From The Genomics Revolution to The Next Frontier in Biology with Cellanome

Announcer:
Welcome to TD Cowan Insights, a space that brings leading thinkers together to share insights and ideas shaping the world around us. Join us as we converse with the top minds who are influencing our global sectors.
Dan Brennan:
Welcome to the podcast. This is Dan Brennan. Today, I'm joined by Omead Ostadan, CEO of Cellanome, and one of the most experienced leaders in the life science tools industry. Omead spent 15 years across Illumina, and its predecessor sequencing company, Solexa, helping guide the companies during the rise of next-generation sequencing and the genomics revolution that transformed biological research and clinical diagnostics.
Along the way, he worked alongside industry pioneers, including Jay Flatley and Mostafa Ronaghi, gaining a front-row seat to how breakthrough technologies became category-defining businesses. Today, he leads Cellanome, a company founded on the idea that, while genomics has taught us a tremendous amount about biology, many of the most important questions still require understanding what living cells actually do.
Cellanome has developed a platform that combines live-cell imaging, AI-driven analysis, and single-cell molecular profiling to follow cells over time and link cellular behavior with molecular state. The company refers to this as understanding the cellular choreography of biology.
But today's conversation is much more than about Cellanome. We'll discuss the state of biological innovation, whether we're underestimating the importance of biology relative to AI, the role scientific tools playing in driving discovery, lessons from building great companies, and where the next major breakthroughs in life sciences may come from.
Omead has spent nearly 30 years at the intersection of science, technology, and entrepreneurship, and I'm really excited to get his perspective on where the field is going next. So, Omead, first, welcome.
Omead Ostadan:
Well, Dan, thank you very much for having me. It is a privilege to be with you, and I'm definitely looking forward to the conversation.
Dan Brennan:
Awesome. So, maybe let's start off with the state of biology and scientific progress. You've spent much of your career during the genomics revolution. What did genomics deliver that exceeded expectations, in your opinion, and where did it ultimately fall short?
Omead Ostadan:
In so many respects, it has actually, and it continues to actually, exceed my expectations. It's hard to believe that... Gosh, it was maybe about 20 years ago. It was almost 20 years ago that we had sequenced, what, maybe three or four genomes, quite frankly, at that time, ever in the history of humanity. And now, probably in the course of the conversation that you and I are going to have, there's going to be three to 5,000 genomes sequenced somewhere.
I mean, it's just hard to imagine that in two decades we've gone where sequencing genomes is a moonshot, and now, quite frankly, it costs less to sequence a genome than it does to go and buy an iPhone. And that is a remarkable testament to the impact and power of not only innovation, but innovation that fits a gaping hole and a need, which was to really understand the molecular biology of life. And that's what genomics enabled.
Areas where it exceeded my expectations, certainly, I think from my perspective, the breadth of impact it has had not only on clinical diagnostics, as you've talked about it, but expanding it to other areas that have ultimately affected and shaped therapeutic development. If I think about things like vaccines, mRNA vaccines, COVID vaccines, all of those types of technologies, to a large extent, I think, benefited from the availability of genomics and sequencing as a tool. And more broadly speaking, if you span that out and look at just about any area of therapy development today, I would say genomics has, and will continue to play, a pretty critical role.
I think where... I wouldn't necessarily say it's fallen short, but I think what it has probably highlighted and crystallized for me over the course of, in particularly, probably the last five or six years, is that, as incredibly powerful as genomic has been and will continue to be, it only allows us to access a part of biology. Albeit a very necessary part of biology, but it is entirely insufficient to be able to allow us to understand the etiology of disease to the level that we can actually actively impact it, in ways that allow us to be able to advance human health, I think, according to our ambitions.
So, as incredibly powerful as it has been, I think what's clear to me, and this is really, in some ways, what went behind the founding of Cellanome, was that we need more. We need broader insights, we need deeper insights, we need more multimodal insights into how biology operates at its most fundamental level, which are the cells. But to be quite frank, the only reason we get to actually have the opportunity to contemplate those questions is because genomics has allowed us to be able to even ponder them and be ambitious enough to take them on.
Dan Brennan:
Well, AI is now the biggest technology driving the world and certainly the stock market, though, very early on, how will it impact R&D is far from clear. Most people assume AI models potentially are the limiting factor. I think you've argued that biology may actually be constrained by the quality of the underlying data sets. Why is that?
Omead Ostadan:
Well, so first, I will just acknowledge that I am far from an expert on AI, or biology for that matter. So, take my answer with a grain of salt, or many grains of salt for that matter. But I think it goes back to just a point of view, which I think is actually a relatively well-accepted point of view, it's not my point of view, is that these predictive models, these AI models are essentially, to some extent, can be only as good as the data on which they were based.
Because fundamentally, what are these models based on? They're based on information and categories of data that then embed a series of rules and heuristics, which then can parlay into predictive analysis moving forward. And so, the complexity and challenge in terms of really activating the full potential of AI in biology is that, for the majority of biology, we actually lack the input data that can clearly delineate the rules of how biology functions.
As complex as biology is, it's extraordinarily rules-based. Hundreds of millions of years of evolution have led us to be, actually, relatively predictive from a biology perspective. The challenge that we have is we don't understand those rules. We don't understand cause and effect and the network of actions that lead to phenotypic outcomes.
And so, for foundation models to be impactful in biology, they have to be pre-trained and ultimately trained on data sets that represent the dynamic complex nature of biology in ways that are sufficiently context-rich, supervised, structured so that you can actually get to the underpinning, if you will, rules of biology, which can then embed the heuristics of your model, and thereafter you can use that model for predictive analysis.
Unfortunately, those data sets, by and large, don't exist in biology. We have large volumes of data in biology. They tend to be static snapshots of either genomes, single-cell transcriptomes... massively useful, but insufficient in terms of actually what we need to be able to do to predict the dynamic nature of biology.
And that's why I think the rate-limiting step in activating AI in biology is the data, because the compute power is there, the ability to generate the tools are there. What's missing, to a large extent, is data.
Dan Brennan:
That's the $64,000 question facing the tools industry today. A lot of investors grapple with this... is, it could really be exciting if there's a lot more data that needs to be generated, because before computers begin to remove steps in the process, there needs to be a lot more spending and wet lab work.
Is there any way to frame, just in any context, about... are we five years away from generating this data, two years, 10 years? It's obviously going to be a constant evolution, but I'm wondering how much more data do you think needs to be generated, and would that be considered a good thing for the demand trends for a lot of the life science tools industry, do you think?
Omead Ostadan:
In a lot of ways, I think of biology as the ultimate big data problem. I mean, if you think about the variables that are involved, the diversity of humanity, the inputs that affect biology, there are just so many variables and just so much complexity that needs to be unraveled. And to be able to get to that, I can't possibly even imagine the volume of data that's needed.
So, think about it this way. If you think about how much data even... I don't mean to diminish them, but LLMs have to be trained on to be able to tell you what the next word in a sentence ought to be, or how to search for a particular match of an image that you're looking for, that volume is already just extraordinary, well-structured, easily accessible. We don't have that in biology, and biology is a heck of a lot more complicated than that.
So, my view, to be quite frank, is that I think the volume of data we need to be able to maximize the impact of AI in biology, will end up being one of the largest data generation, data-gathering endeavors ever undertaken by humanity, period. I really believe that. And I think it's a sort of thing that is going to span, certainly, years, if not decades.
And a lot of it is just... it's not just about... If I take a step back, you need data to first be able to predict the basic principles. Then, you also need data that allows you to monitor, say, patient behavior to particular therapies over the course of time. And this is one of the things that I think is very tricky about biology, is that the evolution of cellular behavior is continuous, and it gets affected by so many other variables in its environment that you need to almost... it's a constant sampling in order to understand what the trajectory of, say, a person's health journey will be. And you need that data to then feed back into your models over the course of time.
So, I think this is one of those things that is going to span decades. I do think it is one of those things that, and you talked about it, we're going to see it happening continuously. I don't think this is one of those things where it's going to go five years and then, suddenly, voila, there is some magic AI model. And we're already seeing it. We're already seeing it in particular areas, some very, very powerful, task-specific models in biology that have been trained on highly supervised structured data. Now, you apply them to a relatively narrow swath of biology, but boy oh boy, they're powerful. And so, they're definitely, for me, existence proof that, with the right data set, you can actually leverage AI in biology in ways that, quite frankly, humanity can't access any other way.
And so, as data gets accumulated, I got to believe that we're going to see more and more, both application and adoption of AI. I actually think it's likely to happen more in task-specific instances, just because the problem set is a little bit more contained, and then, over the course of time, more generalizable model that allow you to ask slightly broader questions in biology, and be able to get predictions.
I think, on the whole, this is extraordinarily helpful for tools. I actually think this is going to end up being one of the biggest driving forces economically for life sciences tools companies that has ever existed, to be honest. I think the power AI is so enormous that we will find ways to generate the data to be able to take advantage of that capability to advance human health, because I don't think we can get there any other way fast enough.
Dan Brennan:
So, a lot of what we've already discussed in this podcast points to the importance of better biological measurements, more data, more insightful biology. So, what was the key insight that led you to believe Cellanome needed to exist in that context?
Omead Ostadan:
Well, so Cellanome was founded in middle of 2020 by Mostafa Ronaghi. And Mostafa is an ex-CTO of Illumina, and I was fortunate enough to span 12 of my 15 years at Illumina with Mostafa, and just an enormous amount of respect just for everything, quite frankly, that he does.
And the realization that he and a group of five other co-founders came up with, and we've talked about it, was that, as incredibly enabling as genomics has been, that we needed to know more. We needed to know and be able to understand and measure more at the level of living biology.
And the insight was, look, cells are like the fundamental unit of life. They're a wholly contained, fully functioning unit of life. And so much of biology really is transmitted, transacted by cells and cellular communication and interaction. And if we could find a way to be able to interrogate large bodies of cells, individual single cells, and, more importantly, co-cultures of cells, and ideally in the context of primary cells, because what you're trying to do is get as close to biology inside the human body as humanly possible.
And so, what that requires is the ability to be able to survey diverse mixtures of cells, including adherent and suspension cells, dynamically over the course of time. And this goes to the point that cellular behavior varies moment by moment, and to ideally be able to interrogate those using multimodal measurements... so discreet, independent measurements that essentially allow you to capture different behaviors, aspects of the cell, and then aggregate all of that information to be able to understand basically how does the cellular molecular biology function.
That idea wasn't necessarily the novel insight. People have known we want to do this for decades, if not centuries, quite frankly. The challenge has been, how do you do it technologically? And that's where I think a number of the prior efforts have, quite frankly, hit a roadblock. Keeping diverse mixtures of cells alive for extended periods of time and then be able to interrogate them multimodally is quite challenging.
And that's where I think the breakthrough idea for Cellanome was, was to come up with a technology approach that allowed us to physically constrain cells in biocompatible chambers and environments, and, in the process, be able to then bring the experiment to the cells. So the cells, after they're sort of physically constrained, no longer move, but you can essentially be able to bring media, perturbations, all sorts of antibodies, and so you can sequentially sample and survey them over the course of time.
So, the breakthrough ideal was the technology aspect that allowed us to be able to then apply that concept to living cells. And we've been fortunate enough to actually realize that ambition, and commercialize a product earlier this year that does exactly that for a broad range of cell biology experiments.
Dan Brennan:
If you had to choose one biological question researchers can answer today using Cellanome, or in the near future using Cellanome, that would've been extremely difficult or impossible just a few years ago, what would you go to?
Omead Ostadan:
Well, it's a tough one. It's like asking me to choose which of my kids is my favorite, and you sort of can't do that. There's just so many of these applications and ideas.
But one that probably, in some ways, encapsulates, I think, the capabilities of Cellanome is the ability to be able to actually interrogate and understand the interaction between neurons and microglia. Neurons are adherent cells. We can actually create very complex, active neuronal networks on the surface of our flow cell, comprising tens of thousands of fully functioning neurons. The process onto itself actually takes a couple of weeks, because you have to actually grow and differentiate these neurons so that they are active and functioning.
And then you can basically introduce microglia, which are suspension cells. And now you can begin to actually interrogate the interaction. You can interrogate them in, if you will, kind of a native condition. You can interrogate them with CRISPR perturbations, you can interrogate them with combinations of drugs and compounds and/or CRISPR perturbations. The range of flexibility you have within our system is nearly endless. You can almost think of the flow cell chamber as an open canvas for experimental design. It creates essentially an open dimension that allows you to basically compartmentalize different combinations of cell-cell interactions, and be able to localize those interactions and study them.
And no matter how homogeneous you think a population of cells are, they're never exactly the same. And so, one of the great advantages of our technology is that it allows you to see the distribution of cellular interaction and response across time and across different modalities of measurement. And that begins to give us insights that are otherwise not possible.
So, this seems like a relatively straightforward, simple idea, but it is an experiment that literally cannot be done today by any other platform. You just physically can't do it. And if you then think about the importance of understanding the fundamentals in not only neurobiology, but its interaction with the immune system, if you think about the range of diseases that we want to better understand and address, whether it's Alzheimer's or Parkinson's or you name it, this is an avenue through which we think people are going to be able to gain the type of insights that we believe, in time, are going to accelerate understanding of the biology and, in due course, our ability to be able to affect it.
Dan Brennan:
Excellent. So, you worked closely with Jay Flatley during Illumina's rise. I wanted to ask a question on leadership. So, what was the most important leadership lesson you learned from Jay that still influences how you run companies today?
Omead Ostadan:
Okay, I'm going to cheat on that one. There's no way I can reduce that to just one answer, Dan. I literally could write pages and pages of things I've learned from Jay, and I continue to learn from him.
First thing I'll start with is there's a reason Jay is just, I would say, very broadly considered one of the most influential, inspirational, impactful leaders in life sciences, or just say business, period. And I think one of the things that has always impressed me about Jay, always impressed me about Jay, he is about as authentic of a human being as you can come across. With Jay, what you see is what you get. His passion, his energy, his empathy, his commitment, it's just off the charts. And that creates a level of, quite frankly, loyalty and trust and admiration that motivates you to always want to do your best in a way. Even though I don't necessarily... it's not in the forefront of my mind, it's like one of those things that always... I was like, I never wanted to disappoint Jay, because I just think the world of him.
But I'd say a couple of things, just learning from Jay, and actually more broadly from... There was a period of probably about a decade where I think the people at Illumina were just about as good as you can get. So, when I talk about learning lessons from Jay, it was Jay and a number of other really exceptional colleagues that I was fortunate enough to work with.
But one was just the power of teams. Put together the best team you can, and put them in a position to be their best versions. Drive professional debate, encourage debate, listen to different points of view, beat the heck out of ideas, because as you sort of polish these ideas, it's like every once in a while a diamond does emerge, but you got to put in the time. And the best way to do it is with really good people who feel empowered and supported to express their points of view, no matter how contrarian they may be. And then out of that, you can come up with ideas that can be exceptional.
The second is never settle for incremental. Definitely not in our industry. Incremental sometimes is safe and comfortable, but in our industry, it's rarely the right answer. Bold is risky. Bold, by definition, means you're going to have failures, but bold is also just about the only way you get to breakthrough ideas that are going to have transformative impact. And so, be humble enough to learn from your mistakes, be quick enough to adapt and adjust. And out of every failure, you can always build success, but bold, because failure in our industry, quite frankly, is not an option. The consequences is pretty significant.
And the third one is... And this is just... it's less of what Jay said, but just the way he lived it, he so fully believed passionately in the mission of what Illumina was out to do. And it was so palpable that he was so driven by just, I think, a desire to positively impact human life that imbued everything he did. And so, for me, it was a privilege to work with him, and I always saw this as just an incredible opportunity that I had to be in this industry, but with that opportunity came the responsibility, knowing that if we did our best, that on the other side of it, somebody could live better, tens of thousands of people could live better. That purposeful work, and making sure that that was always the North Star of why you showed up to work, why you solved the hard problems, was one of the things that I always found to be extraordinarily inspirational. And Jay lived it in everything he did and everything he said.
So, those are some of the key takeaways that, to this day, quite frankly, form how I think about operating within Cellanome.
Dan Brennan:
That's great. Well, I think we should close the podcast out by maybe zooming out here with a question on the future, which we've talked a lot about, but I'll ask you potentially to choose amongst your children again here, Omead. So, if we have this conversation again in 10 years from now, what scientific breakthrough do you hope we're talking about?
Omead Ostadan:
Again, it can't be a short list, but I'll pick two, Dan. So, I'll be slightly disciplined. I mean, I would love to be able to look back and say that collectively, and not just Cellanome, just collectively as an industry, we have managed to understand the causative points of most neurological diseases, and that if we're not able to affect and treat them, that we're pretty gosh darn close or underway.
And then the other is, quite frankly, cancer. As much progress as we've made, and there's just... even with some of the stuff that's coming up in the last couple of weeks, I literally can almost feel and taste it. Like, boy oh boy, we're making such enormous progress. And I would love to be able to look back 10 years from now and say that we're in a state where the majority of cancers can be detected early enough that they can be treated as chronic diseases, as opposed to, in many instances, death sentences for people.
I think if we can look back in 10 years from now and we can put a check next to those two boxes, and if I and Cellanome have in any way been able to contribute, however small, to that, I'm going to look back and feel like it was a privilege to be a part of it all.
Dan Brennan:
Well, it's been a privilege to be able to host you on this podcast and get to know you throughout the process. And obviously, look forward to getting to know you a lot more as we follow Cellanome's success. Omead, thank you very much for being with me and us on the podcast.
Omead Ostadan:
Well, Dan, once again, thank you for having me on, and I love listening to these podcasts personally, quite frankly. And so, for me, it was really an honor to be your guest today, so thank you very much.
Announcer:
Thanks for joining us. Stay tuned for the next episode of TD Cowen Insights.