Latest / HumanifAI HealthTech / Can AI Detect Disease Before Symptoms? w/ Dr. Guru Banavar, Founding CTO, Viome
Transcript
- 0:00What if your body could flag health problems before they show up?
- 0:04The warning signs may already be hiding in your RNA, the molecule that reveals which genes are switched on and what the trillions of microbes inside you are doing right now.
- 0:16Today's guest built his company around using AI to read RNA and reveal what's changing in your body to potentially catch problems long before they become symptoms.
- 0:35Welcome back to Humanify Health Tech, where we explore how AI is rewriting the rules of healthcare and putting people back at the center where they belong.
- 0:44I'm your host, Brian Chi, senior director of content at Net Health.
- 0:48Today I'm joined by Dr. Guru Banavar, founding CTO of Viome, where he leads the AI systems that decode an individual's biology from non-invasive samples, then turns it into a personalized diet supplement and diagnostic plan.
- 1:04Viome's AI system processes 100 quadrillion biological data points against a catalog of more than 100 million genes and genomes drawn from analysis of over a million microbiomes.
- 1:18I think that's about 100 billion millions of data points.
- 1:22Not sure.
- 1:23Before Viome, Guru spent 21 years at IBM, where he founded the Watson AI research team.
- 1:31In 2017, he left what some might call the best job in AI for a startup almost nobody had heard of.
- 1:38These days, Viome is pretty hard to miss.
- 1:41Guru, welcome to the show.
- 1:44Wow.
- 1:45That was an amazing introduction.
- 1:46Thank you very much for having me, and I think you actually stole a lot of my thunder already.
- 1:51But this is good.
- 1:52This is good.
- 1:53Uh- Awesome … so I can pick up from there and then take it to the next level.
- 1:55Well, let me tell you, you have so much thunder, though, because, because I didn't include, like, the Nobel stuff that you've done and the conferences that you've been, uh, that you've- Thank you … attended and that you've spoken at, and all the other things that you've done in, in your varied and, and, and really amazing, prestigious career.
- 2:12So- Thank you.
- 2:12Thank you, Brian… so I think you've got a lot more But be- Thank you.
- 2:16I-- Yes.
- 2:16I'm excited to talk about the topic that you, you have, uh, focused on in this show, which is that, uh, you know, health needs to be everybody's concern, number one, and health needs to-- or healthcare needs to put patients in the center, number two, and number three, I think preventive medicine is the, is the only way to really go.
- 2:37I think we do-- should not wait for diseases to happen and then react to it.
- 2:42I think we should be proactive, and we should figure out what's going on in our bodies before big problems arise.
- 2:49Awesome.
- 2:49Okay.
- 2:50So let's, let's get down to it then.
- 2:51Let's talk about your- Yeah … let's talk about your day job as the preeminent engineer and AI architecture visionary of Viome, right?
- 2:59Thank you.
- 3:00Yeah.
- 3:01Um, what you are building there is an infrastructure that, according to the website on Viome, will create a world where illness is optional, which is such an aspirational, such an aspirational statement.
- 3:13It's, it's, um, it's, it's amazing.
- 3:16Um, it's bold, right?
- 3:19But it does fit, and it fits with your story.
- 3:21So you go, you-- Just to, just to bring the audience up to, up to, uh, up to speed here.
- 3:27In 2015, you were standing on the Nobel stage talking about how Sometime in the near future, experts would one day work with learning and reasoning machines to come up with solutions and answers to human illness.
- 3:44Then fast-forward not even two years later, in January of 2017, you had what you called a sleepless night after meeting with Naveen Jain, the founder and CEO of Viome.
- 3:56Yeah.
- 3:57Walk me through that, that, um, that epiphany and what happened there.
- 4:01What was sleepless about it, and what was so inspiring that it made you walk away from Watson?
- 4:06Yeah.
- 4:07Yeah.
- 4:07Brian, what a great question.
- 4:09So first of all, let me tell you, uh, some personal, uh, anecdotes, and then I wanna, I wanna sort of step back to the whole sort of the global stage, right?
- 4:17I've had a lot of people in my family and in my circle who've had chronic disease.
- 4:23Uh, chronic disease that has resulted in great tragedy, great, uh, suffering, great pain and, and loss of quality of life, and loss of life, actually.
- 4:34So, uh, there have been times when I, I ask myself, what am I doing, uh, you know, building all these great new technologies if I'm not able to help the people closest to me and, uh, help them live a good life, live a long life?
- 4:52You know, that's what we call health span.
- 4:54We want everyone to be able to do that, and I have multiple cases in my own family, um, that, uh, that kind of raised that consciousness awareness in my, in my mind as I was hitting those, uh, years of, uh, uh, understanding the power of AI, which was sort of in the mid-20, 2010s, 2015.
- 5:15And, um, at the same time, what I was noticing was in the global view of the world, um Chronic disease has now become the biggest epidemic, right?
- 5:28It's one of the big human problems that i- is about the same level as, you know, poverty and hunger and, uh, global warming and so on and so forth.
- 5:38Chronic disease is something that every single person that I know either suffers from or knows somebody who suffers from, right?
- 5:45So it's literally like 70, 80% of the population of the planet have some known chronic disease that they're worried about, right?
- 5:53Turns out that in the developed world, it's actually worse than the developing world.
- 5:58You know, there's a lot of reasons for that, including the Industrial Revolution, including the, uh, lifestyles that we all follow in, in the modern sort of environments with lots of toxins around us, antibiotics around us, which destroys things like our microbiome, for example, right?
- 6:14So that was a global kind of perspective that also made me feel like, you know, I should do something that's much bigger than myself, which is to help people find a way to live a more healthy life.
- 6:27So, uh, believe it or not, during my last couple of years at IBM, I tried very hard to apply a lot of things that we've learned from the Watson experience to the healthcare industry, including the cancer, including the genomic work, and including some projects and everything.
- 6:48But I, the, I was not satisfied.
- 6:50I wanted to do that full-time, and I wanted to do it directly.
- 6:54So when I called Naveen and when I asked him about, uh, you know, his advice, yeah, I've known him for a long time, and he's a great friend of mine.
- 7:00And so I asked him, "So what do you, you know, what do, what, what do you think I should do?
- 7:03I'm planning to go into the healthcare world." He says, "Holy I am doing something in the healthcare world.
- 7:08Why don't you join us?" And so he and I and a bunch of other people found a way to meet in San Diego one Friday afternoon around 2:00, and we started laying out what could be an agenda for a startup.
- 7:24And of course, Naveen had already and, and, and other folks have already thought-- came in with their thoughts.
- 7:29And, uh, by the time we were done, I think it was 2:00 a.m. that day.
- 7:33So it's a 12-hour, really intense discussion, which is the reason I couldn't sleep that night because it was just so intense.
- 7:41It was so aligned with everything I wanted to do that, you know, I, I felt like this is the exact right group.
- 7:47And when I got up in the morning and, you know, I, I kind of just rolled out of bed because I did not, hadn't, had not slept well enough.
- 7:54But I, I had to go and meet everybody for, for first thing in the morning coffee and say, "Listen, I'm in. I'm, I'm getting out of IBM. I'm doing this next." So that's my story of how, how I, uh, got into the Viome founding team, so to speak.
- 8:11And, uh, and here we are, you know, about, about, you know, not quite 10 years, but will be 10 years soon.
- 8:17Do, do you, do you think that you're closer to that aspiration that you walked into with today than you- Absolutely … than there?
- 8:24Yeah.
- 8:25A- ab- absolutely, Brian.
- 8:26I mean, you know, the numbers that you mentioned earlier, right?
- 8:28Is, is that those are just starting numbers.
- 8:30And, you know, there is a lot of, I would call it, uh, newness in the direction that we have taken, and the novelty will take time for people to understand and, and, uh, and embrace.
- 8:46And we have now started getting into the healthcare, health, health professional community.
- 8:53And, you know, all the healthcare professionals have been trained in the, I would call it the previous generation of medicine with, you know, with either DNA or with proteins or other markers, you know, chemistry markers that, uh, in many cases were validated in very small scenarios that may not have actually passed the, you know, the bar that, that we apply today for many of these kinds of, uh, illnesses and so forth.
- 9:21So today, given that we have started the metatranscriptomic, the RNA sequencing and AI methodology, I think we've made tremendous progress, and I think over the next few years I can see us taking off, uh, hockey stick-like and, uh, hopefully all of the healthcare professionals that we have started to talk to will, uh, help us get the knowledge base and the credibility within the groups that will get the adoption across the entire industry.
- 9:55That's, that's what I'm hoping for in the next few years.
- 9:58I mean, that idea of, that idea of RNA versus DNA w- that sounds like that was a real key part for you in terms of making that decision 10 years ago, or the, the idea of using RNA- Yeah … to map that, to map this is w- is still pretty landmark, right?
- 10:16And that's a- Yeah um, that, that's the level of complexity is so much greater, but also the ability to be more accurate is also so much greater.
- 10:24Is that right?
- 10:25Yeah.
- 10:25Yeah.
- 10:25Yeah, Brian.
- 10:26I mean, it's, it's a very simple thing.
- 10:27You know, DNA is a static molecule, right?
- 10:29You're born with some DNA.
- 10:31You, you will die with it.
- 10:32You, I mean, the, the, i- if you're healthy last year and if you're sick this year, your DNA did not change.
- 10:39Your DNA is probably exactly the same.
- 10:41I mean, even if there are small mutations, they get fixed very quickly in, in the DNA, but it's more or less the same.
- 10:47But your RNA expression changes dramatically between when you're healthy and when you're sick.
- 10:53So RNA essentially captures the state of your body, whether in health or sickness, different types of sicknesses and so forth.
- 11:02That's one big bet we made when we started Viome.
- 11:05The other big bet we made is that, you know, guess what?
- 11:08More than one cell in your body is non-human for every human cell in your body.
- 11:16So, you know, for a, you know, in, in other words, you have 1.3 microbial cells in your body for every human cell.
- 11:22So that's A tremendous amount of biology that is happening in your body that you may not have even realized, you know, you, you are, you are being influenced by.
- 11:33And there's a-- actually, there's 100 times more microbial genes compared to human genes.
- 11:38So the interaction among the human genes and the microbial genes and understanding the activity and that interaction was a second big bet we made.
- 11:48And the third big bet, of course, is we said, listen, we have to collect a lot of data because without the data, you cannot really understand biology.
- 11:58You know, biology at the end of the day is an information science in my mind.
- 12:01So an information science needs a lot of data, and once you have the data, you can use all the AI techniques that we know today, and, you know, we keep inventing new types of techniques, and we can apply all those techniques to the dataset.
- 12:14So we said-- on day one, we said, "Let's build a, an MVP of a product that we can start selling direct to consumer and then collect the data and then do the R&D in a sort of a flywheel fashion to understand biology across the board." Now, 10 years later, we have developed probably hundreds of pathways from RNA data.
- 12:38We have published dozens of papers.
- 12:41We have patents across the board from the metatranscriptomic RNA processing all the way up to the clinical applications that we've built, which includes things like recommendations, things like disease analysis, things like, uh, glycemic response prediction, you name it, right?
- 12:58Many different kinds of diagnostics and so on.
- 13:00So we have that whole stack built up.
- 13:03And, and then I think we can take that now to essentially every aspect of healthcare.
- 13:09Whichever disease, whichever human system you're looking at, RNA plays a very big role, microbiome plays a very big role, and we now have the data and the AI to show you what is happening in any aspect of, of health that you're interested in Okay.
- 13:26So let me ask you, let me ask you, let me bring this into something I can understand.
- 13:33Okay.
- 13:33Great.
- 13:33So, so the, the R… I understand that RNA gives you a picture of where a human is today currently, right?
- 13:42In their terms of their health and what they're doing and their, um, um, all of the inputs that are going into their, into their lived experience.
- 13:49Correct.
- 13:50Um, how does that though, given the amount of change that happens over the s- over the, over the span of a year, let's say, how does that make what you're doing more accurate to the health and potential markers that you see in a human that could actually result in knowing, knowing about conditions before symptoms are present?
- 14:15Do you see, do you see what I mean?
- 14:17What, what- Oh, yes.
- 14:17Yes, yes … how does that, how does that work?
- 14:20Yeah.
- 14:20So let, let's, let's start with the basics, right?
- 14:23So your body has an immune system, right?
- 14:28You are always looking for a variety of external factors and triggers that you need to defend against and you need to overcome in some fashion.
- 14:40Now, the process of reacting to these external triggers sometimes results in this notion of inflammation, right?
- 14:50Turns out that your body is constantly exposed to the external world, not just your skin on the out- outside, but also your, your food canal that, you know, when you eat through your mouth all the way to the other end, your body is exposed to all of the environmental factors, and there's a lot of immune system, uh, activity across the length of your food canal, right?
- 15:17So think about that system for a minute, and you ask yourself the question When does inflammation help you and when does inflammation hurt you?
- 15:30Because inflammation is happening in that food canal constantly because there's food coming in and there's, by the way, microbes and other kinds of factors in the food that you're eating that is impacting and triggering the, the inflam- inflammatory system.
- 15:45Turns out that if we can detect how inflammation is behaving, meaning is it going up, becoming active, defending, and then going back down, meaning it gets activated and then get, gets resolved, that's a healthy sign.
- 16:02But if it goes up and if it stays in a chronic state of inflammation, low-grade inflammation for a very long time, that's not good because that means your body is working overtime to fight something that it should not be fighting.
- 16:17So that could happen if your microbiome, for example, in the gut, and I can say the same thing about your mouth, about the other parts of your body and so forth.
- 16:26But if your microbiome is not in a good shape, meaning, you know, some people call it dysbiosis, but let's just say that it's not in great balance, meaning the good guys, bad guys, the neutral guys, all of them are constantly kind of living and fighting with each other and learning to live with each other and so forth.
- 16:45But if one of them overtakes the other and, and they end up creating a lot more inflammation, and it's constantly driving your immune system, that results in a low-grade inflammatory, uh, signal that results in heart disease, it results in metabolic disease, it results in neuro diseases, it results in autoimmune type diseases, r- results in lots of different things.
- 17:12So the question is, how early should we detect inflammation in your gut, and what can you do about it?
- 17:21Turns out, if you are looking at RNA, you can see the genes that are expressed on the microbial side, the genes that are expressed on the human side, which are the, the immune system genes that are expressed on the human side, and you can see the interactions among them, and you can see that inflammation is not in a good place, as an example.
- 17:45That's-- I'm just picking one particular function, which is inflammation.
- 17:49I could pick something else.
- 17:50I could pick oxidative stress.
- 17:51I could pick insulin resistance.
- 17:53I could pick other functions as well.
- 17:55But I picked inflammation because it's a pretty fundamental biological function.
- 17:59So because we are able to look at the gene expression, we are able to understand the biological functions that are active inside your body at any given point in time.
- 18:11Same thing on the microbial side.
- 18:13A microbe like Akkermansia, everyone's heard about Ak- Akkermansia I think.
- 18:17Akkermansia muciniphila, right?
- 18:20It can be very good for you for metabolic disease, or it can be very not good for you.
- 18:26It can result in Or multiple sclerosis, for example.
- 18:29So it depends on which part of its genome it's expressing.
- 18:32So if you look at the expression of the microbes and the expression of the human cells, and if you're able to understand what pathways are dynamically active, that can give you a lot of clues about what that could lead to.
- 18:47High-grade in- or low-grade inflammation for a long time can lead to all these diseases that I mentioned earlier, right?
- 18:53So coming back to it, I, and I hope this is making sense By looking at DNA, you will not be able to tell whether somebody's inflamed or not.
- 19:05But looking at RNA, you can.
- 19:07And because you know that somebody's inflamed in your gut, in your mouth, in your blood, and so forth, you can then put those data points together and you can say, this person is potentially going through an inflammatory bout, and that could lead to heart disease, for example, or some other diseases.
- 19:25Inflammatory bowel disease is a very common thing.
- 19:27Inflammatory bowel disease is, is one of the, one of the most important things in the, in the gut that we can detect.
- 19:33So putting that together is, is what AI does, right?
- 19:36Yeah.
- 19:36So what, what is the-- what ha- what has happened with AI that is a, that has enabled this to happen personal-- i-in a personalized way at scale?
- 19:49What, what, what's, what, what is it about AI now compared to, let's say, 10 years ago or seven years ago, that allows this sort of transformation to occur?
- 19:59Yeah.
- 20:00Uh, another great question.
- 20:02And, and so first of all, you have to start with, uh, the data that we generate for each individual.
- 20:07So if, if-- Let, let's just, let's just start with the basics.
- 20:09If you get a Viome test, either through your doctor or as a consumer, you can go to our website and get a consumer test, you would typically send three samples in.
- 20:20You would send a stool sample, you would send a blood sample, finger prick blood at home, and you would send a saliva sample.
- 20:28Now, each of those three samples, when it arrives in the lab, it goes through our RNA sequence, sequencing technology assay.
- 20:36It generates literally hundreds of thousands of data points.
- 20:40So if you, if you look at our app today, you will see that, um, you know, those three samples put together is, you know, something in the range of about 300,000 or 400,000 data points that are generated.
- 20:52Now, these are interpretable data points.
- 20:54There's a- actually a whole bunch of data that is still not interpretable because science has figured out-- hasn't figured out exactly what those things mean.
- 21:01And as we learn more, we will incorporate those as well.
- 21:04But today, you can get a, a few hundred thousand data points that are interpretable.
- 21:10Now, interpretable by what?
- 21:13By whom, right?
- 21:15A human being looking at 300,000 data points on a sheet of paper would never be able to interpret it.
- 21:21So you need AI to be able to put the dots together and connect it to the biological functions that we were talking about.
- 21:29For example, which enzymes on the microbial side and on the human side, when they're expressed together and are interacting with each other Is a marker of inflammation.
- 21:42In order to do that, you need to understand what are the mark-- what are the enzymes in your, all of your samples that are responsible for inflammatory activities.
- 21:53Now, what our AI has done, and this is the, the training that we've done in the last 10 years, is we have learned from hundreds of thousands of people who have some inflammatory disease, what kinds of genes are expressed in those individuals relative to people who don't.
- 22:16So you can actually see the contrast of the kinds of genes that are expressed on the human side and the microbial side with people who have in-in-inflammatory diseases versus people who don't.
- 22:27And so once you know that, you mechanistically understand these are the enzymes and molecules and transporters and other regulators that are responsible for inflammatory activity, and we've trained models to be able to put the data together for you as an individual and tell you what is the level of inflammation, in our case, on a scale of zero to 100, right?
- 22:52So o-our AI basically has taken millions and millions… Actually, at, at this point in time, we've generated, you mentioned the 100 quadrillion, right?
- 23:02That's, that's approximately 10 to the-- Well, it's 10 to the 16, getting to 10 to the 17.
- 23:10That's kind of the, the scale of the data.
- 23:11Just for, for people who may not, uh, see the, the scale of that, uh, all of the data that's on the internet is approximately 10 to the 24 Today.
- 23:25While it's, it's growing, but, you know, going from 10 to the 24 to 10 to the 25 is a huge step.
- 23:29And Viome's molecular data is in the range of tw- 10 to the 16 or 10 to the 17.
- 23:34So that's, that's a humongous amount of data, and, you know, that's petabytes, hundreds of petabytes of data.
- 23:40So we have learned from that data set how to identify, let's say, inflammatory pathways or oxidative stress pathways or metabolic pathways and so on and so forth, and tell you, "Look, your metabolic pathways are not doing great.
- 23:56Your insulin resistance is not great.
- 23:58Your oxidative stress is not doing great, and therefore you may have a risk of this disease or that disease." Now, that is the work that our AI has done over the last 10 years.
- 24:10And, you know, of course, we've built this bit by bit, and we've published most of this data, and, and your, your question of what has the new AI techniques helped us do?
- 24:20The new AI techniques have helped us interpret a lot of the known science and understand the mechanisms of action.
- 24:32So, for example, if you look at all of the enzymes in a particular pathway, and
- 24:37you ask yourself the question: How do these pathways interact with each other?
- 24:41you ask yourself the question: Today's AI is able to help us understand that better than we were able to do seven years ago or 10 years ago.
- 24:48you ask yourself the question: But I wanna be clear that the AI we built in Viome is not a large language model AI.
- 24:55you ask yourself the question: It is a molecular AI because it's looking at your RNA molecules, both on your human side and the microbial side, and we've trained our, our AI with the molecular data from these million-plus samples that we've collected.
- 25:11you ask yourself the question: So we have a molecular AI that builds the pathways and understands it, and once we get the results from there, we can use a large language model to interpret that using today's knowledge of the published literature that we couldn't do as easily before.
- 25:28you ask yourself the question: So now we can do that interpretation part of it and explanation part of it and interaction with users much more easily than we were able to do six, seven years ago.
- 25:37you ask yourself the question: How much of what the finding is, is AI's own judgment here?
- 25:42you ask yourself the question: And how much of it is what you just mentioned, being pulled from published research or published, published, published findings?
- 25:50you ask yourself the question: So I'll- Is that, um… Is there a, is a s- is there a sense of AI sort of making that judg- making those judgments and making those recommendations?
- 25:58you ask yourself the question: Yeah.
- 25:59you ask yourself the question: So our biomolecular AI is trained to understand specific pathways and specific functions in biology, like glycemic response, like inflammation and so forth.
- 26:11you ask yourself the question: Those are all trained models, and they're not non-deterministic models like large language models, which if you ask the same question two times, you get different answers.
- 26:19you ask yourself the question: These are deterministic models that we have trained over the past several years.
- 26:25you ask yourself the question: So if you ask for your sample A and sample B, if you send two samples, let's say from the same biological sample, we divide into two, two technical replicates, we run it through the model, you will get more or less the same answer Okay?
- 26:40you ask yourself the question: And, and I say more or less because there's always, you know, a little bit of assay, uh, variability in there, but it's not significant, right?
- 26:48you ask yourself the question: So you get the same answer, and that answer will tell us what is the level of, let's say, inflammatory activity or whatever other activity that we are looking at.
- 26:57you ask yourself the question: So that's a deterministic AI model.
- 27:00you ask yourself the question: The interpretation part, which is the user interface, which says, "Okay, your, uh, pathway XYZ is at this level, and underlying that are the following specific enzymes, and un- and because of that, you have risks for disease A, B, or C," that story is something that you can use a large language model to help it.
- 27:26you ask yourself the question: So when, when you get into the Viome app, for example, there is a chatbot.
- 27:29you ask yourself the question: You can ask the chatbot, "Hey, what does this inflammatory activity pathway mean?" And it'll put together a, an interpretation of our, um, molecular AI's output.
- 27:40you ask yourself the question: But even, even the large language model chatbot that we have has been trained with our internal data to be able to interpret it in the best possible way.
- 27:49you ask yourself the question: So that's kind of the s- the spectrum of AIs that we have built and, and we have integrated into our system.
- 27:55you ask yourself the question: So you said a couple times, I'm gonna go back for a second 'cause you s- you mentioned a couple times healthcare professionals as well as consumers being sort of your audience or your, your, um, ideal, ideal customers, right?
- 28:06you ask yourself the question: Yes.
- 28:06you ask yourself the question: The, the idea being, um, uh, somebody could go to your website and order, um, and, and, and, and go direct, um, or you go through a healthcare professional, um- Yes … to get the, to get the service.
- 28:19you ask yourself the question: Which, which, which do you prefer?
- 28:23you ask yourself the question: Which-- What is a be- what is a better play here for, for, for, for somebody or for, you know, for healthcare professionals?
- 28:30you ask yourself the question: Well- I can imagine that a clinician is looking at this science and thinking to the, this, these AI models and thinking,"This could really help me t- as a diagnostic tool." Is that, is that right?
- 28:41you ask yourself the question: Absolutely.
- 28:41you ask yourself the question: Absolutely.
- 28:42you ask yourself the question: So the, so I prefer the clinical grade tool definitely for multiple reasons.
- 28:48you ask yourself the question: Number one reason is because the clinical grade tool is sequenced more than three times the depth of the consumer grade.
- 28:58you ask yourself the question: So c- the consumer grade gives you like the first kind of high level picture, but the clinical grade gives you the deep picture three times more, 300% better, um, you know, uh, resolution of the data, so to speak.
- 29:10you ask yourself the question: So I prefer to do that.
- 29:12you ask yourself the question: And then you, you also get a very extensive analysis from a clinical standpoint.
- 29:20you ask yourself the question: You get a clinical report which has all the information that you need.
- 29:23you ask yourself the question: You can look at the high level if you don't have time.
- 29:25you ask yourself the question: You can look at the first page and you get most of what you need.
- 29:27you ask yourself the question: But then if you wanna know more, there's a huge amount of additional data that you can get from the clinical report.
- 29:33you ask yourself the question: And then, of course, if your, your patient wants to see the data, you can ask your patient to also get their own version of the report or the app from which they can get all the information.
- 29:46you ask yourself the question: Now, I'm not saying that you need to absolutely have the clinical grade test for every possible case.
- 29:53you ask yourself the question: There, there could be individuals who say, "Listen, I wanna take my health in my own hands.
- 29:58you ask yourself the question: I wanna know everything that I can know." So you can go to our website and you can directly order as a consumer if you want to.
- 30:06you ask yourself the question: But I would say that if you want the absolute highest degree of clinical grade results and explanation and interpretation and follow through, I would definitely go with, uh, the healthcare professional-oriented version, which by the way is called Viome Precision Health Pro, which also has the three s- three samples I mentioned before.
- 30:28you ask yourself the question: That's diagnostic, that's, that's, that's discovering biomarkers, things of that nature.
- 30:33you ask yourself the question: Whereas- Correct … the consumer side of it is more about wellness recommendations, things like that.
- 30:37you ask yourself the question: Is that right?
- 30:37you ask yourself the question: Yeah.
- 30:38you ask yourself the question: Correct.
- 30:38you ask yourself the question: Correct.
- 30:39you ask yourself the question: Correct.
- 30:39you ask yourself the question: So this is, uh, this is, uh, you know, think of it as a, a clinical grade screening tool that gives you a lot of information about all of the different systems in your body and then tells you what the risks are, and then also provides a lot of recommendations to healthcare professionals about what they could do going forward from that point onwards.
- 30:59you ask yourself the question: So it, it catches everything as early as it can possibly catch.
- 31:04you ask yourself the question: And, and oh, by the way, I should also mention, Bryan, before we end the show, is that we have developed biomarkers for even cancers.
- 31:12you ask yourself the question: One of the, one of the first cancers that we did the biomarker for is oral and throat cancer using a saliva sample We are now working on, uh, pancreatic cancer using a blood sample.
- 31:26you ask yourself the question: And of course, in the future, not too distant future, we're gonna have colon polyps, which are precursors of colon cancer, from a stool sample, right?
- 31:34you ask yourself the question: So from each one of the three sample types, we have these other, you know, very difficult diseases that we look for biomarkers for and do early detection of.
- 31:45you ask yourself the question: And there's all the publications and all of the clinical data on our website about all of these tests, and you can see why the clinical grade test is able to give you that level of insight.
- 31:57you ask yourself the question: Okay, two more questions, and then I'll let you go because I, I know you, I know you're a busy man, and this has been such a fascinating conversation.
- 32:03you ask yourself the question: Thank you.
- 32:03you ask yourself the question: Um, so thank you for the time.
- 32:05you ask yourself the question: Of course.
- 32:06you ask yourself the question: First, first question, though, is this, is at the start, I mentioned value-based care, and I did that for a reason because of our-- because of this conversation, right?
- 32:15you ask yourself the question: This-- That deeper level of w- via Viome Precision, does that offer healthcare professionals, clinicians, providers an opportunity to actually transition into more of a value-based care payment structure than what they're probably currently doing with fee-for-service because of the predictive sort of quality to what Viome is, is producing?
- 32:41you ask yourself the question: Is that-- Do I have that right?
- 32:42you ask yourself the question: Is that- Absolutely … is that, uh, w- tell me a little bit about what y- what Viome is thinking in terms of that, um, that frontier.
- 32:51you ask yourself the question: Well, from a health economic standpoint, right, you wanna look at the total cost of care at, at, you know, through the life cycle.
- 32:59you ask yourself the question: Now, if you wait for a disease to surface through whatever symptoms for the particular disease, and you start process-- you start Uh, treatment plans at that point, and you're reacting to an advanced disease, you are obviously gonna spend quite a lot more money than you would have.
- 33:19you ask yourself the question: Uh, I mean, across this, across the whole system.
- 33:21you ask yourself the question: I'm not talking about any particular individual or entity within the system, but across the whole system, you're gonna spend a lot more on, um, on providing care for an advanced patient for any disease, right?
- 33:35you ask yourself the question: The whole point and the, the premise of Viome is that you wanna catch every one of these diseases as early as possible, and in fact, you want to look for the biological functions that may be leading towards the disease even before you start seeing preclinical symptoms, right?
- 33:53you ask yourself the question: So that's what I mean when I say inflammation or oxidative stress, all of these other factors.
- 33:58you ask yourself the question: So from a, from an economic standpoint, if you handle the testing and the care early through regular safe GRAS-type in-interventions, you will obviously, overall as a system, you spend a lot less.
- 34:14you ask yourself the question: So I think that gives you the value-based care kind of, uh, an argument right away, right?
- 34:21you ask yourself the question: Whatever, whatever you're spending, you're, you're doing it based on the value that you're getting for every dollar that you spend, and for, for patients as well.
- 34:31you ask yourself the question: You don't have to worry about all of the, you know, the, the fee for payment and everything else downstream.
- 34:37you ask yourself the question: Okay.
- 34:37you ask yourself the question: Last question.
- 34:38you ask yourself the question: A-and, and I wanna circle, make it, make-- try to make this as full circle as possible.
- 34:43you ask yourself the question: Okay.
- 34:43you ask yourself the question: Value, value-based care is about quality of care.
- 34:47you ask yourself the question: Mm-hmm.
- 34:48you ask yourself the question: That's the whole life cycle.
- 34:49you ask yourself the question: That's the predictive nature of it.
- 34:51you ask yourself the question: That's where, where you look at the whole experience of a human being.
- 34:55you ask yourself the question: You started this conversation talking about this, what Viome does as being a, uh, being something that could actually lift, uh, a world regardless of station, regardless of economic station, to better health.
- 35:12you ask yourself the question: Yeah.
- 35:12you ask yourself the question: Where are you at with that, and w-where, wh-where do we go to achieve that ambitious aspirational goal, that dream?
- 35:22you ask yourself the question: Brian, great last question.
- 35:24you ask yourself the question: Uh, let me tell you this.
- 35:26you ask yourself the question: If you were-- wanted the level of science and technology that is built into a Viome test and you tried to get that from anywhere else, it will literally cost you thousand plus dollars.
- 35:39you ask yourself the question: M-multiple thousands of dollars potentially.
- 35:43you ask yourself the question: Okay?
- 35:43you ask yourself the question: Now we have brought down that cost to a couple hundred.
- 35:48you ask yourself the question: You know, for a stool metatranscriptomic test, it's just a couple of hundred dollars to, for you to get started with.
- 35:54you ask yourself the question: Now- Will this reduce further in the future?
- 35:58you ask yourself the question: Absolutely, because sequencing cost is going down.
- 36:02you ask yourself the question: I think our operational costs are going down because we're scaling, and at the end of the day, the costs are going to be at the point where anyone should be able to get this kind of a test literally when they think about it.
- 36:17you ask yourself the question: And, and, you know, in my, in my mind, every doctor who does any kind of, um, intake for any patient can get one of these tests to get like a general overview of what somebody's health status is.
- 36:31you ask yourself the question: So it becomes like a least common denominator that everybody has to know about themselves and can tell their healthcare professionals and healthcare team what their current status is across the board.
- 36:46you ask yourself the question: So I look at this as the, you know, the democratizing view of, of healthcare because everybody can get this without spending too much money, and also not waiting until healthcare costs a lot of money, right?
- 37:01you ask yourself the question: I mean, people spend I don't know how many hundreds, thousands of dollars on even on insurance, right?
- 37:07you ask yourself the question: The question is, what is the value that you get back from it?
- 37:10you ask yourself the question: And I would, I would say that doing the preventive care and increasing your quality of life gives you a lot of the benefits and the value than any other form of, of healthcare in my mind.
- 37:24you ask yourself the question: And that's what, that's what I do, by the way.
- 37:26you ask yourself the question: I, I, I do this test every six months, every, every year, and I know what I need to do better, what I need to eat better, what, you know, what kinds of exercise or sleep or what kinds of issues I need to address better, and I try to take care of them as early as possible.
- 37:42you ask yourself the question: So I'm hoping that more and more people get on this particular agenda and, um, healthcare professionals are able to interpret and help other people get on this agenda as well.
- 37:54you ask yourself the question: Okay.
- 37:54you ask yourself the question: One th- I got one, one more question I gotta ask you.
- 37:56you ask yourself the question: You have one more.
- 37:57you ask yourself the question: Okay.
- 37:57you ask yourself the question: So, okay, 11 years ago when you were standing on that, on that stage and you were talking about that, you know, experts and, um, machine learning- Yeah you reasoning, did you, did you envision this?
- 38:10you ask yourself the question: How far away is this from what you th- what Guru from 10 years ago, 11 years ago saw?
- 38:16you ask yourself the question: Oh my God.
- 38:17you ask yourself the question: I think we are very close to what I, what I had in mind at 11 years ago.
- 38:22you ask yourself the question: It's, uh, so great that you asked me that question because I don't get to reflect back on all of these things that we were thinking about.
- 38:28you ask yourself the question: I was talking about back, you know, on that Nobel stage and on many other, you know, the Turing stage and everything else that, that I spoke on.
- 38:35you ask yourself the question: I said, "You gotta have an angel on your shoulder which will help you Live a better life and tell you when you're doing something great for yourself and not doing great for yourself, right?
- 38:48you ask yourself the question: I think, you know, literally this week we launched an AI tool that gives people that kind of guidance on top of all of the analysis and all of the, um, insights that are provided in the app.
- 39:03you ask yourself the question: And this kind of tool, AI tool gives you that, that kind of, uh, you know, whispering on, in your ear, uh, your guardian angel whispering in your ear, both if you are a patient or consumer, as well as when you're a doctor, so that you can help other people and you can help yourself.
- 39:21you ask yourself the question: So I'm very happy with how far we've come in the last 10 years.
- 39:24you ask yourself the question: So I'm really, uh, uh, thrilled to, to see that we are going the right direction, and I hope that we keep going, going in this, in the same direction and get it across to everybody on the planet.
- 39:36you ask yourself the question: Guru, thank you so much for your time and this amazing discussion.
- 39:39you ask yourself the question: I know I learned a lot, and I'm sure the audience did as well.
- 39:43you ask yourself the question: You're doing important work- Brian, thank you … and, and I'm excited to follow where Vieum goes from here.
- 39:48you ask yourself the question: Uh, thank you so much.
- 39:49you ask yourself the question: Your, your questions were amazing.
- 39:50you ask yourself the question: Your discussion points were great, and, uh, I hope that the audience got something from, from all of the discussions we had.
- 39:56you ask yourself the question: Awesome.
- 39:57you ask yourself the question: Awesome.
- 39:57you ask yourself the question: Thank you.
- 39:58you ask yourself the question: Thank you.
- 39:58you ask yourself the question: Thank you again.
- 39:58you ask yourself the question: Thank you very much.
- 40:04you ask yourself the question: All right, recap time.
- 40:06you ask yourself the question: Here are the top three things I took away from my conversation with Guru today.
- 40:11you ask yourself the question: First, DNA knows what your biology is capable of.
- 40:15you ask yourself the question: RNA shows what it's doing right now.
- 40:19you ask yourself the question: That distinction is the reason Viome exists, and it's also a real challenge to a decade of health tests built on DNA alone.
- 40:29you ask yourself the question: Second, wellness recommendations and the diagnostic biomarker work are two different claims carrying two different levels of weight.
- 40:37you ask yourself the question: One is a supplement list.
- 40:38you ask yourself the question: The other could be a disease detection tool.
- 40:42you ask yourself the question: Judge them on their own evidence, not as a single package.
- 40:47you ask yourself the question: And finally, Guru left one of the most secure roles in AI on the impulse of one sleepless night, but the desire of decades.
- 40:55you ask yourself the question: That kind of courage is how progress is made when the stakes are doing things like making illness optional.
- 41:02you ask yourself the question: And that should do it for today's episode of Humanify Health Tech.
- 41:06you ask yourself the question: As always, thanks for tuning in.
- 41:08you ask yourself the question: If you're joining us on Spotify or YouTube, we'd love to hear your thoughts in the comments, and don't forget to leave us a rating while you're there.
- 41:15you ask yourself the question: Once again, I'm Brian Chee from Net Health, and I'll see you next time.