At the Cutting Edge of Computational Precision Medicine, with Rafael Rosengarten
Genialis, led by CEO Rafael Rosengarten, is one of the companies working toward a future where there are no more one-size-fits-all drugs—where, instead, every patient gets matched with the best drug for them based on their disease subtype, as measured by gene-sequence and gene-expression data. Analyzing that data—what Rosengarten calls "computational precision medicine"—is already helping drug developers identify the patients who are most likely to respond to experimental medicines. Not long from now, the same technology could help doctors diagnose patients in the clinic, and/or feed back into drug discovery by providing more biological targets for biopharma companies to hit.
"Our commitment to biomarker-driven drug development is very principled," Rosengarten tells Harry. "There are some amazing drugs out there that, when they work, work miracles. But they don't work that often. Some work in maybe 15 percent of the patients or 20 percent. If you could tell which of those patients are going to respond, then at least the ones who aren't can seek other options, and we would know that we've got to develop [new] drugs for the others."
Please rate and review The Harry Glorikian Show on Apple Podcasts! Here's how to do that from an iPhone, iPad, or iPod touch:
1. Open the Podcasts app on your iPhone, iPad, or Mac.
2. Navigate to The Harry Glorikian Show podcast. You can find it by searching for it or selecting it from your library. Just note that you'll have to go to the series page which shows all the episodes, not just the page for a single episode.
3. Scroll down to find the subhead titled "Ratings & Reviews."
4. Under one of the highlighted reviews, select "Write a Review."
5. Next, select a star rating at the top — you have the option of choosing between one and five stars.
6. Using the text box at the top, write a title for your review. Then, in the lower text box, write your review. Your review can be up to 300 words long.
7. Once you've finished, select "Send" or "Save" in the top-right corner.
8. If you've never left a podcast review before, enter a nickname. Your nickname will be displayed next to any reviews you leave from here on out.
9. After selecting a nickname, tap OK. Your review may not be immediately visible.
That's it! Thanks so much.
Harry Glorikian: Hello. I’m Harry Glorikian. Welcome to The Harry Glorikian Show, the interview podcast that explores how technology is changing everything we know about healthcare.
Artificial intelligence. Big data. Predictive analytics. In fields like these, breakthroughs are happening way faster than most people realize.
If you want to be proactive about your own health and the health of your loved ones, you’ll need to learn everything you can about how medicine is changing and how you can take advantage of all the new options.
Explaining this approaching world is the mission of my new book, The Future You. And it’s also our theme here on the show, where we bring you conversations with the innovators, caregivers, and patient advocates who are transforming the healthcare system and working to push it in positive directions.
For most people, the genomics revolution still feels pretty distant, like something that’s happening off in the ivory towers of big pharma companies or research universities.
But say, heaven forbid, you get diagnosed with cancer next week. All of a sudden you’re going to want to get very familiar with your own genome.
Because thanks to the Human Genome Project and all the new tools for sequencing and analyzing genes, we know today that there are many different forms of cancer.
And each one may respond to a different type of medicine.
So before you and your doctor can decide which medicines will work best for you, you really need to know which genes and mutations you carry and how they’re expressed in your cells.
Drug companies need similar data when they’re testing new drugs. Because if they happen to test a drug on a population of people who happen to have the wrong genes to respond to that drug, they could wind up throwing away a medicine that would work perfectly well on people who have the right genes.
The problem is that all of this gene sequencing and expression testing generates incredible amounts of data. And doctors and hospitals and even big pharma companies aren’t always set up to understand or analyze that data.
My guest this week is the CEO of a company that’s helping with that problem. His name is Rafael Rosengarten. And his company Genialis has built a software platform that organizes and analyzes data from high-throughput gene sequencing and RNA expression assays.
We’ll talk more about what all those terms mean. But what you need to know is that Genialis is one of the companies on the cutting edge of translating genetic data into actionable predictions. Those predictions are already helping biotech and pharma companies get drugs to market faster. And in the near future they could help doctors funnel patients toward the right treatments.
I wrote a whole chapter on this stuff for my new book, The Future You. So it was really fun to talk it through all of it with Rafael. Here’s our conversation.
Harry Glorikian: Rafael, welcome to the show.
Rafael Rosengarten: Thanks for having me, Harry.
Harry Glorikian: For those listeners that don't have backgrounds in, say, computational biology or drug development, could you define a few terms that are probably going to come up later in our discussion? I mean, first, you know, maybe define next-generation sequencing or this term we call NGS. What is next-generation about?
Rafael Rosengarten: Sure, I'd be happy to do that, let me start by just kind of saying what Genialis is with some jargon in the words, and then I'll define the jargon for you. Okay. So Genialis is computational precision medicine. So what that means is we're really interested in matching patients to therapies, right? And we use data about the molecular biology of patients' diseases to do that. And our favorite kind of data to work with come from next generation sequencing. So next generation sequencing, often abbreviated as NGS, although we've been doing that for 15 years now, we probably just need to call it this-generation sequencing, is a technology where you can get the genetic information of the entire, say, genome or the transcriptome, that's the expression [for] which genes are expressed, and you get literally every base pair off of a machine that reads the DNA or RNA from cells in our body. And with that information, you do some fancy computation that, frankly, a lot of that's now fairly commoditized. And it kind of maps all of the individual bits of data into what we think we know about the human genome. And so you can say, OK, we've got this much of this gene and that much of this gene or you can say, you know, Gene A has certain mutations and Gene B has other mutations. And so it allows you to ask whether whether they're mutations or changes in the amount of certain molecules and so forth. But you get to do it for all the genes and not only all the genes you can do it for, [but] for all the space in between the genes in the genome.
Harry Glorikian: Yeah, I you know, it's funny because just the other day there was the announcement that we quote "actually finished" the entire genome, which I thought was an interesting announcement. One more definition. So this term RNAseq, right? So, you know, drawing the analogy of DNA and saying, OK, RNA is the next level. And why has that become so important now in drug discovery?
Rafael Rosengarten: That's a great question, so again, for your listeners who may not live and breathe this stuff, there's a concept in in biology called the central dogma, and it kind of still holds. And the notion is that there are these different levels of organizations or different layers of the onion and peeling back the information that our cells use to conduct business. And the the core of this is DNA, and that's our genetic information that's encoded in our nucleus and it's passed down from parents to children. It's the heritable information, and I apologize to all my friends who do live and breathe this, who are going to call shenanigans on my definition of being overly simplistic. The next level is, as you described, is the RNA. And so RNA is actually a lot of things. But messenger RNAs are the transcription of the genes. So the DNA genes that hold our genetic information are converted through a molecular process into another kind of molecule. And that kind of molecule is RNA. It's chemically similar to DNA, but different, and that RNA tend to be in smaller pieces than the whole chromosomes, and they represent smaller pieces of genetic information, and they can vary widely from, say, one gene to the next in terms of how much RNA is made for that given gene.
Rafael Rosengarten: And then just to fill out the picture a bit more, in principle, then, those RNA molecules get turned into protein, or they are the specific instructions to create proteins, and proteins then go do the work of the cell. What I just told you is mostly wrong, but it's sort of the framework that we think about. So the reason why RNA, the middle layer, is so interesting in drug discovery, and I'm going to add to that, in diagnostics world, is because it's a bit more, let's call it dynamic than the DNA level. So mutations sometimes are heritable and sometimes they arise de novo. But once they've arisen, they're kind of there and they go through from cell to cell, once the cells divide. And that's, you know, that's important and interesting and meaningful information, you can learn a lot about what genes are potentially druggable from that. But it doesn't tell you a whole lot about the state of tissue or the state of disease in this moment, right? It's kind of background information in a way. And so RNA is a bit more dynamic.
Rafael Rosengarten: It changes. It can change on, you know, really rapid time scales, but certainly therapeutically relevant time scales. And so in some ways, it's a little bit closer to sort of what's happening now.
Harry Glorikian: Right.
Rafael Rosengarten: It's also just a different, it's a different class of information because there are these abundances, different genes at different levels. Those relative abundances have biological importance and sometimes therapeutic importance. A lot of cancers, for example, are bad for you. They are essentially dysregulation of gene expression, so they can arise from mutations or they can arise from events at the DNA level. But it's understanding how much of some species of gene is being expressed in the RNA that can be informative or potentially therapeutically actionable. And I'm going to shout out to my proteomics friends, the guys who study proteins. That may be even more therapeutically relevant in a sense, because most of our drugs actually target proteins. And that's quite the key of it. Except for gene therapy, which is a big deal, especially in the CRISPR era, we're not often targeting DNA with our drugs, right? Mostly, we're targeting proteins and occasionally we're targeting RNAs and less frequently we're targeting DNA. Again, all CRISPR bets aside, right?
Harry Glorikian: Yeah. No, we did an episode with talking about CRISPR and, you know, amazing advancements happening there. But now, being from Applied Biosystems, I remember an entire room full of sequencers where we, I think they were like 600 or 800 we had running 24 hours a day at one point. Now I can do that on a desktop, right? But. There's a lot of data that comes off that. T hat's a challenge, I think, for people in drug development to manage that much data. You started at Baylor with a lot of your research. How did how did you personally encounter these challenges in your research?
Rafael Rosengarten: I mean, it was very much this challenge that inspired us to start Genialis. So the conception story of Genialis is my co-founders and I, we really wanted to be able to do advanced cutting edge data science like machine learning, AI type stuff, which I'm sure we'll talk about at some point, in order to really bring kind of the next level of analytics to bear on biomedical problems. And what we realized is that's all well and good, but you can't do any of that stuff unless you get the data in a place where you can work on it. And I remember going to talk to one of the top researchers at all ofe Baylor College of Medicine. This person is top of her field, chair of department, et cetera, et cetera. And I asked her, How does your lab deal with your data retention and your data management, your data analysis? And she said, Glad you asked, this is such a big problem. We just had one of our postdocs leave, and he took his little thumb drives with him, and all of the data from all of his stuff was on those thumb drives. And now we can't reanalyze. I was like, You're kidding me! She said “We had to go and redownload download some of it that he had published and put online.” So, so even top researchers didn't have a clue how to do this. And this wasn't that long ago. I would say that drug companies by now are mostly more savvy and certainly the commercial sector for data management tools is thriving, right? There are some really good commercial products.
Rafael Rosengarten: Genialis has one. There's some others of note. And Big Pharma has invested a lot, obviously, in building in health solutions. But this creates another kind of complication, which is you get all these different solutions and they don't all talk to each other. Even having data on different clouds. Some people may use Amazon and others Google and others still, Microsoft. And those are the three majors. You know, those create silos in a way. So, so you know, the cloud has been super helpful. The advent of software purposely built for biological data management has been helpful. But, you know, there's still a lot of work to do. And I'm going to argue that the kind of next, let's not call it a frontier, but the next big challenge and the one that we encounter a lot, it's not even around the primary data. We're good now. We're good at sucking that off the machines and putting it in the cloud and organizing it and getting it processed really efficiently using distributed computation. Now the challenge is getting what we call the metadata, the annotations of where those data come from. Is it coming from patients and if so, what's the patient information associated with it? Is it an experiment? Getting those metadata consistently curated and attached and linked to the primary data is a big and very important challenge, and it's one that I think will be solved in a similar way through these software solutions. But it takes a lot of will and a lot of manual effort at this point.
Harry Glorikian: Just to summarize, the software that you have is helping biologists and clinicians work with data without necessarily having to become a bioinformatician, if I had to frame it that way, is that is that a decent representation?
Rafael Rosengarten: That is that's one of the softwares we have. So you're referencing Genialis Expressions, which was kind of our initial flagstone software. I'm excited, though, in November, at Biodata Basel, we launched our new software, our newest product, which is called Responder ID. And this is where our dreams of really applying machine learning and AI to these data have finally come to fruition. Responder ID is a software or really, it's a suite of technologies that we use on those clinical data and on those experimental data to actually extract knowledge and very specifically to figure out which patients are most likely to respond to certain therapies. And so the first piece of software is really the kind of about the data management. It's about getting data organized, getting it processed, all the best practices and efficiencies around that. And that was sort of, you know, I don't want to call it last year's problem because it's still a problem, but it was the first thing we did. It's where we started. And it's got some beautiful visualizations and it does let bench scientists like myself work with their own data. But the new stuff is where we're really bringing the application to bear on human health and on value propositions that I think really resonate with pharma, diagnostics, and other biotech and frankly, clinicians and and ultimately patients.
Harry Glorikian: So, well, that's great, I mean, that transition to the new software, I must have missed that in when I was doing my research. I hadn't seen that yet, but what are some of the stories or anecdotes by customers that you can share? What have they been able to say, accomplish with it, so that we can put it into context for the listener?
Rafael Rosengarten: Yeah. So you know, most of our customers are biotech drug companies and we help them solve a number of problems. But the key challenge is that drug development is just an incredibly risky and expensive and time consuming proposition. Most of our work's in the oncology space, not all of it, but it's a good place to make this example. The success rate of a drug that enters a Phase I clinical trial in the cancer space that actually makes it to market is something like three or four percent. It's dismal, and it's among the lowest of any therapeutic area. And there are any number of reasons for that. But the simplest, simplistic one is that biology is complicated and patients are diverse, right? Even within a single disease like, let's just say, breast cancer, there are at least four kinds of breast cancer. There are probably 40 kinds, and there are actually probably more than that. Each individual's disease is going to have its own unique flavors. And so what we allow a company to do, let's say a company that's developing a drug against, for example, breast cancer, is to really try to understand how many molecular types are we talking about, which ones are going to respond to our drug? And can we find those patients ahead of time? And what that lets them do is think about alternative and sort of novel and innovative strategies for designing clinical trials. It allows them, if they so desire, to think about partnering out on diagnostic development with third parties to actually create a diagnostic to go with their drug. That's not, obviously, necessary. You can you can build assays that you run in-house, but that's an alternative.
Rafael Rosengarten: And to make it very concrete, we have one partner we work with a lot. A company called OncXerna Therapeutics. And with them, we've helped develop their first biomarker as part of their biomarker platform to the point not only of clinical trial assay, but also it's been licensed by Qiagen to be turned into a companion diagnostic for their lead drug and a research-use-only assay for scientists writ large around the world. And so, you know, this is a great success story. In about the course of two years, we went from taking a published academic signature, something in the literature—and by the way, there are about a million of these public academic signatures and there are only 46 approved companion diagnostics, so there's a big gulf between them—we went from an academic signature—and this was hand in glove work with them, so I don't want to take all the credit, but we certainly did a lot of the heavy lifting—and we built a category-defining first-of-its-class machine learning algorithm that learned a complex RNA-sequencing-based signature that predicts with uncanny ability patients that are going to respond to a wide array of drugs in a wide array of diseases. So it's pan-cancer, multi-modality, right? This is just it's an astonishing clinical advance, in my opinion, and it's something I'm clearly very proud of and willing to self-promote. But I do think it's an important advance, and I think it shows the power of both the Genialis philosophy around modeling biology and pairing patient biology with potential therapeutics, but also just what you can do if you're really thoughtful about getting the data in the right place, treating the data properly, and then using machine learning and some of these advanced algorithms to decipher.
Harry Glorikian: Yeah, I mean, I think we're starting to get to that cusp of producing the data is getting faster, more cost effective. I mean, if Illumina actually gets down to, I think they, at the last JPMorgan, they said, we're trying to get it down to $60 for whole-genome. But at some point you're getting to numbers that are, I don't want to say a rounding error, but damn near close to that. And so the burden is going to fall on, how do I interpret all this data and what do I do next, right? What's actionable? I mean, I think the treating doctors are like, this is all great data, but tell me what to do, right? And it sounds like your new suite of software might be more applicable for a clinician or to to be communicated to a clinician, than just on the research side. So is is Genialis now moving beyond its original set of customers and moving more towards the clinical space?
Rafael Rosengarten: I certainly think that's, on the horizon, that's something that we're contemplating. You know, the U.S. health system, well, systems, plural, is a complicated beast, right? And so there are certainly big companies that have products that are there for drug companies and products that are there for patients and products that are there for providers and so forth. And that makes sense. I think once you've got a wide enough kind of horizontal, you can stack all these verticals on top of each other. You know, hopefully we get big enough to do that ourselves. But you know, for the time being, we found this really, you know, this really great motion and success story working around certain therapeutic modalities for certain therapeutic opportunities. I actually think what may be the bigger prize is to take what we learn about disease biology from some of these diagnostic models and turn them on their head and say, OK, we've shown this model really captures patient biology and it works. And we know that because look, there are patients and they respond to the drug that we predicted they would. We've definitely cracked something there. Now let's take what we've learned about that patient biology and interrogate this model for new therapeutic opportunities. What about all the patients who don't respond to this drug? What will they respond to? The model still has them pegged as nonresponders. The model understands their biology. We just need to interrogate it for the next generation of therapies. And so I think this is where my vision of precision medicine maybe deviates. Diagnostics is an industry. Drug discovery are an industry. Those are separate companies. Those are separate industries. But to me, precision medicine shouldn't be this kind of linear thing where you start with the target, you end up with a drug and a diagnostic, and that's where it ends. It should be a circle. It should wrap around. And what we learn from patients should feed right into the next round of drug discovery, right? And so I'm interested in playing at that sort of fusion point where the where the ends of the string meet and form a circle. And so we're really interested in partnering and learning more about, for example, discovering new drugs to match the targets, right? And so I kind of see that as where a lot of Genialis's future focus is going to go. I'm not ruling out patient reporting software. I'm not ruling out more clinical products. That would be logical, but my real interest is thinking about helping the patients who just don't have therapeutic options today.
Harry Glorikian: Let’s pause the conversation for a minute to talk about one small but important thing you can do, to help keep the podcast going. And that’s to make it easier for other listeners discover the show by leaving a rating and a review on Apple Podcasts.
All you have to do is open the Apple Podcasts app on your smartphone, search for The Harry Glorikian Show, and scroll down to the Ratings & Reviews section. Tap the stars to rate the show, and then tap the link that says Write a Review to leave your comments. It’ll only take a minute, but you’ll be doing us a huge favor.
And one more thing. If you like the interviews we do here on the show I know you’ll like my new book, The Future You: How Artificial Intelligence Can Help You Get Healthier, Stress Less, and Live Longer. It’s a friendly and accessible tour of all the ways today’s information technologies are helping us diagnose diseases faster, treat them more precisely, and create personalized diet and exercise programs to prevent them in the first place.
The book is now available in Kindle format. Just go to Amazon and search for The Future You by Harry Glorikian.
And now, back to the show.
Harry Glorikian: When I think about this and where we're going with this and the I hate saying it, butthe old dogmatic way of looking at it is very compartmentalized as we look at it in discrete pieces. And these data analytics platforms allow us to look at multifactorial, or almost turn the data into a living organism where we can look at it in multiple ways, and I think it's hard for people to get there mentally. I mean, sometimes, sometimes when I'm looking at something, I realize that my limitation is the information that I have about a particular area and that I need to learn something new to put another piece of the puzzle together. But I think this, let me do this and then let me do this and then let me do this. That's breaking down because of the data analytic capabilities that we're bringing to bear. Applying AI, machine learning, or in reality, sometimes just hard math, to solve certain problems, is opening up a wider aperture of how we would manage a patient and then treat them appropriately. And I think. Hell, I don't know, Rafael, I'm a little worried, I don't think the system is necessarily designed to absorb that next-gen opportunity, right? Because somebody will be like, OK, where do I get the information? Does that go in the EMR? I mean, wait, where is there a code that I can bill for it? I mean, there's these arcane roadblocks that are in the way that have nothing to do with, "I've got this model, and I'm telling you this will work on this patient," right?
Rafael Rosengarten: Yeah, I don't know that I'm smart enough to know the solution to that. I will say that there are some really exciting newish young venture-backed upstarts that are interested in disrupting hospital systems, point of care, EHRs. All of that, is fair game, right? It is, as you described, it's just ripe for disruption because it's so, you know, it's so cobbled together, right? You know, I'm thinking about when my wife and I moved from Houston, Texas, to the Bay Area and then we got pregnant with our second child. We wanted to have all of our medical records from pregnancy number one sent from Texas Medical Center, which is one of the shining jewels of health care institutions, to John Muir Health System in the Bay Area, which, listen, they were changing out the wood panels from the 1970s during all of our doctors' visits. And literally, we asked the doctor if he could just print, print something for us. He said, No, I can't do that, but I could write it down on a sheet of paper for you. Like, you know, it's. But that's that's, you know, I agree with you. There are going to have to be changes top down, bottom up, and there's going to have to be hopefully support for this in the regulatory bodies, you know, at the governmental level.
Rafael Rosengarten: Where I live and breathe, those is really kind of in a life sciences sector of the health care system. So again, we're interested in in drug development, we're interested in diagnostics, we're interested in drug discovery. And those themselves are kind of big things. So where I think about changes and regulatory and systemic stuff is more along, like, what is the FDA doing to to adopt or adapt to these kind of new technologies? What about standards like how are we thinking about data standards, model standards? Genialis is a founding member of and I'm on the board of directors of the Alliance for AI and Health Care. And this is a really exciting and rather amazing industry organization that was stood up at JP Morgan in 2019. And you know, we've got gosh, I don't know what the headcount, the member number now is, but over 50 member organizations, including the likes of Google and and Roche and bigs like that. Some of the more household names in the smaller biotech community like Recursion Pharma, In Silico Medicine, Valo Health, et cetera. And then and then companies like Genialis as well. Big academic centers. So we have a real great brain trust and we're interested in tackling, I'm going to call them, these hard, boring but incredibly important systemic questions around regulatory and standards and so forth. Health insurance, Medicare, all that stuff is a big fish, and we haven't, you know, we haven't set our hooks in it yet, but you know how hospitals bill and those kinds of codes, we’ll have to have to revisit that at some point, for sure.
Harry Glorikian: Yeah, I know that you're a member there and sort of interesting to hear why you got involved in how you see it working. So if you think about the standardization side of this, you know, what is what is the organization sort of advocating for? Because I totally agree with you, but at some point, I think you almost need to reach back towards, how is somebody doing an experiment to make sure that then the data comes out the other side in a standard way, right? Because I used to joke, which sample prep product are you working with? And I could tell you sort of what direction something is going to lean. And that that in and of itself is a problem. So how is AAIHC thinking about some of these problems, I don't know if there's a proposal. What have you guys proposed so far?
Rafael Rosengarten: That's a great question. So we have workstreams around things like the FDA, working with the FDA to propose guidance for a good machine learning, practice guidance for software as a medical device, AI as part of software, as a medical device. So a lot of this, it's less concerned with can we rein in and constrain the experimental part? Because again, that's that's a huge world. And maybe it's not really where the constraints need to be. But rather can we come up with a common set of guidelines for how you evaluate the quality of a data set, right? Recognizing the data are going to come in a lot of shapes and sizes and flavors, and even two different RNA sequencing data sets that are produced on different machines or with different kits may have slightly different flavors or tints to them. That's fine so long as you have some guidelines for characterizing those differences, for appreciating those differences and then for knowing what to do with the data, given those potential differences. A lot of the concern around AI in a regulated setting is that, the whole promise of a machine learning approach is that it gets smarter the more data it sees, right? So these should be, these algorithms should evolve in a way they should be living and breathing. But if you have a regulated product that's to work on patients, it's got to work the same every time or, you know, can't get worse.
Rafael Rosengarten: So this is, there's a tension here, but it's not unsolvable. It's not insurmountable. For example, you know, a regulated AI doesn't have to evolve in real time. It can be updated over time, right? Right. And it can be it can be locked and then operate, and then you can improve it and update it and redeploy and relock. So building the plans, what are the change plans? How do you demonstrate that the retraining or the improvements are actually improvements? These are the kinds of things that at least we can sink our teeth into today. And then we're also interested in the standards problem. I think the organization is not necessarily going to be dogmatic about recommending exactly what the standards are today, but what we're trying to catalyze those discussions, right? And we're trying to create frameworks where those discussions can actually lead to some actionable tools. And there are examples of organizations that have done this in other fields. So we do have some blueprints. But it's a lot of work. And frankly, that's the privilege of being in the organization. It gives you the opportunity to roll up your sleeves and build the industry of the future, to build the industry you want to operate in.
Harry Glorikian: Yeah. And this has got to be in lockstep with the regulatory authorities and everything to make sure that everything is, everybody's on the same page so that when you come up with a golden solution, they're ready to accept it. Because we can't have, you download the latest software for your phone and then it breaks, right? That's not an acceptable update that you can do, right, and somebody has to release a patch to get it to fix. You know, that's that doesn't necessarily... I'm sure it happens in our world, but it's. It's really not what you'd like to see happen.
Rafael Rosengarten: Yeah, yeah. You know, I can tell you from having had to invest in a lot of the kind of procedures around clinical reporting in software and so forth, and, working with some really top tier point of care software providers, it's not foolproof. But boy, there are a lot of hoops to jump through, right? Like things do get tested the whole way. And I would just, I would argue, although, you know, let me not be overly full of hubris, that there are plenty of other failure points that are a lot more likely to fail than the AI software that's predicting a biomarker not working in a particular instance, right? Given the room for error in things like biopsy collection and human handling. There's a lot of stuff upstream of that where human error is more likely to play a part. That that may or may not be sweet solace, right. That might not help you sleep at night. But I think that the regulated environment, especially around regulating computational tools, can be rather bulletproof.
Rafael Rosengarten: So is there anything else going on that at Genialis that that we would want to know about that and directionally or what's next, that you can [share]?
Harry Glorikian: Yeah, I mean, the exciting stuff is really twofold. It's, you know, just going deeper with our partners, right? So clinical development, as I mentioned, is is a long game. And you know, we like to start working before the drugs in the clinic, right? So these are meant to be long partnerships. And the other piece of this is we're doing a lot more internal R&D. A lot more internal R&D, a lot more work with our academic colleagues. And so we're really, really excited to just, you know, to innovate our way out of some of these hard problems.
Harry Glorikian: Well, that's necessary in this field, right, you're always going to run into some, I like to call them speed bumps because I don't believe that they're like insurmountable problems, but they're speed bumps that you need to like innovate over or around.
Rafael Rosengarten: Mm hmm. Yeah. So, you know, I want to give you something meaty like, you know what to look for from Genialis. So, sometime soon, my hope, knock on wood, is that we'll have first patients enrolled in clinical trials that are the biomarker I described to you earlier. This is the OncXerna trial. First patient enrolled, that's going to be super exciting. It's a Phase III trial and we're going to be stratifying patients with the biomarker. I mean, just the gratification of actually having our technology potentially impacting outcomes is huge. We've got a lot up our sleeves in terms of internal development improvements to Responder ID, but also, you know, some biomarker work we're kind of doing for ourselves, digging deeper into some pernicious problems in cancer that others haven't adequately addressed, in my opinion. And some some exciting partnerships, hopefully around, kind of…. we'll call them data partnerships. We talked a bit about just the scale of the data challenge, though, is it lives all over the place, right? And so there are different ways of getting your hands on it. And one of the ways a lot of companies have gone about is to become the testing companies, right? There are some giants out there that sequence literally millions of patients a year, and they've got big data warehouses, right? We haven't done that ourselves. And so we rely oncollaborations for a lot of our data. Not all of it, but we're building some of these collaborations, and I'm hoping we can talk more about that in future episodes or in other forums.
Harry Glorikian: Just for a second, so people understand the magnitude. This Phase III trial, how many how many patients would you say are in it?
Rafael Rosengarten: I need to be super careful not to misrepresent someone else's trial. It's going to be on the order of several hundred. You know, it's a properly powered Phase III and it's got two treatment arms. And so, you know, so it has to have quite a number of patients. And that's, you know, I would say that's a typical sized trial of for this stage in this kind of disease.
Harry Glorikian: Yeah, I just want people listening to sort of get an idea of like, these technologies are, you know, can affect lots of people and then if that drug comes through and then the technology is utilized afterwards to sort of stratify people or the biomarkers, then there's an even larger population of people that then gets affected by the work that you guys are doing.
Rafael Rosengarten: Yeah, yeah. I think that's right. And you know, in a way, you know, our commitment to the sort of biomarker driven, you know, drug development, it's very principled. It's based on this idea that patients deserve to have the best treatment option, right? And there are some amazing drugs out there that when they work, work miracles. But they don't work that often. Right? And some of these drugs have, you know, first line approvals in dozens of diseases. But again, in some of those diseases, they work for half the patients, and that's great. And that's probably how it should be. But in some, they only work in maybe 15 percent of the patients or 20 or whatever the threshold is, because they were better than the alternative, right? But if you could tell which of those patients are going to respond, then at least the ones who aren't can seek other options. Or you know that we've got to develop drugs for the others. So it's very principled, although it's complicated because from an economic standpoint, if you have the ability to sell your drug to everybody, of course you're going to do that.
Harry Glorikian: Yeah, look, I drank that Kool-Aid. I mean, Jesus, 20 years ago, right? I mean, you know, why wouldn't you want...I mean, if you were a patient, you'd want the best drug you can get, right? Because the data says that you respond to this particular drug. It's getting the system to that point. And I have seen, I have had stories where the data said one thing. They put the patient on it. They looked like they were responding. A new trial opened up. And somebody suggested that they go on the new trial, even though the therapy was working. And they switched and the outcome was not positive. Right. And so it's one of those things of like, I don't understand. The data clearly pointed in a particular direction and you deviated from that, and that doesn't make any sense to me. As a science person is as well as an investor, if the data is showing something, you better respond to the data or you're not going to be happy with the outcome. It's just seeing that implemented in a way that makes it very actionable for everybody, and they embrace that. That's where I sometimes, I find, you know, the biggest problems. But I totally agree. I mean, I have a whole chapter in my new book about that whole dynamic of why you want the data, how the data impacts you as a patient. What are the sort of questions you should ask, et cetera, because if you don't have that information, you're making suboptimal decisions.
Rafael Rosengarten: Yeah. No, and that's absolutely right, I think the point you make there is probably the key one, which is a lot of biotechs and companies like ours, we operate with kind of a world view of our own research and our customers’. But we have to remember that the reason we do this, the reason we get up every day and the reason we toil is it's because we can impact patient lives. And if you actually want to really foment that change, then that subset, that stakeholder, needs to be involved, right? A patient needs to understand what are my choices? And so if a patient comes into the clinic and has a grave illness and the doctor says, well, this is the approved drug, but there's a test that could tell you if there's something else. I mean, if I'm the patient, I want to take that test. I want to know what my options are. And I think that frankly, it's unrealistic to expect publicly traded companies to not try to maximize revenue. That's just kind of the system we live in. But it's also incumbent upon us to to engage patients, to help them understand what their options are, to engage physicians the same and to say, there are multiple approved drugs, maybe, or this is the one, but there are some investigational drugs that haven't been approved yet that may be better fits for your disease. Remember, your disease isn't necessarily the same as someone else who happens to have it in the same tissue. And so I think that's a big deal, and I do think that there are any number of exciting organizations that are really focused, doggedly focused on this point of patient engagement and especially patient engagement around data.
Harry Glorikian: No, I mean, I always I tell every one of my guests, “Hurry up, go faster,” because I'm not getting any younger and theoretically like, you know, statistically, I could end up in that place. I want the best that I can get when I get there. So Rafael, I know it's getting late where you are. So really appreciate your time and the opportunity to talk about what you guys are doing and the impact that it's having on not just drug development, but downstream on patients.
Rafael Rosengarten: Well, thank you, Harry, for having me, for giving me the opportunity. This has been a lot of fun to connect over this.
Harry Glorikian: Excellent. Thank you.
Harry Glorikian: That’s it for this week’s episode.
You can find past episodes of The Harry Glorikian Show and the MoneyBall Medicine show at my website, glorikian.com, under the tab Podcasts.
Don’t forget to go to Apple Podcasts to leave a rating and review for the show.
You can also find me on Twitter at hglorikian. And we always love it when listeners post about the show there, or on other social media.
Thanks for listening, stay healthy, and be sure to tune in two weeks from now for our next interview.