How I Built This with Guy Raz - HIBT Lab! Immunai: Noam Solomon

Episode Date: December 22, 2022

According to the CDC, 6 in 10 adults in the U.S. have a chronic disease like cancer or diabetes. Meanwhile, the process for discovering new drug treatments is critically expensive and inefficient. About 90% of drug candidates fail to gain FDA approval; the average cost to develop a new drug is over $1 billion; and testing can take over 10 years.Noam Solomon is on a mission to change this. His company, Immunai, is using artificial intelligence to create an atlas of the human immune system. This week on How I Built This Lab, Guy talks with Noam about how Immunai’s immune system mapping is accelerating the development of new personalized drug therapies. Plus, Noam shares how Immunai’s culture of ‘not knowing’ drives scientific innovation. See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

This podcast is brought to you by Squarespace. I talk to entrepreneurs all the time who are looking for a way to upgrade their digital footprint. Well, whether you're just starting out or you're scaling your business, Squarespace is the easiest way to build a great website that stands out. It's an all-in-one website platform that gives you everything you need to claim your domain, showcase your products, and get paid. Anyone can use Squarespace's cutting-edge design tools to build an online presence that truly reflects what makes your business special. There are templates, intuitive drag-and-drop editing, and even an AI-enhanced website builder. Then, Squarespace's built-in analytics tools help you make smarter business decisions. Review website traffic, learn where to focus engagement, and track revenue all in one place.

Looking to grow your business? Squarespace even offers fast, easy business financing through Squarespace capital. Go to Squarespace.com slash built for a free trial. And when you're ready to launch, use offer code built to save 10% off your first purchase of a website or domain. Loans issued by Celtic Bank and serviced by Stripe. All loans subject to credit approval. This show is in partnership with Airbnb. This past summer I took my family to Vienna, and it was incredible. We spent our days wandering the old streets, stopping for coffee and pastries,

visiting museums, and just soaking up the history of one of the most beautiful cities in the world. And one of the things that made the trip so special was the whole. home we booked on Airbnb. It had tall windows, beautiful old details, and plenty of space for all of us. And being in that home on Airbnb, right in the middle of Vienna, walking distance from so much of the city made it feel less like a visit and more like we were actually living there. Plus, taking a trip is the perfect time to host your space on Airbnb. Your place, with all of its personal touches and its amazing location, could make someone else's vacation even better. Your home might be worth more than you think. Find out how much at

Airbnb.ca.ca. slash host. Hello and welcome to how I built this lab. I'm Guy Raz. So if we've learned anything from our experience with COVID over the past few years, it's that we still have a lot to learn about how the human body fends off and fights disease. My guest today is on a mission to change that. In 2018, Noam Solomon and his co-founders launched a company called Immunei, and it's working to create what's essentially an atlas of the human immune system. There are thousands of clinical trials that take place each year around the world where people with certain diseases and conditions are given experimental medicine. Immunei partners with hospitals and universities to gather cellular level data from those clinical trials. The company's massive computing

power and artificial intelligence programs, then analyze that data, which can then help figure out which experimental treatments work and which ones don't. No one believes that eventually immunized technology can help pharmaceutical companies develop more effective treatments for some of the most serious diseases, including cancers. Immunei has offices in New York, Tel Aviv, and in Europe, and it's raised over $200 million for this work. Between them, immunized founders of backgrounds in computer science, advanced mathematics, and biology. Noam, the CEO, joined me from his office in Tel Aviv. So I knew I was going to be an academic even before I was 10 years old.

And the plan was to be a faculty member and doing research in math and computer science. And I have a change of plans. From what I gather, you ended up in the United States, around 2017, you came here to do some post-doc work in mathematics and computer science at Harvard. And then you went on to MIT to do more post-doc work in mathematics. And that's really kind of where your, I guess, your career path was derailed in a good way because you did not go in the direction of mathematics and computer science. Tell me what happened.

What started the idea that would eventually become immuni? So about five years before then, for my second PhD, I wanted to have some, you know, additional income. So I worked in a few startup companies as a data scientist. We developed some algorithms in different disciplines and fields it was not in biology. So e-commerce, ad tech, different types of web applications. And it was mainly to supplement the income. working as an academic, especially in Israel, it's a challenge. A few years later for my postdoc, my co-founder and CTO at the time, we met as friends.

He told me about his grandfather. At the time he had cancer, he was getting a combination of immunotherapies, and it was working for him. So the cancer was regressing, but he couldn't stand the side effects, the adverse event. so he stopped taking the medication. So eventually, a couple of years later, the grandfather died. At the time, I was very, very interested to understand if you could use data science and mathematics

to better find the right therapy for the patient. So this was the starting point for me, whether you can use some of the knowledge and background that I have to maybe do something that can impact patients. What was the opportunity you saw that you could contribute to in developing technology that could really transform medicine? Yeah, so it's a deep question. Immunotherapy is like the name are drugs that are targeting our immune system.

Our immune system is an incredibly complex system that is essentially governing our health, the way that we respond to therapeutics, the way that we cope with different bacteria and viruses. Of course, the entire world now understands after the few years that we had with COVID, how important the immune system really is. But how can you know if a patient is going to respond to an immunotherapy or not?

And maybe one patient is going to respond to a drug and another patient is not going to respond to the drug. Can you personalize the treatment? So this was the starting point for us. How can you create the knowledge base to answer this question? You've described immunized work as sort of like creating a Google Maps of the human immune system, right? More or less? Yeah.

So what does mapping the human immune system mean? Right. So the different cells in our body, we have trillions and trillions of cells, they're different. You have cells within your immune system, within the different organs. And what is important is first to understand the makeup of our cells, but also, the way that they interact with one another. So let's say somebody is being exposed to a virus or a drug. It's not going to be one cell. It's going to be many cells being impacted by this. How do you measure on a system level the way that cells interact with one another? So this is the

type of challenge that I don't think we were able to cope with five or ten years ago. But I think the increase of technologies, for example, single cell technologies allow us to measure cells individually. So we can map every single cell separately. Every cell is giving you the ability to look at 20,000 different genes. And then you can look at hundreds of different proteins. You can look at many, many different characteristics and variables. So for example, we are going to generate one terabyte of information from every sample that we take from a patient. So it's a massive amount of data that we generate and then we're trying to make sense of this data. But the reason why I called it Google Maps is because we also want to help people get from

where they are today. Sometimes it's a state or city of disease and they want to go to a state or city of Q. How do you find the way there? And this is where the mapping of the human immune system is a means to an end. And the end is to find novel therapies that can help patients get to where they're trying to get. Noam, has the sort of the absence of a map or maps of the human immune system essentially held scientists back, researchers back, from developing effective treatments for, you know, for a variety of diseases? Yes. I think today the way that medicine is being discovered and developed is based on a trial and error of scientists in the lab.

Having a map of the human immune system will enable researchers to apply data mining tools to come up with novel ways to develop and discover therapeutics. What's been the biggest challenge? I mean, why hasn't the human immune system been mapped? What's been the technological obstacle that has prevented that from happening until now? I think technology is changing very rapidly. What happened in the past five years was in some, sense, a perfect storm where compute power has grown dramatically. We have seen more and more,

and we keep seeing more and more technologies in biology, in chemistry, that allow us to measure things in unprecedented and resolution granularity. And we have more sophisticated algorithms. So I think this perfect storm of technology in science allows to do things we couldn't have even dream like 15 and 20 years ago. Yeah, right. My assumption is every single human has a somewhat different immune system. Our immune responses vary widely from human to human. And essentially, if you mapped the human immune system and everybody had access to this mapping, you could develop personalized treatments more quickly. It is true. And similarly, I think that, you know, the colors of our eyes are

slightly different, but there are categories. And I think that when you map thousands of people, of patients. And you will find a lot of similarities. So I think the similarities allow us and make us confident that we could measure the human immune system and try to understand how patients and that have breast cancer or non-smone cell lung cancer, open-creatic cancer can be clustered into groups. And those immune systems profiles will allow us to figure out what will be the right therapy for them. I believe that the world is going into a more personalized way to understand our health. And I really hope that medicine is going to go into preventive care and not only into medical care, right? I mean, when you're very tired in the past few months, maybe there is an immune

reason for that. So when we understand the immune system better and better, and we map more areas of the immune system. We will be able to cope with more rare autoimmune indications, orphan disease, and other, you know, cancer indications that only few patients have. And, you know, there are certain indications that are not getting enough funding because not enough patients have them. The vision that I have and the hope that I have is that maybe in 15 years will be able to use our immune profiling capabilities to help people remain healthy.

So, Noam, can you sort of lay out what the challenges right now are with developing immunotherapy treatments? So, first of all, drug development is becoming increasingly inefficient. It is super expensive. So to develop a new immunotherapy, it takes more than $2.6 billion. It takes more than 10 years of research and development before market. And more than 90% of drugs fail to get an FDA approval. All right. So it's a massive investment of money in something that is uncertain. So basically, I'm assuming that when pharmaceutical companies take a swing like that, they have to be pretty sure that there's a good chance it could work before they're going to put $2.5 billion into something. Right. And so essentially what you're saying is, if we can begin to map the human immune system with increasing granularity, you could start to solve some of these development and research.

challenges? So I will tell you something that a great scientist told me as we were starting. He said, we can cure every type of cancer in mice. And this really gave me a pause, and I've been thinking about this for years, understanding how to bridge the gap between the human immune response and the way that drugs are being tested in animal models is, I believe, the biggest challenge that the biopharmine and biotechnology industry should face and will face in the coming decade or two. And this is our vision. Our vision is to improve the drug development process through an engineering first mapping of the human immune system and through it the human immune response. We're going to take a quick break.

But when we come back, more from Noam Solomon on how his company Immunei is using artificial intelligence in the fight against cancer and other diseases. Stay with us. I'm Guy Raz. and you're listening to How I Built This Lab. Welcome back to How I Built This Lab. I'm Guy Raz, and my guest today is Noam Solomon, whose company is combining data science and biology to develop more effective treatments for all sorts of medical conditions.

You know, I want to ask you about the business side of this for a moment, because this is a business that you're building. You've raised almost $300 million. And that's not enough, right? I mean, I'm assuming that the research and staffing costs are incredibly high. So, I mean, how much money do you think this will take? So there are a few aspects to the business here. First is that we are already working and making money by working with our partnerships.

You know, large biofarmaceutical companies are paying us good money to help them further develop and improve the development of their drug candidates and of their drugs. They're paying for your data and your insights. They're paying us for not only for data and insights, but for generally. the data on their samples and gleaning insights that are going to improve their clinical trial design and the chance for getting their drugs approved in the clinic. So this is already something that we are doing and the nice thing about it is that our machine intelligence keeps getting better and better and better. The other aspect is the core vision and the

mission statement of the company is about building something unique and differentiated and it is super risky. So our premise is that we are going to map the human immune system, but the implications of this is going to make a real dent in the way that we discover and develop therapeutics. And, you know, it's been four years. We have had a lot of progress and successes, but the risk is still here with us.

And I think that everybody that joined Immunei and our investor that invested a lot of money, they understand that this is still a very ambitious scientific project. Yeah, I mean, and you're not the only company out there that is doing this work. Do you see the other companies as competitors or as potential collaborators as both maybe? So first, I really like competition. I think there are many competitive companies in our landscape and I think some companies are really making progress and for that I congratulate them.

and I think it's good for all of us. I think that at the end of the day, companies that are building real technology and real science will succeed, and I think that we are working with some of the more ambitious ones. We are very collaborative in our premise. We have over 30 partnerships. Some of them are with companies such as ourselves,

earlier stage companies that are building new technologies. Some of them are more mature than us. But I believe that every company that is able to, you know, bring a scientific vision that is real and make progress towards an important problem will survive. I always say, lady science has to dance with us. We have to prove our premise. We have to show that mapping the human immune system can lead to a real ability to improve patient's outcome and improve the clinical trial design. So the collaboration is part of I believe the new drug development paradigm shift.

What does it mean for the kind of insights that you're able to provide to pharmaceutical companies and companies that are developing these therapeutics? So take a drug candidate. Some biopharmar or biotechnology company invested a lot of money in this drug candidate. There was an FDA approval to start giving this drug as a clinical trial set up to the patients and you start treating patients that have serious disease. What we are able to do is measure the human immune response of the patients and try to see whether this immune response resembles something which is good

that is effective. And we do it on a way that allows you to leverage the entire power of the database and only a handful of patient samples from the clinical trial. This is kind of the secret. So you have a huge database on the back end, and then you can use this huge database to glean insights by looking at only a handful of patients in a clinical trial setup. Help me understand what you're doing that. I mean, essentially, are you collecting blood samples from patients all around the world? I mean, how do you actually collect the data?

So there are a few ways. We have many partnerships with academic centers and biopharma and biotechnology companies, and every one of our partners are sending some. samples to us. The samples can be blood samples. They can be preserved immune cells from tissues, from solid tissues, from bone marrow. And our work is to try and create the human immune profile of the patients from the different sources of tissues and blood that we are getting. And the more partnerships we have, the more biological samples that we get. And every one of the biological samples that are getting to our lab in New York is being profiled and sequenced in our lab and this data is being uploaded to our database, which is called a MECA. It is the

largest of its kind database of the immune system with a single cell resolution. So today we have a database that spans over 100,000 patients, over 500 different disease indications, thousands of studies. And we're trying to put as much clinical context and clinical metadata into the patients. And the more data that we have, the better our mapping of the human immune system is. So I understand that recently that you've been able to show sort of definitively that there are certain treatments that have improved patient outcomes in certain cases. Can you tell me about that? Yeah. So a lot of our work with our partners is to, help them with their clinical development. So they have a drug, a drug candidate in clinical trials,

and they do a lot of experiments. So they usually have multiple clinical arms. They are trying different combinations of treatments. And we were able to show quite definitively that our platform can identify the preferred combination arm that will demonstrate a superior immune response. So it happened already, and it's something it gets both us and our partners very. excited about the potential to really leverage this immune profiling capability. Okay. Help me understand how you can develop treatments and the types of treatments you might develop from what you're building now. Yeah, I'll give you a couple of examples. So first of all, when there already are immunotherapies that are being in the clinic treating patients,

and for cancer indications like melanoma and non-smous cell lung cancer, we see our effective they are. But even in the most effective setups, they don't kill 100% of patients. Right. What we do, and we have been doing for a few years now, is that we will map patients that are being treated with an immunotherapy or combination of immunotherapies and chemotherapies, and we are going to try and find separators between responders and non-responders. And the interesting thing is that when we do our immune profiling of patients, pre-imposed therapies, we see that some very subtle nuanced differences between responders and non-responders, explain or at least partially explain the reason.

So we can find that a specific immune cell type and a specific gene within this immune cell type is correlated with non-response to a therapy. So this can become a novel target. for a novel therapeutics. And this is what we are doing. So the main point is to try and say, when you have one patient that is responding to a drug and another does not respond, it doesn't say much. But if you have 100 patients that respond and 100 patients that don't respond, you can start

apply data science. And when you have the right data sets and the right technologies to measure the biology of the patients, you sometimes find very interesting. patterns and explanations. Okay. So if in fact you could map the human immune system with, you know, incredible granularity at scale, and it sounds like you're on the path to doing that, help me understand what that means for patients, for diseases, 20, 30, 40 years in the future. I mean, how does this impact how human disease is treated? I mean, realistically, could you imagine a world in even 50 years from now where all cancers are?

are treatable diseases where people don't, don't die from them? Yeah, so it's a very interesting question. And I'm going to say things not in the most accurate way, just to make sure that I can communicate a message. You take a patient with a cancer, let's say a solid tumor, and you take the biopsy of the tumor, or you dissect the tumor out, and you are going to do it in a dish

or maybe even give it to a mouse, and you will be able to cure the cancer outside the patient. You're going to cure it by finding a new antibody or a new small molecule. But what you really want to know is what will happen within the patient. And the thing is that cancer almost has its own intelligence. So there are different escape mechanisms and different ways that the cancer can use in order to survive or to grow. And the big question is, can we use this mapping of the human immune system

and use different animal models and different in vitro models, different lab models, to try and predict what would have happened to the patient? So it's a combination of measuring cancer patients that have different types of cancers and apply a lot of computer-aided simulations in order to predict what would have happened to a patient where they treated with this therapy. And this is the challenge. So the thing is that the human immune system and the cancer,

they have this sort of battle. We are trying to understand how to help the human immune system win the battle, and it's a long, long, long road ahead. We've got to take another quick break, but when we come back, more from Noam Solomon, co-founder and CEO of Immunei. I'm Guy Raz, and you're listening to How I Built This Lab. Welcome back to How I Built This Lab. I'm Guy Raz, and my guest is Noam Solomon, whose company Immunei is using artificial intelligence to map the human immune system.

So with your background in mathematics and computer science and your colleague's background in biology are essentially tackling this challenge from multiple perspectives. and obviously with an emphasis on AI and on just massive computer power. Right. So, you know, we have about 150 people that are between New York, San Francisco, Tel Aviv, Zurich and Prague

that are spanning multiple disciplines, immunology, medicine, technologies in single cell, machine learning, computation biology, software engineering, data engineering, And the secret is not to have those efforts done separately. The secret is to try to create multidisciplinary teams that are going to work together towards a common goal. And mathematicians, such as myself, are going to have to study a lot of immunology and biology

in order to collaborate with their colleagues that come from biology and immunology. And similarly, immunologists are going to study the background they need in engineering or machine learning. And I believe that this is the new type of science that is more about connecting the dots between different disciplines and finding the bridges, which makes the challenge much harder, but so much more rewarding when you can figure out a way to use data science to answer a question in a different field. And as a mathematician, somebody that was never used to answer questions in biology, the satisfaction to be able to understand. something that can explain why a certain patient may not respond well to a drug. And another patient that is very similar to this person will respond amazingly well to the drug is very fulfilling. Noam, you wrote an article in a publication that I read, and I really love the piece.

And the title was something like, When Not Knowing is a Methodology. And you say that Immunei, not Knowing is a methodology, that essentially there's much more that we don't know about the human immune system than we actually do, which I love. I love that idea, that essentially, you don't really know, but that's sort of, that's sort of the roadmap that you pursue. You kind of, you follow this idea that we don't know and let's figure it out. Right. So first of all, it's a personal methodology.

I like to ask questions. I started the company. I didn't know anything about biology, medicine. business development, and I had to ask a lot of questions, and it became part of our culture as an organization. Because the company has such a bold vision, and we had to bring experts from many different disciplines, it became part of the way that we interact with one another, that asking questions is not only allowed, but it is encouraged. And in some places, you are expected to know the answer.

I think for us, being able to ask the right questions really led us to make a lot of progress and became part of our not only research culture, but also the way that we do business. I'm always open-minded to collaborating with companies that are bringing something novel because maybe the union of the two approaches will be better. Maybe one plus one will be three. And yeah, sometimes you just don't know the answer, but yes, the right question, it will come out. I read a piece by your co-founder, Luis, and he talked about when Google developed an AI

program that could beat humans in the game Go, and the computer beat humans by using this counterintuitive move called Move 37. I don't know the game well, but I read the article. And I guess the premise of the piece is that the move was so counterintuitive, so out of the box, that a human would never come up with that move on its own. I mean, when we're talking about the AI that you're using, could it eventually, I identify sort of a Move 37-like approach that helps us cure cancer? I mean, is that what we're talking about? Yes, and I would say that what we are doing, this mapping of the human immune system,

is much more higher dimensional than the problem that you just mentioned. I'll give an example. So we are measuring from one patient. We're taking, let's say, we're drawing the blood. So we will measure 10,000 cells. For each of the cell that we are going to map, we are going to measure 10 or 20,000 different genes and proteins and other measurements. Now, we are doing this for, let's say, 1,000 patients. We are putting this information into a machine. This is not the equivalent of goal. No human being will have the intuition to analyze all the relevant data. You have to use a machine to analyze the data. And I think this is where the movement or the unique opportunity comes.

There are certain types of problems where there is no human expertise to guide your ways. And until now, we have been only using a very, very small sliver of what is available. And so to what Louis said, there are many counterintuitive things that humans would never come up when they develop or discover therapeutics. And I think a data-driven way to mine a large data set is the way to go. You know, there is this analogy between what you're doing and the work of developing autonomous vehicle technology. And with AVs, the data that they collect, right, is used to make improvements to vehicles because it's collected from vehicle to vehicle and then the vehicles are automatically improved and the systems that govern them

become improved with the increasing data. How is what you're doing similar to that? Is it the same principle, essentially? So it is similar but also more complicated. And maybe the similarity is exactly what you just said. We are going to train our models on more and more data coming from clinical trials, measuring the accuracy of our algorithms. When we are wrong, we are going to recalibrate the models and do it again and again and again and improve the accuracy. But the place where this is really different is that for autonomous vehicles, you have experts.

And every driver that can drive a car, you can ask whether an autonomous vehicle drives as good as a driver, like me and you. for drug development, even the best scientists in the world, they don't know. As we mentioned earlier, only 10% of drugs are going to succeed. And the question is, how can you build a machine that is not going to be as good as a scientist? It's going to be much, much, much better. Noam, you mentioned this idea of risk and uncertainty with what you are working on and with the business that you've built.

But it seems like where you are already four plus years in is promising and that it seems to be on the right path. Why do you think that there is still so much risk involved with what you're trying to do? First of all, I think it's a good mindset. I think that when you understand that the risk is high, you are more likely to take steps that are going to be non-conformist and non-conformal. and you're going to think about the problem in a different way. What I'm trying to do with Immunei, what I'm trying to do with our board of directors, with our future investors, with our team,

I want everybody to know that we are trying to do something extremely difficult, very unlikely to succeed. Not because internally I don't think we're going to make it. I think every year, every quarter, every month, we are making a lot of progress, I'm becoming increasingly more confident that this problem is solvable and we're going to make it. But the mindset of thinking big and taking risks is the mindset that I want to have when we are, you know, waking up in the morning, when we don't go to sleep late in the evening. And I know, at least it's my mindset and I think it's very productive.

That's Noam Solomon, co-founder and CEO of Immunei. Noam, thanks so much for being on the show. Thanks, Guy. Thanks for having me. Hey, thanks so much for listening to How I Built This Lab. Please do follow us on your podcast app so you always have the latest episode downloaded. If you want to follow us on Twitter, our account is at How I Built This, and mine is at Guy Raz. And on Instagram, I'm at guy.roz. If you want to contact the team, our email address is hibt at id.wondery.com.

This episode was produced by Sam Paulson and edited by John Isabella. Our music was composed by Rumtin Arablui. Our audio engineer was Neil Rauch. Our production team at How I Built This includes Alex Chung, Casey Herman, Chris Messini, Elaine Coates, J.C. Howard, Liz Metzger, Carla Estevez, and Carrie Thompson. Our intern is Susanna Brown. Neva Grant is our supervising editor. Beth Donovan is our executive producer. I'm Guy Raz, and you've been listening to How I Built This.