy of, uh, you know, suddenly discovering the liver. I use a similar analogy when I describe to deans of medical schools, uh, the the the lack of nutrition education in, uh, in in in medics. Um, you know, I I say like, well, it's kind of like you woke up and you realized, you know what, you forgot respiratory medicine on the curricula. Uh, so, you know, we need to squeeze it in somewhere. It's not like you're just like, well, they'll learn that on the job. You you need to really have nutrition as a really core tenant of of medical education, particularly as it comes, uh, becomes so important with the with the the aspects that we've just been talking about today. Um, I wanted to there's so many things I I really want to ask you about, but I'm just conscious of time, uh, Jonathan. Um, the future for Zoe. Um, so right now, you know, you have a a number of different tests. It's a it's a I believe it's a couple of weeks. Um, you send those insights, you get a really detailed report. My report, I think was like 40 pages or something. I'm still going through it. Um, and and coaching as well, which I think is fantastic. And it's definitely that sort of behavioral element that we need to lean into because as a primary care physician, someone who's seen, you know, thousands of people, I really think the nudges are really important when it comes to really having that impact, uh, rather than just the information. So it's it's good to see that Zoe is doing that as well. Um, where do you see this going in the future? What what do you see the next sort of suite of products? What what is the ideal solution that you see and how will Zoe integrate in healthcare systems worldwide?
Jonathan Wolf: So I mean, I guess the first thing to say is that we need to deliver that first bit well. There will be people listening to this podcast saying, well, you know, that's great, but I signed up on the wait list ages ago, um, and I still haven't got my product. So, so the first thing is just scaling this up and for anybody listening, you know, there there is a really big wait list. We've had 200,000 people sign up for the for the product, which is amazing. Uh, it is taking us time to, um, get through that. And critically, we don't want to drop our ability to support the coaching and everything that we believe is necessary for people to, um, to succeed because, you know, what we're not trying to do is deliver something which is just like, oh, there's a bit of interest. So Rupy, when you say how I got this report, that's great and I want to work through it and don't think like for me that means we've failed because we haven't actually shown you something that you're understanding how to take on board and change in your life. So I'm going to follow up on that with the team because that's that's actually not, you know, that's not success. Success is to understand, you know, like in my case, it was like, wow, my blood sugar control was way worse than I would have expected. And, uh, you know, for those of you on on audio, I'm not, um, you know, I'm not heavily overweight. Um, I think there was no external reason and, you know, a number of the scientists were quite surprised. My blood sugar control is really bad. Um, and as a result, that has some quite strong guidance, um, uh, on on my food. My my microbiome was quite depressing as well. So overall, I have to say it was not it was not a perfect picture of of, uh, of health, but understanding what does that really mean about how to change your diet. I think that's what's exciting. And my diet has sort of been transformed, uh, through this process and I'm always like the first guinea pig whenever we come up with something new or someone comes with new research, I'm like, oh, I need to put that into my diet. I need to make that change. So I I I, um, I really like that. So I'd say the first bit to be honest, Rupy, is just making sure we properly deliver out to everybody. Um, the second bit is making it easier to understand and easier to follow because I think the first version of the product when we first came out is like, hey, we've built this brilliant like machine learning thing that will tell you the score for any meal and all you have to do is just, you know, follow these scores that were sort of 75 plus, which means you can sort of eat them as often as you want, you're great. What we hadn't really realized is well, we were making the user basically explore like the millions of possible foods to try and figure out what to do. So that was a terrible first solution. That was where we were 18 months ago. And now it's much more, you know, there's a recipes, um, but even today, we are still I think quite early in the ability to make it very easy to give you personalized recommendations, really good guidance on how to make changes. So I think, uh, I hope that in the not too distant future, Rupy, you'd say, okay, here are your meals that you did during the test, this is what's good and not. And by the way, here are these recommendations that are spot on for you about exactly what you could do that is going to be better and that you will like and you will want to eat. So that's important. In the longer run, I do believe, you know, coming right back to what you were talking about at the beginning about putting this data back into health care, I do believe that just as we did in COVID, where, you know, if millions of people are working together, you can do this amazing cutting edge science that can save people's lives. That is ultimately what really excites me here. I know that's what excites Tim and and Sarah and others, is this idea that as we're collecting more of this data, we're going to start to understand, uh, whole new things about the way that, uh, what we eat, um, and not just what we eat, but how we eat, um, our other lifestyle factors, whether that's sleep and exercise, uh, our microbiome interact. And, um, you know, I really do think this preventative health thing is real. And just like that analogy we said a long time ago about the car, like it's it seems inevitable that at some point, we really will be able to, you know, if it as it were, you know, take me in to the car, uh, dealership, check check me over, look at where I am and say, actually, you know what, like these are the things that are, you know, really risk factors now, and here is the advice that you can specifically do and these are the things we need to watch out for, um, in a way that we do, you know, for a few things that we can screen for today, right? So you might do, um, you know, breast cancer screening or others in particular populations, but we've historically been very cautious about doing this for the population overall for fears of like a lot of false positives, right? And, um, and making mistakes. But ultimately, you know, we do this for everything else in the world, right? You you know, think about how the plane is constantly serviced and looked after. Like that has to be the future for humans as well. And so I think in the long run, that's I think what really excites the scientists who are working with us is this idea that instead of trying to intervene at the point when, you know, you've got cancer or you've, um, you've got diabetes, like actually it should be a much smaller intervention if you could do it much earlier and that ultimately we all know now, I think that what we want is increased quality of life, right? I think we all see and know people, maybe in our own families who've who've lived for a long time but where quality of life was poor for a long time. And I think we all know that, sure, we'd like to live for as long as possible, but above all, we would like to live with like a good quality of life for as long as possible. And a lot of that is down to what you eat and these other lifestyle factors. And yet the advice continues to be confusing and unclear. And part of that we believe is not not that, you know, these nutrition scientists are stupid, but actually, well, if there's this very, you know, if there's a lot of variation amongst people, then actually you've you've got to help to solve for that as well as making sure you understand it.
Dr Rupy: Yeah, yeah, absolutely. Well, if you're looking for recipes, we've we've got thousands in the Doctor's Kitchen. So I I definitely see a collaboration at some point.
Jonathan Wolf: We will we will follow up, we'll follow up straight after this podcast and see if we can figure something out, Rupy.
Dr Rupy: Absolutely. Um, let's let's end with, uh, you mentioned your blood sugar was was poor, um, your microbiota wasn't looking too happy either. What are what are three key takeaways that you can share with with our listeners that you've put into action, um, based on everything that you've learned over the last few years?
Jonathan Wolf: So three is challenging because there's a lot, but I would say, um, the first thing is that breakfast is one of the places where you can most easily make a big change because breakfast is always a meal you make for yourself. So there's no compromises with the rest of the family. It's something you can eat at home, so unlike lunch, which can be a challenge if you're out, but at breakfast you have control. So that doesn't mean you have to eat breakfast, but I think the plus of it is like you can have a lot of control. So the first thing I think I've learned is so I've I've radically changed my breakfast and I think that in general that's something where, you know, if you can make that change, actually that can be a big impact on your, um, your overall day. The second thing is thinking about eating for your microbiome. So all my whole family, you know, my kids, I said I have a a son and a daughter, you know, my daughter's a little young to understand, my son really knows. He's like, yes, dad, what I eat is all about feeding my microbiome. But you know what? It works because instead of just saying, oh, you need to eat your vegetables, why or anything like this, it's like, okay, I understand. I've got like a trillion helpers living inside my gut. So it's like having a dog. And if you don't feed your dog, right? If you don't feed it right, you know that it's going to be sick. And I think when you have that mindset shift, it's really important. You know, you're laughing, but I actually think it sounds silly, but it's a really big mindset shift to say, I'm not just doing this to look after myself because like that's boring. It's like I've actually got it's like I'm in the zoo, you know, and you can't feed the lion plants, right? And you can't, uh, you know, feed something that lives on plants red meat. It's the same thing. So you've got to actually give it, um, the food it wants. And at a very simple level, that's fiber, that's, um, and that's something that our diet is tends to be stripped of. You know, I now start to think again, I just I ate the worst diet for like almost 40 years and I didn't even realize, you know, I think for the last 15 of it, I thought it was really good, but there was basically no, um, no fiber in it. Um, and, um, I guess the the third thing is diversity. This is something that Tim talks a lot about and I think this is what we really understand, which is that a healthy diet where you're actually basically eating the same thing every day isn't a healthy diet, even if the individual food seem like they're good. And I think what's clear again as you start to look at it through the microbiome is you need to feed them all of these different plants because there's so many different microbes that are all optimized for eating different food. And that diversity helps to then create these vast numbers of different sort of chemicals. And that, you know, we were all brought up with this idea that, you know, there's a certain number of vitamins, right? Uh, that you have to eat. And I remember, you know, there was there was times when you would have cornflakes that said, hey, it's got like these five vitamins, despite being frankly one of the worst foods you could possibly think, but you're like, oh, it's good. I feel like I'm I'm giving this to my children and I'm a good dad because it's, you know, it's got vitamin B5 or whatever it ever it is. Actually, I think there's been this big shift and again, a number of the scientists we we talked about where they now understand there are more than 20,000 different chemicals in the foods we eat, right? So that blows your mind. Like you're worrying about five or six vitamins. Clearly, that's just the wrong way to think about this. There's just this vast complexity in the food. And so what you need to try and do is have diversity. What's probably most important about this is is plants. So we often talk about sort of 30 plants a week. Um, how do you try and get these different plants? And the good news, plants is is also nuts and it's seeds and so there's various ways that you can sort of sort of cheat. But I I think about that as a way to try and create more variety, um, across across what I'm eating. So those would I guess be my top three.
Dr Rupy: Those are those are epic. And I was laughing earlier because it reminded me of a book that was written by a colleague of mine, uh, who you've probably come across, Professor Felice Jacka. She was, um, one of the lead researchers for the Smiles trial. And, um, she's written a children's book called There's a There's a Zoo in My Poo. So it just kind of reminded me of that. And I think it's a really important concept to get across to kids because it gives them the the reasoning as to why they need to eat their vegetables rather than just because it's good for you. And I think that fascination with the bugs that live inside us is something that we should, uh, cultivate, excuse the pun, um, from an early age. But Jonathan, thank you so much. This is brilliant. You are, I can tell you're a a budding, uh, podcast host because you talk so so well eloquently and, uh, this is this is super easy for me.
Jonathan Wolf: Thank you, Rupy. You're you're very kind. I've enjoyed it very much and, uh, we will definitely follow up on those recipes.
Dr Rupy: Okay. I'll catch you, mate. Cheers.
