The Quantified Body

(00:05:43) [Damien Blenkinsopp]: Welcome to both Eran and Lihi Segal onto the call. Thank you both very much for joining us.

So I just wanted to jump straight into your research on the glucose response, and all of the other stuff you’ve been doing in the last couple of years really because it’s all kind of related. Why did you focus on the blood glucose topic in particular?

[Eran Segal]: That’s a really good question. When we started a few years ago, we wanted to take a science-based approach to nutrition.

We thought very hard about that problem, and what we should examine. And if you think of the most common approaches in most studies in nutrition they usually consist of some dietary intervention, and then they look at weight loss, or they look at a change in some marker of a disease. And that’s great because ultimately these are the parameters that we’d like to have an effect on.

But, the challenge we found with this approach is that it then takes weeks or months for these parameters to change. You know, a parameter that measures your diabetes level, or weight. And at the end of this, you get a single measure. It takes weeks or months to change, and that measure is affected by multiple things that happen to you during those weeks or months. Both the diet intervention that you give, but also many other factors unrelated, which can be then confounding to what you’re measuring.

So, we thought that maybe one of the reasons that it’s very hard to do nutritional research, and why many researchers are failing, is because they’re looking at this single measure effected by many things. So we didn’t want to go that way. Even if we see an effect, you’re not sure you can attribute it to the diet, and if you don’t see an effect it’s very hard to troubleshoot what went wrong.

So we thought very hard about this, and that led us to look at glucose levels. More specifically, the glucose levels after a meal, what’s called the postprandial glucose response, or post-meal glucose response.

So by that, what I mean is what your blood glucose levels look like in the two hours after you eat a meal, which we can also quantify using the area under the glucose curve into a single measure representing the response that you had to that meal.

[Damien Blenkinsopp]: Right, so that’s like the total area under the curve is the total amount of glucose that was in your bloodstream during that area of time.

[Eran Segal]: Yeah, you can think of that as an approximation. I’ll tell you in a moment what we really are hoping that this is actually measuring, but that’s quantifiable into a single measure. But now we have to think about three aspects, or three features of this that really led us to conclude that this is what we want to follow.

So in a nutshell, what they are is that we were convinced by all the existing literature that this post-meal glucose response is really key to weight management. It’s really key to disease – diabetes, but not only diabetes, I’ll talk about those.

Finally, and not least importantly, that it’s very easy to measure and it’s something that, not within weeks or months but within a week, we can obtain not one, but even 50 quantitative measures of healthy nutrition in a single individual.

So first, why is it important for weight loss and weight management? This is very basic, and there’s been a lot of literature on this.

When we eat – and I’m talking about healthy people, even people who are glucose intolerant, but let’s say not insulin dependent Type I diabetics. When we eat, our body digests carbohydrates in the meal and releases them into the bloodstream.

After that, there is a response of the body by secretion of insulin, whose job is to lower the glucose levels. But in addition, what insulin signals, also, is it signals the cells to uptake the sugar that’s floating around in our blood.

And then excess sugar is converted into fat for storage because it initially is converted into storage of glycogen, but our stores of glycogen are highly limited. So very fast the remainder will be stored as fat. And this is actually known as one of the primary mechanisms by which we gain weight. In other words, this action of insulin.

So, in a sense, we would have liked to even measure directly at a continuous rate insulin, but that’s technically not possible. But in healthy people – and there’s been lots of research – by measuring glucose levels you’re actually looking at a proxy for a measurement of insulin.

And there’s been work showing, for example, that if you eat a meal that spikes your glucose levels compared to a meal that does not, then after a meal that does not you have more oxidation of fat, more burning of fat.

So the connection to weight loss is very well established. There’s also a lot of literature looking at very low-carb diets, which I think as a dietary regiment it’s incompatible with life for most people. But if you look at some of the studies when you eat a low-carb diet your glucose levels are low, and in general, those have the most beneficial effect on weight loss.

So that’s item number one why we focused on blood glucose levels because it’s very important for weight loss and management. The second is disease, and the most obvious is, of course, diabetes.

In fact, diabetes is diagnosed and defined by glucose levels. It’s defined in two or three different measures; either by the hemoglobin A1C, which measures your average glucose over a period of three months or by the glucose levels that you have two hours after you eat a meal. So something very similar to what we’re measuring.

And so, of course, you directly are playing with and improving the measures by which you diagnose diabetes. By that we can manage better the disease; manage it better in pre-diabetics, even possibly reverse it in this population. And, of course, for all the people with normal glycemic levels, we can prevent or delay the onset of diabetes.

So that’s one area where it’s important, but then separate from diabetes there’s been a lot of links to cardiovascular disease, to cancer. So in cancer, this is known as the Warburg effect. We know this for 90 years that cancer cells have a very different metabolism that much more heavily relies on glucose.

And so the thought is that by limiting the amount of glucose that you provide, you deferentially affect the growth of cancer cells compared to normal cells. And there’s been associations in the literature between blood glucose levels and cancer.

There are also been associations to overall mortality. There’s one paper that tracked over 2000 people for 30 years showing that if you responded more highly to a glucose challenge 30 years ago, you’ll live longer, basically. So there’s been links to many diseases, and so we’re very confident that it also has a strong association to disease.

And the final point is what I made before that because of the technologies with continuous glucose monitors we can now really in a single week measure 50 quantitative measures of healthy nutrition. And they’re quantitative of health nutrition because of the two points I made before.

[Damien Blenkinsopp]: So you felt that it was basically the continuous glucose monitor was a game changer because you’d be able to gather a lot more data quickly, and eliminate somebody’s potential variables coming in from the longer term studies which you can avoid.

[Eran Segal]: Absolutely. So if you think about it, we actually focused on examining the direct effect, one of the ways by which food directly affects you, and this is your glucose levels.

And from everything I mentioned before, we also believed that this is really a very critical clinical marker for weight loss and disease.

(00:13:30) [Damien Blenkinsopp]: Right. Okay, great. So you focused on the post glucose response to meals specifically, but you did mention Hemoglobin A1C. Is that something else you tracked and you found useful in these studies?

[Eran Segal]: So that’s something that we measured. We found it useful for predicting how different people respond to different foods, but it’s nothing something that you measure as a direct effect of a meal.

It’s one of those single parameters that takes many weeks to change that, again, would be very hard to develop a dietary regiment that would affect that directly because of all the confounders that I mentioned before.

So in fact, glucose levels is, as far as we know, the only reliable quantitative measure that is really super relevant that we could track, and that’s why we focused on it.

(00:14:17) [Damien Blenkinsopp]: Right. And you mentioned the area under the curve is the part that you’re interested in. So I’m guessing that you’re looking at a benchmark to what’s okay, and what goes too high in terms of that area.

You said to me when I tried to give an analogy to explain that to the audience that it wasn’t quite right. How would you explain the utility of that?

[Eran Segal]: We can just say that it’s basically looking at your glucose response and it’s quantifying how much you had spikes for glucose levels after the meal. And these spikes, as I mentioned before, is what is strongly linked to everything else.

(00:14:53) [Damien Blenkinsopp]: Right. Thank you very much. How did you find the continuous monitoring technology? Did you use a specific device, and how sensitive or accurate did you find it? There’s various monitors out.

We’ve spoken about these before, and I know people that have been using them for diabetes management and so on. So I’m just interested in your opinion on where that technology is right now, if research can be improved maybe later as it advances, or is it already as good as it’s going to get.

[Eran Segal]: So I think it was very good for our purposes. Not without problems, but I think even finger pricking is problematic, and can be variable. But, there’s also progress.

There’s a recent device by Abbot that we are now shifting to using because it’s more convenient, mainly. It’s probably as accurate, possibly even with higher accuracy – that’s what the company claims. But it’s just much more convenient, and it doesn’t require the finger pricking anymore.

But overall, they definitely capture the trends. I will say that when we measure responses to 50,000 meals you really have a very large data set, and you can afford to have some inaccuracies here and there, which all the technologies have. And still you correct for that in the algorithms.

(00:16:10) [Damien Blenkinsopp]: Great, thank you for that. Moving on a bit to what you discovered is actually driving these blood sugar regulation changes. What are the examples of the most unexpected things that you saw in the data?

[Eran Segal]: Are you talking about the factors that affect it, or even just before the surprising responses that people had?

[Damien Blenkinsopp]: I’m interested in both. If we start with what you saw that maybe you weren’t expecting, and then what you think drove that, or what you discovered drove that.

[Eran Segal]: So the first key result of the study was – and this was initially very surprising – we saw that when you give different people the exact same meal, they have very, very different responses. And this is in contrast if you eat the same meal on two different days, which is what we’ve tested on an unprecedented scale of 1000 people. This is 7000 different meals standardized that we provided.

When you eat the same meal on two different days your response is going to be very similar, but when you and I will eat the same food, our responses will be dramatically different. You can eat bread and have zero response, and I can eat bread and have a higher response than what I would have if I even ate pure sugar. So it was really all over the place.

And this was even before talking about our solution, this was very surprising. And we realized also that it has a lot of implications.

Because if we realize, again, the importance of blood glucose levels to our health and weight, then what it directly means is that general dietary recommendations are always, no matter what they are, going to have limited utility. Because for any single food that we tested, we had people who had a high response and others that had a low response.

So you can’t really make a general recommendation about food. Now there are trends. There are foods that lower glucose levels on average, for some people. And that is known; it’s what’s called the glycemic index.

I think you even touched upon that in your questions. And we also saw that in the data. So whatever foods have been reported with lower glycemic index on average they have lower responses also in our data. But if you look at all those numbers that go into making that average, they’re all over the place.

[Damien Blenkinsopp]: So there isn’t a cluster around the mean, it’s widespread.

[Eran Segal]: Exactly. It’s very spread across it. And when you measure enough people the means will be significantly different, but there is a wide spread across the means.

Meaning that we can take ice cream, for example, which on average induced relatively low glucose levels, and we can take rice, which on average, induced high glucose levels, but you will still find people that respond more highly to ice cream than to rice.

(00:18:49) [Damien Blenkinsopp]: So it’s quite surprising in those terms. So, in terms of what you’ve found or discovered that drove that. I know you tested for a lot of different things. What sort of things did you also test for in order to try and find the pattern of what was driving this?

[Eran Segal]: So we looked at many different things. We looked at body measures, anthropometries, height, weight, waist for instance and so on. We looked at several metabolic parameters in blood. We looked at questionnaires.

So we had a medical background in food frequency and lifestyle questionnaires. And the most novel component that we integrated into the study is the microbiome. So we measured all of those. In fact I will say that we found an association, a strong correlation, between variability and the response to food in all of these different groups of parameters that we measure.

And then the next step was to take all of these parameters and integrate them into rules, or an algorithm, that basically given your inputs to all of these factors, which vary from person to person, we would be able to predict how you would respond to each and every single food or food combination or complex meals.

And we showed that that actually works very well, and predicts personalized responses with very high accuracy. In fact, the accuracy that we think is even feasible because, even when you eat the same meal on different days, I mentioned your response is going to be very similar, but it’s not going to be identical.

So there is some inherent biological variability, and our predictive power is approaching that variability.

(00:20:30) [Damien Blenkinsopp]: Okay, great. The microbiome was the novel part of this. What exactly did you look at? Because there’s a few different approaches to looking at the microbiome right now.

What were you looking at and trying to map with it?

[Eran Segal]: So we looked at the most comprehensive in terms of resolution, which is just doing shotgun sequencing. So that’s basically sequencing the entire content of what we find in a stool sample. That mostly consists of bacteria, but this type of sequencing is really the highest resolution.

It allows us to identify individual genes in the bacterial composition, of which there are several millions in each and every one of us. It allows us to identify not just species, but also specific strains of bacteria.

And so there are many of these different factors that we integrated together, and used them in the algorithm.

[Damien Blenkinsopp]: Great. Is that cost prohibitive verses some of the other technologies that are used out there?

So you have the 16S, which is just looking at one part which some of the projects like uBiome are using right now to enable them to serve many consumers and make it a lower cost so people can afford it right now. Are the costs much higher for what you were doing?

[Eran Segal]: So first of all, for 16S, I will say that we didn’t want to go in that direction because science-wise I don’t think we would have gotten as predictive power.

And in fact we even showed that to ourselves in the study because it doesn’t have the resolution, and in many cases it doesn’t allow you to go below even the genus level of bacteria. So you can have the pathogenic E. coli or non-pathogenic E. coli will have identical 16S; you won’t know what’s in there. Just to give an example.

So we went for the shotgun sequencing. It is indeed much more expensive. If you talk to researchers they’ll tell you that it’s way more expensive.

I will say that what we have been working on in our labs for many years prior to this study, and then as part of the study, is to optimize this process very extensively using automation and using robotics.

We’ve substantially reduced the cost; it is still significantly more expensive than 16S. But I think our margins of error are much smaller than other researchers, and this is probably also why we were able to profile at that level.

(00:22:53) [Damien Blenkinsopp]: Okay, great. So, in terms of the microbiome – because we’re talking a lot about the microbiome and the other factors – is there a stronger weighting of the variability? Are there variants associated more with the microbiome, or are there some other factors that are really important?

The other thing that is interesting is the microbiome actually does change, and we’re trying to change it and improve it and so on in many clinical situations now. Whereas your height, age aren’t changeable.

So if you could give me a bit of background on what you found is the biggest weighting there, and maybe which is most actionable?

[Eran Segal]: Those are two very good questions.

Related to what is most important, every component that I mentioned before we can show has significant predictive power. Now of course, in terms of predictive power, some of these components are somewhat redundant with each other.

So for example we found that when you add the microbiome and some other components, then we can do without all of the blood tests, and in fact we don’t need them at all for the predictive power. They add really something negligible.

Of course we think that blood parameters are predictive; it’s just that in the context of many other parameters, they’re somewhat redundant because they can be explained and correlated with several other parameters. And so likewise with the microbiome we found that actually unlike blood, in every context that we apply the algorithm, the microbiome always had a significant contribution to the prediction.

I will say though, that of course the microbiome has the most significant contribution when you add it by itself. As soon as you add more and more parameters, this is expected. It’s marginal contribution. And also, I believe this is an area where with additional research we can dramatically improve in the future.

We already have started this process because we have a lot more information and a lot of smarter ways by which we can handle this data, which is not true for BMI, weight, blood parameters, which are very limited in the amount of information they have.

[Damien Blenkinsopp]: Right, because there is basically truckloads of data we’re going to be taking out of our microbiomes, because there’s so much in there.

[Eran Segal]: And when we and others continue to research and identify key genes in the microbiomes that are helping in the breakdown of certain products, production of different metabolites that affect us, and we know better how to zoom in on different features, we’ll be able to improve the predictive power from it.

(00:25:25) [Damien Blenkinsopp]: Great. So in terms of the level, you mentioned that the technology that you’re using goes right down to the strain level, and the species, and genus, and so on. But where do you see the patterns?

Is it on the genus level, the species level? Is it just one species that can completely change how we respond? Or is it at a very high level like bacteroides, or something like that?

[Eran Segal]: So there are significant associations on all levels.

And I can say that it’s not a single species that is really dominating. We actually have this in our paper; we have many different features from the microbiome each make a contribution to the overall prediction, but together there’s dozens of these features. Together they make a significant contribution.

[Damien Blenkinsopp]: Right. It’s really a multifactorial analysis.

[Eran Segal]: Yeah.

(00:26:10) [Damien Blenkinsopp]: Okay. You did a paper before 2014 on the artificial sweeteners, which also got a lot of coverage. That was interesting also.

And in that one I believe it was the high bacteroides and the lower clostridiales which showed that you had a higher propensity to gain weight, wasn’t it? Rather than just blood glucose regulation.

[Eran Segal]: Yeah. So yes, we did see an overall effect there. But also there we developed an algorithm that could predict susceptibility, in that case, to consumption of artificial sweeteners. And that was also multifactorial basically using dimensionality reduction of essentially all the species that we had in the sample.

(00:26:53) [Damien Blenkinsopp]: So the most recent paper you are looking at is also looking at regaining weight after dieting.

For example, if you go on a diet and there’s this typical yo-yo effect where someone goes on a diet and they just regain it all back. I’m wondering is that related to the microbiome or what’s going on? So if you could relate what you’ve been looking at there and what you found?

[Eran Segal]: Yeah.

So we study many different aspects of the microbiome as it relates to our health. And this is another study where we studied another very basic phenomena, the yo-yo diet that you mentioned. And what we showed there is actually that even after you complete a diet and you lose weight, your microbiome doesn’t go back to what it was.

So it’s very well known that as you gain weight your microbiome changes, and what we showed is after you lose weight your microbiome doesn’t revert back to the original state. And that memory, if you will, of the microbiome is in fact sufficient to induce and enhance weight gain once you stop the diet.

So I would say it’s another work further establishing the causal link, and providing more insights into mechanisms by which the microbiome plays a key role in our health, and specifically with respect to metabolic states and diseases; in this case relapsing obesity.

[Damien Blenkinsopp]: In that study did you find any mechanisms? Is it specific species? I think you were talking about metabolites in there as well.

[Eran Segal]: Yes. So this work was in fact work in animal models; this was work in mice. And the advantage of is that we can really go deeper into mechanisms, unlike in humans where it’s much harder.

And so there, we also did a metabolomic profiling, and we identified metabolites that were missing after you lose the weight. And when we administered these molecules back, we in fact were able to cure the mice of the phenomena of relapsing obesity.

[Damien Blenkinsopp]: Wow.

[Eran Segal]: And more important we actually showed that these metabolites in fact regulate genes in the host, in the mouse, and they regulate genes that affect energy expenditure. So these mice, when they have less of these metabolites which are broken down by bacteria, when the bacteria break them down, these mice are going to have less energy expenditure and therefore more weight gain.

[Damien Blenkinsopp]: Wow. So I guess you don’t understand why that energy expenditure is going on. There’s probably quite a complex downstream process that follows.

[Eran Segal]: Right. That’s quite complex, but we also had some insights in the paper as to that as well, and we found some genes that regulate that process in brown fat tissue that are directly affected by these molecules. And these molecules are made less available because the bacteria in mice that had a previous history of obesity, in fact, were breaking down and taking away these molecules more.

[Damien Blenkinsopp]: Wow, so it’s actually the introduction of new bacteria for the weight gainers, which is taking away these substrates.

[Eran Segal]: So in this case, it was metabolites. So there are specific metabolites that are broken down by bacteria, which we showed here, we call that post-biotics as opposed to pre-biotics.

[Damien Blenkinsopp]: Right, by adding the bacteria that’s missing or making taking away the ones that are causing the problem.

(00:30:17) [Eran Segal]: Yeah. Those can be technically more challenging in some cases, but in general yes.

I also want to relate to, you asked me before about the stability, or how much the microbiome changes. And we have several studies on that; in fact, some are not even published. What we find is in fact the microbiome is actually much more stable, perhaps, than most people think.

So in fact your microbiome, unless there is very dramatic change in health or weight, is probably going to be very stable even across many years. We have data on that. And what I mean by stable, it means you will still look more similar to yourself even after following some dietary interventions, at least in the short term, than you will to other people.

Now, having said that, we also found that short term dietary interventions in fact do change the microbiome, also in consistent ways, across different people. So while you’ll still remain in the neighborhood of what your microbiome is, still some functions will go up, some will go down. Those can be consistent across multiple people who consume the same type of dietary intervention.

[Damien Blenkinsopp]: Right.

Just as a takeaway from that, do you think the microbiome is going to be an important area of work? Basically learning how to modify it, push it in another direction in order to solve things like weight gain, blood glucose regulation. Is that your hope?

[Eran Segal]: Absolutely.

So the more we find causal effects for the microbiome on our health and weight the more this should be a target for intervention. But of course that will require further studies to understand what is casual and also how to change it.

And I do believe that with – and this has also been shown – that with long-term changes in diet, you will in fact achieve changes in the microbiome. But with short term dietary intervention the changes will be consistent, but they will be more subtle and you’ll still remain in your own neighborhood.

And what that means in terms of the research that we did, it means the algorithm is going to give you essentially the same predictions, even in a very stable fashion, across even some small, short term dietary interventions because your microbiome is essentially going to be very much the same.

[Damien Blenkinsopp]: Right. So if I test one month, and then I test six months later after doing a series of interventions – maybe not too intense, something like courses of antibiotics, things like that might be more intense.

[Eran Segal]: Antibiotics is probably a different story. That can have a dramatic effect.

I’m talking about even if you change your diet for a few months, your microbiome is not going to change a lot. If you maintain a very different diet after a prolonged period of time – I can’t give you exact numbers, but a long time – then you will see change.

And at some point, those changes may be large enough you may want to test yourself to make some modifications to the diet. But, for a very long period of time, without dramatic interventions it should stay pretty much the same.

[Damien Blenkinsopp]: It might be interesting if you do a course of antibiotics, because people have to from time to time, to redo the test and see what it predicts afterwords. Maybe some of the food responses are going to be different.

[Eran Segal]: Absolutely. And I think after antibiotics you will have very significant changes, and those could affect the prediction.

(00:33:37) [Damien Blenkinsopp]: Yeah. So the last thing, just going back to the artificial sweeteners we spoke about. Because they did see that those had an impact on the microbiome over time.

Do you think smaller things like that, basically micronutrients or small fibers, not necessarily macronutrient profiles, but those kind of things could have longer term impacts on the diet?

[Eran Segal]: Absolutely. I would say some of them could even have bigger effects than macronutrients. So fiber, for example, is something that is digested solely by our gut bacteria, so definitely could, and this is known, have alternations and will overtime have sustained effects. So yeah, absolutely.

I think the way we think about it now, and even drugs. We and others have shown that the drugs that you take actually also affect your microbiome. Any substance that you intake, although depending on the substance, might just go through your gastrointestinal track, meet the trillions of bacteria that are there.

They have 100 times more genes than we do. They could definitely break down these products, they could convert it into other products. I would think of it right now, anything that you intake could definitely affect your microbiome.

(00:34:50) [Damien Blenkinsopp]: Yeah. Alright. Thank you very much for that. Just a last few things.

A lot of people take xylitol and stevia. It wasn’t in your original study, and I was just wondering if you knew anything about that. Because the other ones, aspartame, saccharine, and there was another.

[Eran Segal]: Sucralose.

[Damien Blenkinsopp]: Sucralose. Yeah. It was a bit of a negative view on them in terms of what they were doing to the microbiome. Have you got any information or did you see anything on the other two?

[Eran Segal]: We are studying those now.

[Damien Blenkinsopp]: Great.

Eran thank you so much for your time. It was really useful.

[Eran Segal]: Okay, great.

(00:35:19) [Damien Blenkinsopp]: Excellent. Okay, Lihi, let’s talk about DayTwo and what you’re doing there.

So basically you’re taking the work done by Eran and his co-researchers and you’ve been turning that into this algorithm service to help optimize people’s diets. Could you give me a bit of an overview, how you look at it? What the company’s doing and how you see it going forward over the next year or so?

[Lihi Segal]: Yeah, so we licensed the technology in an exclusive way about a year ago, in the summer of 2015.

And then what we’ve been doing since then with the help of both scientists, because our founders are scientists and they’re on the management team and very deeply involved in the company. And so there’s a lot of hand-holding in that sense on the scientific level as well.

But what we’ve been doing, we built a team up of machine learning experts in DayTwo and also developers, and we really dove into the algorithm.

As you heard, on the research level the first thing they took 30 metrics in the blood, they did the microbiome, both 16S and the full shotgun. What we really tried to do is once we have all the results is really look into the algorithm and see what is that minimum set of features that we need, and write it to consumer. We don’t want to send them to get anything that is redundant.

So looking into that features into the algorithm, and looking to see what we really need, how to commercialize this. So we went through a kind of learning period when we’re looking to see how we define the product, what do we need. Do we need to freeze your stool? Do we need to send you to a doctor to get blood tests, yes or no?

And where we ended up is by looking at a really minimum set; because as you heard Professor Segal say, the microbiome was very significant in any constellation that they took, and made other things redundant. So really where we ended up with on the product side is that it’s all online, almost.

So you come online and you fill in a lot of questions – not a lot, I think a 10 minute questionnaire. But, of course it has to do with your anthropometrics and your food preferences and your medical history. Any information you just fill in your questionnaire. And then we mail home a kit; just a box. In that box there is a small tube and you take a stool sample at home.

So we use DNA Genotek as our supplier of the kit. If you know them, they’re out of Canada. This is really kind of state of the art microbiome collection kit. You don’t have to freeze it, you literally just take it when you can, when it fits you. You don’t have to time it. It’s there, you take it, and then you just mail it back to us by regular mail.

[Damien Blenkinsopp]: Is it a quick swab, or are you actually taking a sample?

[Lihi Segal]: We tested a bunch of other alternatives as well, but this company really gave us the most stabilized microbiome in extreme temperatures.

It’s really important for us to stabilize it and then send it through the mail. And you don’t have to freeze it and all that. So it made it much easier on the consumer side, and it’s also very important scientifically to get the microbiome at the state it was as it was collected in Day Zero.

So we did a lot of trial specifically on that to see that what the company claims is actually right. And so we send you this kit, you mail it back to us, and then we sequence it.

We chose to sequence, as Eran said, on a full shotgun basis because we found that that resolution rate gets us the prediction into a higher level and a very good level. So we decided to do that despite the higher costs that it has.

But again, we try to put a product on the market that is very good; it’s good scientifically, we don’t really cut the corners there. So although the cost is still higher, we do expect it to go down a scale, both on the full shotgun basis and the kits.

And then what we do is give you a mobile application. So you get a personalized mobile app that you download, and it’s tailored just for you. And it gives you three things initially.

It gives you a microbiome report, because we did it and we have it. Not all our users are going to love it, but a lot of them may be curious to open it up and see. And so there’s a lot of information there.

We’re giving you your top food and meal recommendations. So what that means is that we really look into different categories. You have your top breakfast, your top lunch, your top dinner, your top fast food, because even when you eat fast food once in a while you can still choose healthier fast food than others.

We’re really trying to bring this into your day-to-day and make little changes and not turn your world upside down. And then there’s whatever alternatives with pasta, alternatives with rice. That’s really general.

And we’re really giving you your top A+ meals and scores all the way to your worst list, which has up to C-. So we’re trying to educate you through that stage. You could always go to see what your top breakfast is, what your top lunch, and all that, but then you also have the ability to search.

If we didn’t say something that you eat and you want to know what your score is, you just search for it in our database. In the US we are based on a database of MyNetDiary. So we have 400,000 different foods that are US based foods.

In Israel we are have a different database that has Israeli foods in it. So people can really find what they eat in there.

[Damien Blenkinsopp]: Right, so these are actually branded products you can buy. Is that what you’re saying?

[Lihi Segal]: Yeah, there are a lot of branded there as well, but there’s also, for example, an apple without skin.

[Damien Blenkinsopp]: Okay.

[Lihi Segal]: You also get your just general food as well, but you would find your specific brand of whatever, yogurt, that you’re eating in the specific territory. And then, so that’s the second thing. The third thing is the search and also a build your own meal kind of possibility.

So the whole point here is that we’re not scoring nutrients. We’re not saying carbs or proteins, and we’re not even going into a family of pasta versus rice. It’s very different if you eat a pasta with cream sauce or a pasta with meatballs, or you eat a pasta with macaroni and cheese.

You have to be able to score complex meals, and that is where our kind of secret sauce is, we’re really looking at your personalized response to these complex meals. And so you can just search for those meals if you want. If you’re cooking or if you’re sitting in a restaurant and you’re able to get your scores on the foods that you’re eating.

(00:42:00)[Damien Blenkinsopp]: Yeah. So just to clarify, this is just focusing on glucose management? So lowering…

[Lihi Segal]: Right. So what we aim to do is balance your blood sugar levels. So when you go on and you eat your A+ or A- foods and you eat that on a consistent basis, and you keep portion control.

So it’s not a kind of blank check to eat as much ice cream or drink as much beer as you want, unfortunately. But it does allow you some flexibility with foods that are surprising. Things you thought were unhealthy, all the sudden you understand you can eat them. And vice versa, so it’s surprising in both ways.

And then if you eat that consistently then yes, you’re going to see that we’re helping you balance your blood sugar levels.

And as Eran mentioned, balancing your blood sugar levels has an importance both in minimizing the risk for diseases of all kinds. Even as a healthy person, you don’t have diabetes but it is really important to keep your stable blood sugar levels. And also the whole thing about weight loss.

It helps you, it encourages weight loss in that sense. So you need to have a restrictive diet; you can’t eat whatever you want and think that you’re going to lose weight with this. But it does help you lose weight, it helps control your hunger, it helps control your cravings. And so it really helps you to plan and choose your foods right. That’s what we’re aiming to do.

(00:43:25) [Damien Blenkinsopp]: Okay, great. So, just to be clear. In terms of the inputs, it’s mostly filling in a questionnaire. Is there any other test apart from the microbiome sample? Or is that just the only one that they need to do?

[Lihi Segal]: No, the basic thing is that we need the microbiome and we need your questionnaire.

Now if you do have addition information, if you have your HBA1C levels then we’ll be happy to take them in. If you have more blood tests it’s always good to take in. But it’s not as significant enough so we’ll say you have to do it.

[Damien Blenkinsopp]: Yeah.

[Lihi Segal]: But on a general level, as much information as you’re willing to give us, it will always help, yes.

[Damien Blenkinsopp]: So in your algorithm, it will just take that into account as well?

[Lihi Segal]: Yes.

[Damien Blenkinsopp]: It’s just that in terms of the cost, you don’t want to add to the cost or be inconvenient.

[Lihi Segal]: Again, as Eran mentioned, it becomes redundant at some point.

And so if you have it, great, but we don’t want to get people – the cost is not that much for an HBA1C, it costs like 20 dollars in the US today. So that’s not really the issue.

It’s more just this is the basic package; you send it home, you send it back. But as we’re looking at our future products and as we interact with you throughout your day, the app is going to allow you in future versions to report to us what you ate.

And we have a lot of insight on your sleep and on your exercise. That was not published, but we have it in the data, and they haven’t published that data. He didn’t mention it, but in the research they actually had people logging in their foods, but also their sleep and also their meditations and their exercise. They had a Fitbit on everyone.

So there’s a lot of insight that we’re going to be able to give you. And when to eat your biggest meal, because people have a certain rhythm and that’s personalized as well. So when would be preferred to have a large meal of the day. In the US usually it’s dinner. In Israel sometimes it’s lunch, sometimes it’s dinner.

Certain foods that you should eat at certain times of day. So we can really interact with you over time if we have more information on how you slept last night and how much fiber you had in the past 24 hours. There’s a lot of things that go into the algorithm that, if we don’t have them, fine, but if we do it can even help us give you better results.

[Damien Blenkinsopp]: So you’re integrating these lifestyle factors as well into the computations to tell people when to eat. That’s great.

[Lihi Segal]: Your stress levels, all that.

(00:45:52) [Damien Blenkinsopp]: So I was wondering, are you able to tell the status of someone?

Say I’m glucose intolerant to an extent already, when you get the data from people without getting the HBA1C, for example, are you going to be able to know this person’s going to have to be more careful? Is any of that kind of information coming out?

[Lihi Segal]: We’re not at any point a diagnostic company, so whatever we see we will not tell you.

[Damien Blenkinsopp]: Oh, okay.

[Lihi Segal]: We don’t do health assessments on you. We’re giving you your recommendations under a predictive model.

And for example if we find things that we think you should know, then we would probably say maybe you should see your doctor, or take these results to your doctor or something like that. We would never go into actually giving you any medical advice.

(00:46:35) [Damien Blenkinsopp]: Right. The same usual thing. There’s a lot of blood glucose dis-regulation that goes on way before you get to diabetes, as Eran was saying.

So I’m just sort of interested from an algorithm perspective. I know you’re not going to publish it because there’s a medical borderline there that you don’t want to go near, but I was just interested from an algorithm perspective – can it tell how far you are along that line? Because everyone’s got a little intolerance. I’m just curious, does it offer any information?

[Lihi Segal]: I can’t.

[Damien Blenkinsopp]: Okay, fine.

[Lihi Segal]: I can’t answer that question.

But as Eran mentioned, we’re looking into on the road map for DayTwo that’s not just for the people who want to buy it right now but we are looking into various things we can do with the data that we have, the data we collect, and the things that we learn. And of course diagnostics and therapeutics are a part of that whole agenda.

And so there’s insight that we’re looking into and collecting, and can very well come out with additional products that are related.

[Damien Blenkinsopp]: So as a first stage it’s basically a food recommendation engine as the output, and of course your microbiome data.

Do you have an idea of what type of microbiome data is going to be given? I know we talked about uBiome, for instance, in the past. We had Rob Knight from some of the other tests.

We’ve looked at a few different ones in the past. Have you got an idea yet, or are there pictures or anything of what it’s going to look like in terms of the data you provide for the microbiome?

[Lihi Segal]: I can definitely go back and send you some information about how it’s going to look, more or less.

[Damien Blenkinsopp]: Alright, cool.

[Lihi Segal]: But we’re trying to go into a lot of detail. Again, we’re doing full shotgun so we have additional insight. We’re not at just a very high level; we are looking into specific types of bacteria and trying to link them. We’re looking at studies and just general information about them.

Again, we have to be a little bit careful and not tell you anything that you may be alarmed with, or if you think that you have this and you’re going to be Type II or anything like that. So of course we’re being careful in the way that we present it. But there’s a lot of interesting information.

We’re also looking to do this in a very cool way that’s going to be, at least on the web – on the mobile it’s going to be a little flatter – when you sign into your web, there’s a report that’s going to be very interactive. You can dive in and go all the way down to the strain level, and then come up. So it’s going to be really cool in that sense.

[Damien Blenkinsopp]: So is there going to be, basically are you going to give all of that data?

My audience tends to be on the high quantitative side, so some of them tend to be people who download the data and start playing around with it in Excel. So will you have that kind of data?

With uBiome, for example, they have two aspects of that. They have the raw data they provide for you to download, and then you can put it into software to actually interpret yourself, like biometrician software.

And then they give you graphs which are basically summarized. So there’s not all of that information there, it’s a bit different, and it’s according to their perspective. So in comparison, what will you provide?

[Lihi Segal]: No, I don’t know to tell you that we’re going to give you all of the raw data. We probably could, but we haven’t finalized that down to the core of it. But again, we have it.

We’re going to have, as I said, the report and the very interactive tool so you can explore it. And the infographics is really cool. People are just playing here with it when they’re too tired to code. So they go and start planning that. But we could also provide the raw data, for sure.

Again, I think our users as opposed to uBiome users, uBiome users are mainly people who purchased it because they were curious about the microbiome. Our users, most of them, if I need to kind of guess or what I see, the microbiome is what gets them to say, oh this is really interesting.

This is personalized for me, I have my personalized microbiome; these people are scientific based, it’s not just that somebody came up with a diet based on my blood type, there’s science here. I don’t think that a lot of them are going to be very interested in downloading the file of the microbiome and things with it.

But we could definitely allow that, or be able to do that, if we see that there’s a need for that from our users.

(00:50:58) [Damien Blenkinsopp]: Yeah, cool. Alright. I saw there was a mention of a Mayo study on your site?

[Lihi Segal]: Where did you see that mentioned, by the way? I’m trying to figure out how did that get to you. We didn’t publish…

[Damien Blenkinsopp]: Well I don’t know, I think it was just mentioned. Oh, I know where I found it.

I was looking through your FAQ and there were some directions for Mayo study people on how to find the information.

There’s a leak there.

[Lihi Segal]:L: No, it’s not a secret by far.

We are recruiting people in the Mayo clinic now, and DayTwo is all over it. We just didn’t issue the press release saying that yet. But that’s been approved and it’s on it’s way as well.

So, what we’re doing, I’m happy to share, it’s no secret. But what we’re doing with the Mayo clinic is a clinical trial that is very similar to the clinical trial that The Weizmann Institute has done in Israel.

And so we’re recruiting 500 people and going through the same process of putting exactly the same device that was used in the trial in Israel and giving them test foods that are American foods, like a bagel and cereal, and really having them log their foods and providing all that information, and a lot of blood tests. So we’re really replicating the trial.

We’re just going to do that because we wanted to make sure we’re providing relevant recommendations after we have a basic cohort of US people. It doesn’t have to be the entire 500 completed, but we just, as the Israeli one was all Israeli, with Israeli microbiome and Israeli food, we just wanted to make sure that we’re able to calibrate the algorithm and it also works on a US based population with US foods and all that.

So we’ve already kicked that off. It’s a great collaboration for us to do this with the Mayo clinic, obviously. So we’ve already connected people. If any of your users are Rochester or Minnesota based people they can go and be part of that clinical trial.

[Damien Blenkinsopp]: Right. And it will be literally a copy of the other study so they could look at the other study to see what it would entail as well.

[Lihi Segal]: Right. There’s a bit of new information there as well. So that’s the reason we’re doing that. And also to start a collaboration with the Mayo clinic for other things as well.

(00:53:14) [Damien Blenkinsopp]: Great. Do you have a timeline for that? In terms of when you might get results eventually?

[Lihi Segal]: The timeline for US, it’s opened for pre-order. I know you probably entered through the UK, so you didn’t see that, because it’s IP based.

But if you were in the US you would see a pre-order. If you were in Israel, you could also buy and start getting it. So we started selling in Israel already.

The US is open on a pre-order basis, and we’re going to start shipping kits out to people in the beginning of 2017.

[Damien Blenkinsopp]: Okay.

[Lihi Segal]: It’s just around the corner.

[Damien Blenkinsopp]: Okay. So there are people already using this service in Israel, and it’s functioning.

[Lihi Segal]: In Israel we started the whole process of getting the evaluation, the kits, out to people and getting them back and sequencing them. We’re just starting to get, we’re in the final stage of getting the application finalized, and then getting the recommendations for people.

But there are a lot of people already who are using it because they got recommendations, whether from the Weizmann Institute Study or through us.

They’re not using the fancy application with the ‘Build Your Own Meal’, but the results and all of that have been around and have been used. Actually the BBC had a great show – I don’t know if you’ve seen it.

[Damien Blenkinsopp]: No.

[Lihi Segal]: The BBC has a show called ‘Trust Me, I’m A Doctor’.

[Damien Blenkinsopp]: I don’t watch TV here, unfortunately.

[Lihi Segal]: Oh, okay. So anyway, ‘Trust Me, I’m a Doctor’, it’s a doctor that has a show and she features clinical trials. And so she actually participates in the clinical trials that she features on her show.

So after the publication itself, she approached the scientists. She came to Israel with her colleague and was profiled and went though it, got food recommendations. Then she went back home and only ate what she was supposed to eat, lost weight and felt great, her energy levels [were up].

She was all psyched about it, and featured it on the BBC in a great show. I’ll send you the links so if you want you can see them.

[Damien Blenkinsopp]: Yes, please.

[Lihi Segal]: So there’s a lot of people who are using it, but outside of the clinical trial setting as well.

[Damien Blenkinsopp]: Okay, great. So it’s already getting around.

[Lihi Segal]: It’s getting contracts. Yes, we see the results are there.

(00:55:23) [Damien Blenkinsopp]: Yeah. Okay, so in terms of just how it’s going to be available, you’re only shipping to the US. So is no one in Europe is going to be able to do this?

[Lihi Segal]: Well, soon. We get a lot of approach on our support.

After the show was aired there was like 10,000 people hitting the website. So we know that there’s a lot of people interested. And we really want to go into selling in the UK as well. We’re just trying to be [safe], being a start up and not to jump too far ahead.

[Damien Blenkinsopp]: One thing at a time.

[Lihi Segal]: Right. So we did Israel because otherwise people will kill us here if we don’t bring it home. But we didn’t even translate it into Hebrew, it sold in English.

And we’re opening in the US because it’s an important market to start in. But we have concrete plans to get into Europe in 2017. So, soon. At least in the English speaking countries.

Really, logistically it just means that we need to get this box to people, but it’s not that simple. We will need a local database of food. So there’s some work on the server side to give you your foods and the database that fits you. We don’t think we’re probably going to need a trial to do that.

So in terms of the microbiome what we see is that the changes are not that [significant]. So there’s changes in the territories in the microbiome, but they’re probably not that apart compared to where the recommendations are. So you and I are very different in the way the algorithm predicts for us.

The microbiome is different, but it’s not that different. Anyway, it works on people. It could work on the US even without the Mayo trial.

[Damien Blenkinsopp]: So it sounds like that’s a validation effort.

[Lihi Segal]: Right, exactly.

[Damien Blenkinsopp]: I haven’t looked at studies of comparison of different countries and their microbiomes. There are some?

[Lihi Segal]: There are, if you look at the [57:12 check, unclear] that they have their graph there. So these show the US and there’s overlaps between the US, Europe, and Israel.

There are differences as well, but the differences, the way it reflects it in the algorithm is not that significant. So it works.

(00:57:33) [Damien Blenkinsopp]: Do you know when the Mayo trial, how long that’s going on for?

[Lihi Segal]: Oh, the Mayo trial will take a while. But we don’t need to complete the trial before we’re able to give recommendations. So we just need to validate it in a smaller group. But we’re there collecting data.

It’s more, you know in the US you can’t put a continuous glucose monitor on people at all if you’re not diabetic. Except under IRB kind of trial setting. So on a consumer level we couldn’t find any provider that would allow us to put continuous glucose monitors on healthy human beings without prescriptions. It’s a diabetic label from the FDA.

So we don’t have the device, and in order to really collect that data in the US we need to have a clinical trial set up and get the appropriate IRB and all that. So part of the whole doing of the Mayo clinic is because we just want more data, relevant data with glucose monitors and logging of food.

So we don’t need that to continue to start operating. I don’t even want to stop it after 500, so we’re talking about opening Arizona as a site, and Florida as a site. It’s really good just for our internal research purposes to continue to get more data.

(00:58:53) [Damien Blenkinsopp]: One quick question. I’ve noticed that Arizona comes up a lot in lab testing. I’m just wondering, as you brought it up just then, is there any reason?

[Lihi Segal]: Because Mayo has a site there. So when I’m collaborating with Mayo clinic, they have additional sites other than Rochester, Minnesota. So they’re thinking of expanding this to there and I’m more than happy to get more data.

[Damien Blenkinsopp]: I was just on holiday in Arizona and I just noticed that there are a lot of lab testing companies there.

[Lihi Segal]: It’s probably due to relevant man power and cheap, and something like that.

[Damien Blenkinsopp]: I think there’s maybe some state regulations or something that make it a little bit easier. Something like that also.

[Lihi Segal]: But again, when you sell outside of Arizona then you’re going to have to comply with the state laws anyway. So I’m not sure if that’s going to help you. But I don’t really know.

(00:59:41) [Damien Blenkinsopp]: So right now for the US is it $299 for the pre-order?

[Lihi Segal]: The price is going to be $399 but we’re opening up at $299, that’s a pre-order discount. But once we stop reordering, we’re probably going to go up to $399.

In Israel it’s 500 dollars, but we’re also doing a premium product in Israel; we’re giving continuous glucose monitors to people in Israel. So we’re giving them a fancy report on their blood sugar levels and all kinds of other stuff. We can because the device that I talked about in Israel you can put it on humans that are not sick.

[Damien Blenkinsopp]: Right, wow. That sounds like quite a service. If someone would pay 1000 dollars or more…

[Lihi Segal]: No, no, 500.

[Damien Blenkinsopp]: Oh, and they’re getting that premium service with the glucose monitor?

[Lihi Segal]: Yeah. It’s a lot, 500 dollars. It’s just more expensive than the US because of the continuous glucose monitor that we’re putting on.

[Damien Blenkinsopp]: They’re quite expensive, those things.

[Lihi Segal]: Well, they cost a few hundred. I guess in the UK it’s about 80 Euros. And then the reader and then the patch cost a little bit more.

[Damien Blenkinsopp]: I looked into getting one for myself; not for medical reasons, just to play around with.

[Lihi Segal]: Abbott Freestyle. Just take the Abbott Freestyle Libre. Just look for it. Freestyle Libre and then just order it online. And I think it costs 100 Euros or something.

[Damien Blenkinsopp]: Okay. And it’s got consumables on it too.

[Lihi Segal]: And then you have a patch. You get a round patch that you put on for two weeks. It’s good for two weeks. And then you have a reader.

[Damien Blenkinsopp]: And this is the one that Eran was talking about earlier that they’ve started using.

[Lihi Segal]: Right. So you can get that online.

We bought a bunch of them online ourselves in the UK before we found it in Israel. And once we found it here in Israel we decided to go with this product that we can also collect from people their blood sugar managements and give them all the fancy reports on all that. So it’s cool.

[Damien Blenkinsopp]: Yeah, it sounds quite exciting what you’re doing in Israel, because you’ve got more flexibility there. Are you publishing anything, maybe a bit later, about that on your customer base?

[Lihi Segal]: Not yet.

[Damien Blenkinsopp]: Okay.

(1:01:51) Is there anything we haven’t covered about the service, that we’ve missed?

[Lihi Segal]: Yeah. I think that this is kind of our direct to consumer approach. So we’re selling to you directly, but what we’re really working on is partnerships. Because what we really believe is that the way you’re going to use this is also very personalized.

Some people, the fact that we give them a fancy application that’s really cool and has a report on it and teaches them what to eat and what not to eat. There’s going to be a diet planner at some point on this. So you can really be independent in the way you manage your food.

For some people that’s going to be great, but some people really need more support. So maybe they go to Weight Watchers or they use other weight management services. And once you know as a user that there’s specific recommendations for you that are personalized for you, you really can’t tolerate generalized information anymore.

I’m saying this for myself. I go to this Weight Watchers group – it’s not Weight Watchers, it’s a local Israeli group. But I can’t hear her say to me, you should eat pretzels as a snack. 100 calories of pretzels are your snack. I’ve been doing that for 15 years, and then I found that it was my number 1 spiking snack.

And I moved to a different, totally different corn-based snack that was much better for me if I’m eating that 100 calorie snack already. So I’m snacking on that.

And what we’re thinking of doing is really opening an API with a lot of services. And so you as a user can share your information with your doctor, or with your nutritionist, or with your weight management group. Or when you take out food you want to be able to get a score. You want to log in, connect to…

[Damien Blenkinsopp]: So you could plug into a meal delivery site.

[Lihi Segal]: Think of this. Let’s say you’re ordering take-out of your food. We do this every day at lunch, just because in Israel is how it works.

And so I want to log in and connect with my DayTwo account, into that service, then get a menu and my score, A, or B. I’m already in a great restaurant, I’m eating food or I’m taking it out, I want to be able to get a score and choose right.

In the US specifically there’s a lot of employer wellness programs. All of those wellness programs provide nutritional advice, but it’s generalized. I, as a user, want my personalized advice to go with me.

So, that’s kind of the partnerships that we’re doing. Some will bring us customers, some we will bring our customers to them. And we’re building a marketplace around this.

So literally, think of that that we’re not competing with anyone. That’s the strategy that we built. We want to enable anyone who wants to use this personalized service to use it in their application and services.

[Damien Blenkinsopp]: Great, to make the information more widely available.

Lihi, it sounds great. I’m sure there are insurance companies and so on who would be interested in that as well. I know they’re getting more interested in these wellness programs.

[Lihi Segal]: Of course.

[Damien Blenkinsopp]: Okay well thank you very much for your time today. I really appreciated it.

[Lihi Segal]: Sure. Thank you so much.