theCUBE + NYSE Wired: MedTech Unplugged, the Future of AI in Healthcare & Life Sciences | Thomas Laur, DNAnexus

Clips
More from theCUBE + NYSE Wired: MedTech Unplugged, the Future of AI in Healthcare & Life Sciences

Thomas Laur

CEO

DNAnexus

play_circle_outline Accelerating Health Informatics: How DNAnexus and AI are Transforming Omics Data in the Cloud for Scientific Advancements

play_circle_outline Evolution from DNA data to multiomics in personalized medicine and drug discovery.

play_circle_outline Challenges of organizing dense omics data across different formats and schemas.

play_circle_outline Discussion on collaboration between biotech, AI firms, and pharma for drug discovery.

Info
Transcript

Thomas Laur, DNAnexus

Thomas Laur

CEO DNAnexus

In this interview from the theCUBE + NYSE Wired: MedTech Unplugged: The Future of AI in Healthcare & Life Sciences event, Thomas Laur, chief executive officer of DNAnexus, joins theCUBE’s John Furrier to explain why “omics” is becoming a core data object in modern care – and why the hard part is not novelty, but execution. Laur describes DNAnexus as a secure, compliant cloud platform built for multiomics at scale, where organizations can ingest multimodal data, normalize it, collaborate with controlled access and audit trails, and run increasingly sophisticat... Read more

explore Keep Exploring

What is the purpose and functionality of DNAnexus as a health informatics company? add

What inspired the initial creation of the company and how has its focus evolved over time? add

What challenges are associated with managing and organizing omics data at scale? add

What is the definition of precision health or precision medicine as described by DNAnexus? add

bolt Powered by CUBE AI

Thomas Laur, DNAnexus

search

John Furrier

>> Hello, I'm John Furrier with theCUBE. Welcome back to our program. This is our healthcare series here at the NYSE as part of theCUBE's NYC Wired program, a CUBE original where we feature the leaders in technology and capital markets who are bringing technology to the market and changing our lives. Thomas Laur's here. He's the CEO of DNAnexus, pioneering some great work. Welcome to theCUBE here at our NYSE studio.

Thomas Laur

>> Thank you, John. Thanks for having me.

John Furrier

>> Love the name DNAnexus. We are at a nexus right now. We are at a transition point in society and technology as we see super computing really be democratized. What used to cost hundreds of millions of dollars, now you can get for a million and the prices are dropping. GPUs, all this infrastructure is intelligent. It's really creating a lot of change in areas that were hard to do before, healthcare, medical technology. Explain what you guys do, because I think you guys are hitting on this key AI shift where the end user value and the scientific breakthroughs are kind of happening.

Thomas Laur

>> Yep, that's right, John. So look, DNAnexus was established 17 years ago by a talented team out of Stanford University. And the company has established itself now as one of the leading health informatics companies on the cloud purpose built for omics. We serve hundreds of customers around the world. And what we do for them is we help them ingest very large quantum of multimodal data with an omics component. The data gets prepared for analysis on the platform. And then we have a workbench in the platform where you can essentially expose the data to a very large number of algorithms, workflow engines, our tools, third party tools. You can bring your tool as well into the platform to derive insight. And as such, we support a large number of use cases with our customers. You can also collaborate on the platform. We have a system of controlled access, governance, audit trailed in order to support safely and securely collaboration initiatives between companies or inside a company between different therapeutic areas. The way you can think about us is essentially we are a safe, highly secure, and compliant across most jurisdiction platform for you to derive insight from multimodal data where omics is in play. And we're seeing on our platform customer data increasingly being exposed to more and more sophisticated algorithms. And so we're really excited to see the breakthroughs happening on the platform.

John Furrier

>> You guys really are an example of where the innovation is right now. And hearing you talk about the company, my brain kicks in, oh, it sounds like it's a data science company, data engineering, machine learning. It sounds like a tech company, obviously you are, but you're in an area where, go back 15, 20 years, the state-of-the-art technology was different. Can you share or scope the change that's happened in your area where these breakthroughs and some of this innovation? I mean, you're talking about collaboration, that's development on data, secure access, sovereignty across potential geographic boundaries and/or regulations. These are not easy things to solve. So just scope the order of magnitude of where we were 15 years ago and today.

Thomas Laur

>> Yeah, absolutely. Look, the company was initially started on the inspiration that molecular data would increasingly be present in bioinformatics. And as such, that reality, that trend would render compute and storage capabilities on premise nonviable over time. That was the right calculus. And the company grew like this on the back of next generation sequencing. The first 10 years were really about the DNA, bringing DNA data, genomic data per se, around research use cases. So early drug discovery, research use cases within health systems. What we're seeing over the past few years is really the emergence of this omics genomic, but increasingly beyond DNA, truly multiomics, genome, exome, transcriptome, protein, metabolomics, coming together increasingly out of research into production use cases. So you're seeing increasingly that omic data object becoming a standard object in drug discovery, increasingly being required at the point of regulatory approval, and of course, at the point of care. That is a reality that is challenging typical IT stacks. There's a lagging of readiness at the infrastructure level because the omics data is obviously very dense. When you egress it, it's a big job. The data is not organized around one data schema. Omics data is a variety of different formats, so it's very difficult to organize this data at a large scale. We are enabling all of that. We're taking on that complexity of ingesting enormous quantums of data. We have today over 130 petabytes of omics data on the platform, by far the largest representative of genomic data in the world. We help our customer normalize it around a standard data schema so they can enrich it with other data assets and create larger data sets on which they can run their informatics. And we help them as well deploy scientific frontier language models or their own AI models for whatever use cases they want to use. The customer we serve, just to clarify a little bit, is we serve pharma customers. So we do the drug discovery and drug development, pharma and biobanks. We help diagnostic companies, molecular sequencing companies. We also assist health systems at the point of care. Population health, so we operate large biobanks, as well as regulators, and the FDA being an important customer for us. So we have a view from the platform and our customers, we can see essentially the entire continuum of care from drug to care to regulation, really embracing this notion of personalized therapeutics, personalized therapies, not just anchored in your pheno profile, but also your omics profile, which is very exciting.

John Furrier

>> I want to dig into some of the workflows because it's a platform, you have the workflows. There's a tech angle there. I want to put a pin in that. But first explain for the folks that don't know what omics is because you are at a nexus, and explain the nuances of omics as a category, what's involved. Obviously you got DNA, you mentioned some of those things. What is omics? Because I think this speaks to the whole multimodal data, which means there's a boatload of data of different types that you're synthesizing in the platform. So explain omics real quick and then I want to get into some of the workflows and some of the data engineering.

Thomas Laur

>> Absolutely. Well, look, historically, healthcare has been organized around the symptom that you expressed visibly that can be measured by the eye and essentially your pheno data, your clinical data. And the paradigm of care is based on a fairly unsophisticated statistical arbitrage against phenotypic data. With the Human Genome Project, over the past 15 years, we've increasingly improved our understanding of the biology of life at a cellular level. The first decade was focused on DNA data, and now we're moving really from DNA data towards really gene expression. So we're going deeper and deeper into the biology of life at a molecular data. The future of healthcare and all of this precision medicine discipline will be linked to combining pheno data with omics data at a scale that allows you to really run sophisticated AI model. And we'll shift over time from a paradigm of healthcare which is reactive, where the drug calibration is based on trial and error, when you try a drug, we see if it works and then you move on to the next drug, to a standard which will be far more predictive where preemptive interventions will be possible and where the paradigm of therapeutic calibration is not going to be trial and error, but we'll be able to kind of identify the right drug for the right patient at the right time. So omics is, it's not just a plus one element into the recipe of healthcare, it's a complete revolution of how we think about care historically.

John Furrier

>> Again, back at the nexus here, the applications are multifold and it's hard to imagine beyond just medicine, people think, okay, personalized medicine, all those benefits you're mentioning. Take us through what the impact is because when you think about healthcare, we hear about proactive, preventative. So there's certainly the medicine side of it, that's key, diagnostic disease, understanding drug resistance, things of that nature. But there's also other factors like biodiversity, what's in the environment, the food supply, things that are contributing. There's a lot of cross-connect. And I think you guys are at that nexus. Explain where the application, some of the breakthroughs. Does it impact these other areas? Is it primarily just medicine and pharma? Explain the vision because I can see this translating with the data you have into resilience around food supply.

Thomas Laur

>> Yep, absolutely. And so the first point is the way we define precision health or precision medicine at DNAnexus is the development of new drugs, the approval of these drugs, personalized drugs, and then the delivery of the drugs. So it's not just drug discovery and clinical trials. It goes much further than that. The platform really, when you think about what we're enabling for our customers, it's obviously focused on speed, accelerate development of these drugs, shrink the cycle for regulatory approval, and then provide tools to assist physicians at the point of care. To your observation, a lot of people think about precision medicine and precision health in narrow terms, in terms of just clinical development of drugs, but it goes much further. Think about just the steps that you need to execute to get your drug approved. If it's one compound, it's one thing. When you talk about personalized drugs, every drug is a bit different. So the standard processes of regulatory approval and bringing all of these omics enabled evidences challenge standard workflows for regulatory approval. And so we're working at DNAnexus to try to solve for that problem. Think about drug delivery at the point of care, bedside between the physician. Take oncology. Today, when you look at the workflow, the principal workflow of care delivery around the EHR, EHR as we know them today are not equipped really to support physicians to order the right molecular test at the right time. They're not really equipped to help physicians make sense of the interpretation of whatever comes back and then calibrate what drugs to use for that individual patient. So what we do goes way beyond just the notion of drug discovery and powering this next generation of AI native biotechs. It is truly a industry-wide exercise to really help all of the different steps of the industry adjust to this new reality of omics being a standard object in care.

John Furrier

>> It's an industry-wide opportunity, but it's a global biological advancement, I mean, to oversimplify it. I mean, that's what this is being enabled. So there's so many pieces to unpack. So I want to just zoom out for a second. If you were asked by a friend or someone who smart, may not be in the industry in the level you are, obviously you guys are pioneering many new things, how would you explain the impact of this AI revolution? I mean, we saw the technology wave from print to the internet to mobile cloud, supercomputers kind of been around since the '80s, faster workstations. Now we're in this AI revolution where it can be data AI to autonomous, programmatical, situational analysis where you can say, "Hey, John, you live in New York City and California, this is your health, here's your diet." You could literally prescribe a life plan.

Thomas Laur

>> Yeah. So you made the point that we're seeing the multi-modality of healthcare related data getting richer and richer. You're describing sensor data, environmental context data. That's certainly coming into the fold. The reality here first, let me just add a caveat, which is that in order for AI to fully unleash its potential in the context of healthcare and life sciences, we're going to need far greater amounts of organized data that is omics enabled with those new modalities. There is a little bit of a deficit of scientific omics enabled data today. The phenotypic data gets organized and we have large quantums. When you overlay omics, the samples get far more narrow. So we still need as an industry to continue to platform omics and organize it. That will come with time. And it is our hypothesis at DNAnexus that based on the amount of data that we platform, we see kind of this platforming happening fairly fast. But we still need more of that omics enabled dataset. It's still a little bit of a gating factor today.

John Furrier

>> Well, it's early on. I mean, it's first inning, as they say. But in talking to the leaders at Nvidia, Dell Technologies, the folks that are making all the hardware to make these AI factories work, you're seeing two things happening that I want to get your reaction on. One is simulation and synthetic data. And you mentioned earlier about collaboration. Another signal we're seeing in our coverage is with open source and academic AI models coming together, there's also a community of practitioners that are coming in from research with kind of this open model of collaboration. So you got simulation at scale of synthetic data in the absence of more data, but then you have this collaborative piece to it. So what's your reaction to that? Is that in line with the innovation equation? What's that turn into? Obviously more real data is better, but how does that move the needle?

Thomas Laur

>> So I would agree with your premise that the advancement of AI, and first of all, we need to recognize that healthcare and life sciences historically have been lagging technology progress. And in the context of this AI revolution, we're seeing healthcare and life sciences being really the first mile, first adopter. So that's incredibly exciting for us. I also agree with you that we're seeing hyperscalers, we're seeing pure AI native companies, as well as biotech increasingly focusing on drug discovery together through partnerships, which is wonderful. I'm just back from JP Morgan where I had a chance to speak on a panel at the Bain AI in Healthcare Summit and talking to some of these innovators at the apex of AI and genomic and omics, I mean, the feeling is that we are truly close to a breakthrough, a leapfrogging in the understanding of the biology of life. When you see these young companies being able to essentially create mathematical model to map out the biology as a closed system and then create digital point and synthetic data to then kind of run this as synthetic clinical trials, I mean, this is absolutely groundbreaking and we're going to see it's a new wave of innovation. The work being done at a cellular level with those increasingly sophisticated scientific oriented models, it's beyond imagination. I had the privilege to listen to a founder who's looking at how you leveraging AI and a lot of omics data, how can you slow down aging at a cellular level? How could you reverse aging on certain types of cells? We're seeing people working on preventive vaccines for a certain type of cancer. I mean, whether it's next year, in three years, five years, hopefully not 10 years, but we are going to see I would say over the next five years, a tremendous amount of leapfrog for preemptive therapies for complex diseases, and this will change the game.

John Furrier

>> Yeah. And I think that's going to really provide great value from these factories and these supercomputers. The JP Morgan conference, healthcare conference in San Francisco, we did hear a lot about the playbook or the blueprints for preventative. So share your thoughts on that. And also, we heard that you were talking about N plus one. So I want you to explain that. So first, what's the blueprint for preventative? And then describe and define what N plus one means.

Thomas Laur

>> So the blueprint is still being defined a little bit, so I'll give you a directional answer. Preventative healthcare will happen once we're able... We need to do more work on organizing data at a larger quantum with new multi-modality. We need to see continued progress in terms of developing what I call scientific frontier models, which can really execute very specific scientific tasks. We're going to see progress in terms of the mapping of biology as a system for additional twins, which will help us understand preemptively when a trial is going to be successful or not, and we'll be able to significantly accelerate clinical trials hopefully. There's a whole context around it around regulations, regulatory approval that needs to be revised and accelerate. So there's still a lot of work, but the recipe is more data and then you need an infrastructure to be able to do the work. You're talking about a high compute environment, a lot of storage as well. You need those environments to be flexible as well because we're seeing pharma today truly retool their data and their IT estate for the AI era. So everybody is creating it their own way. There are some convergence. But when you talk about truly precision medicine as a discipline with omics, it's a completely different ballgame. You need a different layer of infrastructure. So we're having at DNAnexus a lot of conversations with large pharma. We happen to serve about 15 of the top 20. And we're trying to guide them on, okay, what do you need to build yourself versus what you should procure, because it's very complicated. It took us 17 years to get there and we know the obstacles and challenges. And so we are in an era where while data is starting to accumulate almost organically and accelerate, there's a little bit of a rightsizing of the IT infrastructure that's still ongoing that needs to happen. And then you're right, the theme of collaboration is critical. You need one platform, one environment where you can shrink the distance between patient, trials, regulators, and I think that's happening, frankly. We're seeing-

John Furrier

>> And you guys are doing a great job there. Congratulations on your platform. Final question. Explain the N plus one revolution that you talk about.

Thomas Laur

>> Well, the N plus one revolution is it's healthcare at a hyper-personalized level. It's the future of healthcare that we're trying to contribute to. Hopefully we'll be successful in making it a reality for all very soon. It's the ability to understand your own biology at a cellular level and change the paradigm of healthcare. Organize a set of prediction models, disease predictive models around each individual, being able to tailor and apply the right therapeutics for the patient so there's no trial. We gain time. And for me, the N plus one is we're all fighting. Oncology is the big villain. We want to try to basically find a solution for cancers. And I think the N plus one is the bridge that will get us there. The biology remains very complicated. Deep biology is the deeper we go, the more we realize that we still have a ton to learn. And oncology is a, cancer is a tough fight. But the N of one hopefully will give us a bridge in I hope within a decade towards being able to start winning some serious battles against cancer.

John Furrier

>> Well, Thomas, very inspirational. We appreciate what you do. And again, this is the beginning of a revolution, a sea change. The transformation has happened. The supercomputing's getting faster and cheaper every day, so we're going to see more breakthroughs. Put a plug in for the company, where you guys are at. What are you looking to do? What's your optimization plan? What's your focus? How big are you guys? Put a plug in.

Thomas Laur

>> Yeah. Well, look, we're growing. I mean, we are industrial, like hundreds of customers around the world. Just to give you a little bit of a couple of statistics, we have tens of thousands of bioinformaticians and health informatics professionals connecting on the platform. Platform has 50 billion patient files, 55 million hours of compute. So we are truly industrial and operating at a scale. And also in terms of compliance and security, I think unmatched. We serve the largest institutions in the sector. We're increasingly really the backbone and the environment on which precision health has been stood up as a production standard of care. And look, we're just going to continue. We're learning along the way with our customers. We take on big challenges with them. We're moving with humility. We know that we have to be able to be flexible as a platform. There's no need to be opinionated in a world where language models will be decentralized. But we're working very, very hard to try to help our customers do incredible things.

John Furrier

>> Well, it's super inspirational. Again, we really appreciate what you're doing. Thank you for spending the time on our AI Healthcare Leaders series. Really appreciate your time.

Thomas Laur

>> Thanks, John. Take care.

John Furrier

>> Okay. This is theCUBE here at our NYSC CUBE studio, NYC Wired program, a CUBE original. I'm John Furrier, host of theCUBE. We're featuring the leaders in AI and healthcare and medicine and medical tech. As the technology gets bigger, better, and faster, the breakthroughs will come in. We're in an era now we're going to see complete change, we haven't even fathomed use cases before, and hopefully make our lives safer and more productive. Thanks for watching.