We go around walking through life, not knowing that the paths we take might blow up in our face.
And when I’m talking about explosions-blowing-up-in-the-face, I’m talking about chronic diseases. Chronic diseases are long-term illnesses that are mostly caused by the paths we take in life: our lifestyle. This means that most chronic diseases, like heart disease or diabetes, are 90–100% preventable.
But chronic diseases are the number 1 leading cause of death and disability in the US.
This is absurd.
How is something that is almost entirely preventable the leading cause of death in the United States?
Well, there’s really only one major obstacle:
It’s human nature for people to continue their current lifestyle without knowing a personalized, specific alternative. It’s human nature for people to ignore the warnings of those who know better. Without any stimulus, without any personal reason to change, people follow Newton’s First Law to the letter: they stay at rest.
And our only solution against these preventable chronic diseases right now are the basics: eat, sleep, drink, and exercise well on a daily basis.
But we ignore these basics. On a daily basis.
And these basics are just too broad so they fail against human nature. They don’t provide that personal reason to change that human nature requires. What does it even mean to eat healthy? Is my best option to go for a Keto diet or Mediterranean? (What’s the difference?) Is my version of eating healthy different from yours because of our different genetic makeup?
This question highlights one of the two big hurdles in our fight against preventable chronic diseases. Simply put, we don’t understand our health and all the complex genetic and epigenetic factors that affect it in order to generate personal health plans — the steps a person needs to take in order to stay healthy.
Personal health plans not only provide the map to avoid preventable chronic disease, they provide that personalized reason to change and thus are a huge help in our fight against human nature and preventable chronic diseases. With a personal health plan, people won’t ignore the steps they need to take healthy because those steps are literally built for them.
If we were able to have this understanding of the causal factors between genomics and epigenomics, we could generate personalized health plans and recommendations based on the data of individual patients. But we need a little something extra too.
While having a personalized health plan is a huge step forward towards living healthier because it provides that personalized reason to change, it does not solve all our problems. The human psyche is a little bit more complex than that.
In order to give that extra push, that extra nudge, our solution is not the only thing that needs to be personalized. The provider of that solution needs to be personalized as well. An Olympic level swimmer can’t reach greatness just through personal training plans. They need a coach.
Your own personal AI health assistant that will empower you to see the future of your health and guide you to change it for the better.
Topi’s more than just a robot. Topi is your guide.
Topi solves both of the biggest hurdles in our fight against preventable chronic diseases and human nature. It’s able to create personal health plans with complex causal modeling of genomic, epigenomic, and lifestyle data, and it knows how to best deliver and communicate those plans because of its meaningful relationship with the people it serves. Topi comes with a wearable biosensor/communicator that is able to keep track of biomarkers in order for it to a) gain a better understanding of your body and b) be able to communicate with you at all times in order to enhance your personal relationship and keep you on track.
Topi’s mission is to:
- create a causal model of your genome, epigenome, and lifestyle through an understanding of the relationships between these three key factors and disease risk
- create a personal health plan from this causal model with an efficient communicator-biosensor that is able to keep track of biomarkers and lifestyle factors in order to build on Topi’s causal model and communicate with you at all times
- build a meaningful relationship with the people Topi serves by understanding personality, language, and human nature in order to be optimized at ‘nudging’ people towards their personal health plan
Our vision is to create a world where 100% of preventable chronic diseases are prevented.
How does Topi work?
Almost 32.3 million people in the United States have Type 2 Diabetes, which is a chronic diabetic condition primarily caused by lifestyle. Type 2 Diabetes is a great example of a preventable chronic disease that Topi aims to eradicate. How would Topi actually achieve this task?
The Personal Health Plan
Let’s say you’re a pre-diabetic, and you’re trying to change your lifestyle, but everything just seems too far away. You still want to eat sugar, but you don’t know what’s the minimum amount you can get away with.
Topi tries to balance your health and your comfort by creating a personal health plan that makes sure you follow the minimum steps you need to not get Type 2 Diabetes.
Genomics: Topi first needs an understanding of your gene and the causal factors of your genome. Right now, there are over 120 gene variants that are associated with Type 2 Diabetes, but the genetics of T2D is still largely not understood. In the future, given data from a whole-genome sequence, Topi will be able to a) figure out all the variants associated with T2D you have in your genome to understand the underlying risk you have for T2D and b) understand how your genes affect how you process sugar. For example, the protein glucose-6 phosphatase catalytic subunit 2 (G6PC2) regulates blood glucose levels, and a deficiency in this protein obviously means higher blood glucose. If Topi found out the DNA coding for G6PC2 (along with similar proteins) was mutated, then it would know that your body does not handle sugar well.
More generally, Topi analyzes your genome in order to a) understand the underlying, baseline risks you have of preventable chronic disease and b) understand how your body interfaces with outside factors like sugar, sodium, lack of sleep, etc.
However, Topi needs more information than just your genome sequence to be able to perform these tasks. It needs to understand your epigenome as well. Your epigenome is basically a bunch of chemical compounds that tell your genome what to do. If you think of the genome as a blueprint of your body, your epigenome are edits to that blueprint. The epigenome plays a large role in the expression of genes that affect both your underlying risk for certain diseases and how your body interfaces with outside environmental factors. Additionally, your epigenome is affected by your lifestyle and those outside environmental factors. Topi analyzes your epigenome through epigenome mapping which means analyzing DNA methylation and histone modifications.
Once Topi understands the underlying, baseline risks you have of preventable chronic disease and how your body interfaces with outside factors, it’s able to take your lifestyle data (data that is retrieved from the biosensor and conversation) and build a personal causal model of your life.
The causal model takes in the baseline risks (both genomic and epigenomic) that you have of chronic diseases, the relationships between your body and its markers (sugar consumption, sleep, etc.), and the actual factors themselves (lifestyle data) as inputs. The causal model builds relationships between these various factors and thus is able to predict the state of your health in an amount of time given that data.
For example, in the case of T2D, Topi might know that you have 67 genomic/epigenomic variants that put you at risk for T2D which might mean that you might have a 3% chance of T2D if you maintain a healthy lifestyle. Then, Topi also knows that given ur current genome and epigenome, your body does not regulate blood glucose well and thus is more sensitive to sugar. Topi might also know that your sleep affects your blood glucose as well. Given that Topi understands how what you do affects other markers in your body, it would be able to predict that your current lifestyle (eating x grams of sugar everyday, sleeping y hours, exercising z minutes, etc.) would increase your chance of T2D 10% every year and that you currently have a 33% chance for T2D.
Given these predictions, Topi also knows what kind of lifestyle would decrease your risk for T2D, and is able to generate a lifestyle recommendation that would do so. This recommendation generation involves some data processing/semantic analysis so that the recommendation is semantically meaningful to you. For example, instead of saying “Eat less than 8 grams of sugar every day” Topi might say, “Your maximum amount of sugar for today is equivalent to an apple.”
The wearable biosensor is a key component of Topi. Not only does it provide Topi most of the lifestyle data through constant biomarker monitoring that Topi needs to generate a personal health plan (sleep, blood sugar, sodium levels, etc.) it provides a constant bond of communication between you and Topi, enhancing your personal relationship.
This biosensor will be able to keep track of your health at all times, even when Topi can’t physically see you. Through the monitoring of areas like blood pressure, heart-rate, and other individual areas of interest (eg. insulin for diabetics) Topi can guide your health.
Most current biosensor technology is only able to monitor biomarkers such as pathogens, metabolites, and proteins through blood sampling; though ideally non-invasive measures are the best. Most non-invasive measures currently only involve radio-imaging or some sort of light-diffraction technique in order to measure molecules in the bloodstream. This technology needs to be innovated on and enhanced greatly in order to create a multi-purpose wearable non-invasive biosensor.
The Meaningful Relationship
Topi is a stationary device that is able to see what you’re doing in a given space. Topi utilizes this visual input to a) understand your behavior and personality and b) have meaningful conversations based on what you’re physically doing. This visual input is what Topi needs to do personality recognition.
Personality recognition is already a fairly mature technology. Current machine learning methods are able to recognize the Big-5 personality traits just through videos of eye-movement. Topi can use its understanding of personality in order to better communicate with its patients.
The relationship aspect of Topi is arguably the most important part of its technology. People don’t listen to people they don’t trust, and Topi won’t be able to nudge people towards the right decisions if they don’t trust Topi. So how can we build this personal relationship?
Humans can develop personal relationships with anything: even rocks. In current solutions, like health coaching, where the relationship aspect is a key factor in guiding the patient, insecurity and fear play large roles in lowering the effectiveness of these solutions. However, insecurity doesn’t exist when the robot is physically incapable of judging you. By consistently providing personalized, valuable advice and recommendations, Topi is able to build a meaningful personal relationship, and effectively nudge its patients towards their optimal health.
You might have two big concerns after reading the capabilities of Topi.
- You’re worried about privacy. What will happen if this data, the data that literally defines your health, get’s leaked. Not to worry: your data is held on a secure blockchain. Blockchain is a decentralized way of holding information that is pretty much un-hackable. Find out more in this article.
- You might think that Topi is too powerful. You don’t think a machine should have that much control over someone’s life. But let me ask you this: “Why not?” We already trust doctors, who are prone to error, about our health; in fact, we sometimes even trust those sketchy sites that pop up claiming that they’ve cured cancer. Topi will have all the qualifications of a doctor and more. It can’t lie because it’s a machine. It will do exactly as it’s programmed to do: to help others.
The Future with Topi
You have a friend or someone who just gets you?
Topi is that friend.
Because of Topi’s total understanding of your genes, lifestyle, and behavior, it knows what to communicate to you in the best way possible.
You’re about to eat a cookie.
You think it’ll be the best cookie you’ll ever eat.
Then Topi tells you what’s going to happen if you eat that cookie: It’s 7:PM which means your body is ready for sleep, yet eating that cookie will wake your body up again. Eating that cookie will break one of the habits you’ve been working on for weeks: trying not to eat food before bed-time. Topi knows from past experience that if you eat that cookie, you won’t be able to build the habit up again for a long time. Enough time to seriously mess up your digestion cycle and sleep schedule. Topi tells you that if you break the habit and live that reckless-midnight-snack-lifestyle, your chance of T2D will increase by 7% in the next 4 years.
So you don’t eat the cookie.
Topi gets that.
Or maybe as you’re trying to eat that cookie, Topi instead tells you that you’re on the border between being health and unhealthy. Topi knows that junk food is fine every once in a while, but it also knows that you’ve been ignoring that rule for the past two weeks. Topi tells you how if you continue down this path, the chances of you getting type 2 diabetes will increase to 80% in the next two years…. So you don’t eat the cookie.
Or maybe, you’re downright crazy.
Topi knows that.
So Topi creates a 2-minute song for you that’s based off your two favorite genres of music: Romantic-period harpsichord and heavy metal. The song completely wipes away any desire you ever had for that cookie…. So you don’t eat the cookie.
I think you get the point I’m trying to make. Topi is more than just Siri turned health-assistant. Topi is that friend. That person who just gets you.
The Path to Topi
Topi is a complex moonshot project that implements many technologies that have not been created yet or are not at the level to be efficiently adopted. This section aims to break down what technological barriers need to be overcome in order to create Topi.
Causal Modeling and Data
To be able to implement causal modeling, Topi needs a lot of data. For example, in order to make counterfactual predictions about what would happen to your health in the future, it would need lots of longitudinal data (data that track the health of people over time). Huge advancements need to be made in current causal modeling, data science, and health data accessibility in order to make Topi feasible.
I am working on a proof-of-concept that shows that causal modeling is able to generate lifestyle recommendations in the specific case of hypertension. Check out my progress here.
Topi needs a complete understanding of gene-trait associations in order to understand risks of diseases and make genetic-based health-style recommendations. The ENCODE project is working towards a full understanding of functional genomics, yet this project may take as long as the Human Genome Project, or longer.
Topi needs real artificial intelligence in order to communicate the recommendations it creates for you and build a relationship with you. In order to understand you as a person, it needs to be able to remember things you’ve done in the past as well as analyze those events. One area of interest that might be applicable here is neuromorphic computing. Neuromorphic computing is a type of computing based off the brain. Because it’s based off the brain, it possesses key capabilities of the brain, like energy efficiency, spatio-temporality (able to remember events by time and space), and latency (speed at which data is processed). Topi could be based on a neuromorphic computer that enables all these key capabilities.
All of these technologies need to be at industry-level in order for Topi to be created. And once they get there, we need to do further research in order to actually create Topi.
But we’re not waiting for these technologies to get there.
We’re actively learning and growing with these fields to reach our vision of a world where everyone can live healthy personally.
And once we get there…
Say hi to your new friend.