Layne Sadler founded KeyBio to help pharmaceutical companies better identify genetic traits in patients that may explain differing responses to cancer treatments. KeyBio’s computational approach promises to add a new dimension to precision medicine, which now looks at the genetics of the tumor, by also incorporating the genetics of the patient to improve outcomes in clinical trials.
Highlights:
Sal Daher Introduces Layne Sadler
What KeyBio is Solving
"... The revenue is solely focused on cancer drug discovery and the reason for that is because that's where all the money is in the pharmaceutical industry..."
"... 90% of clinical trials fail and the cost averages around 792 million, spending about a billion dollars to run a trial..."
"... I developed this AI platform and I wanted to start collaborating with other principal investigators or heads of labs in the Boston area..."
"... The DNA of tumors, over time, it mutates and it becomes very different from the DNA of the host..."
Why Layne Chose Entrepreneurship
ANGEL INVEST BOSTON IS SPONSORED BY:
Transcript of “KeyBio”
Guest: Layne Sadler
Sal Daher: I'm really proud to say that the Angel Invest Boston podcast is sponsored by Purdue University entrepreneurship and Peter Fasse, patent attorney at Fish & Richardson. Purdue is exceptional in its support of its faculty, faculty of its top five engineering school in helping them get their technology from the lab out to the market, out to industry, out to the clinic. Peter Fasse is also a great support to entrepreneurs. He is a patent attorney specializing in microfluidics and has been tremendously helpful to some of the startups, which I'm involved, including a startup that came out of Purdue, Savran Technologies. I'm proud to have these two sponsors for my podcast.
[music]
Sal Daher Introduces Layne Sadler
Welcome to Angel Invest Boston, conversations of Boston's most interesting founders and angels. Today we have a very interesting guest. His name is Layne Sadler. Say hi, Layne.
Layne Sadler: Hey, Sal, thanks for having me.
Sal Daher: Great to have you on. Layne is a well-established product person but with data analytics and life science focus. He is self-taught in the area of data analytics and he's also self-taught in the area of life science because he's worked in that space. He has started a company called Key.Bio, which is getting a lot of interest because what it's doing, and it is hoping to do on an ongoing basis, is to help pharmaceutical companies in drug trials or researchers have a much broader understanding of factors that might contribute to the success of a clinical trial of an experiment.
One of the problems that Layne sees as I understand it, and, Layne, please, stop me from saying nonsense if I'm talking nonsense here, but I just want to set the stage here, there is precision medicine that has been focusing on understanding the genetic makeup, for example, of a cancer, of a tumor. We are now able to take a sample of the tumor and then look at the genome of the tumor and try to target treatments aimed at that particular tumor. However, tumor behavior is not determined solely by the genetic makeup of the tumor. It is also determined by genetic variations and, basically, mutations in the makeup of the person who has the tumor.
Layne has identified, out of the 24,000 genes, places where you could look to see if variations in those areas could explain better outcomes in clinical trials than the current standard of looking at only the genetic makeup of the tumor, so look at the genetic makeup of the host, look for those abnormalities, and then look at the interaction. This is where data analytics really shines. Am I explaining this correctly?
What KeyBio is Solving
Layne Sadler: Yes, that's right. You mentioned those differences in our genetic makeup and I think the past few years of the pandemic has taught us that each individual has a different immune system. When these tumors arise in our bodies, we're each going to respond to those differently. It's not just about the content of the tumor, although, that does play a role. It's also about how we as hosts are equipped to respond to that cancer, just like any other disease.
Sal Daher: Phillips Kuhl, friend of mine, he's retired now, but he ran a very successful series of conferences in the space. He used to say that initially, we were really excited about sequencing the genome. When we sequenced the genome, we figured out, "Wait a second, now that we know what the genes do, we need to know about the proteins, the proteins that the genes code for. What are they doing?" Then they said, "Well, we have to go the next step," is proteomics to figure out the proteins surrounding those genes. It probably gets bigger.
Then they discover but the action of the proteins, sometimes they're mediated by some sugars, glycomics, and the thing just kept going and going. Then they're looking at the metabolites, which are various types of product from these metabolic reactions. The scope of looking at this, it just keeps expanding. Then you have influences which point us in the direction of epigenetics, which is influence of the environment, various things that cause expression of certain genes, or not expression of certain genes, it is a massively, massively complex problem. That is why there's so much variability on therapies that they work at 30% of patients they don't work at the other 70% and we don't understand why.
There are all these layers to the onion, and you're attempting to unwrap one particular layer here by doing something which is just allowing to look at a multivariate understanding. Instead of looking at just one variable, when they're analyzing that likelihood of success, is this gene present or not? In the cancer, you're looking at initially 50 of these variables that are the most likely places, very promising. How does this platform, how is it instantiated now? Is there an API that users can use to upload their data and then to clean it and do all the data processing that has to be done?
Layne Sadler: The platform itself is called Artificial Intelligence Quality Control, AIQC, and that QC is a trigger word for people in the biotech space, the bioinformatics space. When I say bioinformatics, that's just the science of working with biological data. It plays by its own rules in comparison to some of the other types of data analysis. This platform that I developed is actually open source and what it helps you do is rapidly, rigorously, and reproducibly conduct artificial intelligence training experiments.
You can train algorithms really fast and really well. I mentioned that it's open source, yes. How do you install it? It's just out there and anyone can go and download it, but the secret sauce is what I've built on top of that for bioinformatics in the oncology space and that's where the patent-pending technology of KeyBio really comes into the picture.
Sal Daher: Okay. We've had Jay Batson on to talk about Drupal and the open source, the business that exists around Drupal. The code is available. It's open, but there's a whole community around it to help people use it. In your case, you have specialized tools that make it more effective.
Layne Sadler: That's right. I like to look at open source as a distribution strategy. If you look at it from a marketing standpoint, and as a product developer by trade, you come to learn that when you're adopting any piece of software and it looks and seems perfect, it'll probably get you about 90% of the way there. Then in order to really extract value from it, you have to work with the developer and that's where that commercialization relationship comes into play in open source.
Sal Daher: That additional customization to make it really valuable for your application, for your use. Is your revenue model going to be one where you're going to be getting fees for building these specializations or customizations?
"... The revenue is solely focused on cancer drug discovery and the reason for that is because that's where all the money is in the pharmaceutical industry..."
Layne Sadler: The revenue is solely focused on cancer drug discovery and the reason for that is because that's where all the money is in the pharmaceutical industry. Everything is based around drugs and I want to be the one providing the data analysis, not giving customers a tool that they can use to go and conduct experiments because you have to really idiot-proof things, you have to jump through a lot of hoops with compliance. It's a lot easier when working with pharma if you just provide the services yourself because I've actually found in engaging with them in previous roles, even when you're providing them a tool, they push you to go and do all the services work for them anyways.
Sal Daher: Right. How scalable is that? If it depends entirely on you creating a customization, how many times can you do that in a given year?
Layne Sadler: The data types of bioinformatics and drug discovery are actually really standardized. There's a few companies that control the protocols and they're really well documented and pharma likes to copy each other. In the bioinformatics space, the data is structured very well. A lot of these experiments are designed to be reproducible. It's just a matter of what angles you're coming at from the data and what questions you're trying to answer. The workflow that we've patented is going to be very generalizable. It's more just about building relationships with pharma. When you say how scalable is it, when you look at the drug revenues and you consider the fact that we're building a royalty-based business model, you really only need a few enterprise relationships to get this going.
Sal Daher: The rationale is that by using the data in this much, much better way, you are going to be improving the odds of the clinical trials working out. You're going to be matching patient to medication in the trial so that they have better results in the clinical trial.
"... 90% of clinical trials fail and the cost averages around 792 million, spending about a billion dollars to run a trial..."
Layne Sadler: That's right. You talk about the economics a little bit, 90% of clinical trials fail and the cost averages around 792 million, spending about a billion dollars to run a trial. Some of those estimates go to 1.5 billion. 90%--
Sal Daher: 90% failings. If you can get it from 90% failing to 88% failing?
Layne Sadler: Right.
Sal Daher: [laughs]
Layne Sadler: The interesting thing is that the biomarkers that we're looking at, and we can talk about what a biomarker is, but 90% of the variation in our experiments that drives algorithm accuracy has been overlooked by the rest of the industry. They've just breezed right past it and intentionally filtered it out of their experiments.
Sal Daher: Okay. Unpack biomarker.
Layne Sadler: Right. If you look at diabetes as a disease, blood sugar is a biomarker but when you look at things in terms of DNA, it becomes much more complex. You mentioned that we sequence the human genome but when we did that, what we got was just essentially a really long book in a language that we didn't know how to speak, and still many of the genes, we figured out a few of their functions, but there's 24,000 genes and that only represents 2% of the human genome. There's a lot going on that regulates these different genes as well. 24,000 genes, how do you find the most important biomarkers? That's where the value of the protocol that we developed comes into play.
Sal Daher: Biomarker so then would be some kind of biological indicator that can be of value in predicting the outcome in an experiment.
Layne Sadler: That's right. You've got a gene that gene produces a protein, and that protein can either positively or negatively impact survival.
Sal Daher: Survival of the patient during the clinical trial.
Layne Sadler: Right.
Sal Daher: Very interesting. Do you have some kind of a ballpark estimate as to how much you can improve the survival in a clinical trial, in a typical drug trial?
Layne Sadler: That would be really hard to estimate, but where I would start is that when you get a cancer patient and you're a doctor and you're trying to look at the health outcomes, a lot of that right now is really guesswork. You've got the age of the patient, you've got the stage of their tumor, and you've got basic statistics that come up with survival percentages, and there have been methods like liquid biopsies that will look for the presence of tumor DNA in the blood, they can get in the range of 70% accurate. The methods that we're using right now are upwards of 95% accurate. Obviously, we want to expand our work to a larger patient size but just see a signal that strong in our preliminary data is really encouraging.
Sal Daher: Going from 70-something to 90-something is a big improvement?
Layne Sadler: With a little more granularity because we're looking at very specific biomarkers that can help group patients. It's not just like, "Oh, you're this year's old and you've got tumor this stage, you're going to be this likely going to survive. You've got ctDNA in your blood, you've got these specific genes, upregulating and downregulating you and therefore you might be a better match for this kind of drug," or, "You're high risk so we need to be more aggressive with your treatment." It's much more precision.
Sal Daher: The pharmaceutical company that is running the clinical trial has an incentive in having success during the clinical trial. They don't necessarily have an incentive in how the drug performs aftermarket. They don't want to be necessarily testing that but during the clinical trial, they want to have the best likelihood of success. That is the time that is of most value to you for you to come and say, "I can improve your likelihood of success and pay me based on the improvement success that you have. Give me royalties based on the success that I have on this framework that I'm going to create, which is specific to the experiment you're going to run."
You're going to invest your time and your knowledge, experience based on your open-source platform and then customize it. Drug company runs the clinical trials and then it gets approved. How do you state that in negotiations with the-- If you have survival of above X only this percentage of royalties; is that how you anticipate getting paid?
Layne Sadler: More so just the flat rate on the drug, as well as the tests that match patients to the drug.
Sal Daher: Okay. If the drug gets approved, you get a certain amount per use of the drug, and then you also get a certain amount-- Explain this matching again?
Layne Sadler: There's a concept called a Companion Diagnostic. You've got a cancer drug, that drug might cost $100,000. It might have deadly side effects so who is and isn't going to be taking that drug is of great interest to not only the doctors and the patients but also the government and the insurance providers who are helping foot the bill for-- There's a lot of cancer patients out there for it to cost $100,000.
Sal Daher: Right. I'm just imagining here. That if you pick the right biomarkers that give the best indication for whom those drug can be helpful, then you're going to end up with more sales of the drug because you're going to fit the drug to the exact population that it helps. You're going to capture everybody in the population that could be helped by that drug in a positive way where it pays to use this drug. Then you're also paid on that percentage, let's say, or-- What is that term again you said that you used?
Layne Sadler: Companion Diagnostics, CDx.
Sal Daher: No, but the criterion.
Layne Sadler: The biomarkers?
Sal Daher: No, not the biomarkers, but the criterion for compensation for payment.
Layne Sadler: Oh, royalties. You could also call it a licensing strategy and that's a vetted model for startups.
Sal Daher: Okay. Based on the success of the test ratio, what is that-- There was some other term that you mentioned.
Layne Sadler: Clinical trials themselves, you need to establish clinical benefit versus the standard of care. It could be displacing chemo as a first-line treatment. We've defined a lot of terms. I'm trying to figure out which ones are relevant.
Sal Daher: Well, we talked about, one, is royalty based on the drug being approved, the number of sales, and the other one is on test results-
Layne Sadler: Right.
Sal Daher: -only you're having a set of tests that accurately matches the drug with the right patients.
Layne Sadler: Right. Neat. That's the Companion Diagnostic.
Sal Daher: The Companion Diagnostic, exactly. All right, very good. I think it's pretty clear. Now I understand that you have a collaboration going on right now or more than one collaboration.
"... I developed this AI platform and I wanted to start collaborating with other principal investigators or heads of labs in the Boston area..."
Layne Sadler: Yes, there's a few up in the air right now. I guess the most relevant one to talk about is, the first, where I developed this AI platform and I wanted to start collaborating with other principal investigators or heads of labs in the Boston area so I reached out to Dr. Guru Sonpavde at Dana-Farber. He's since moved to AdventHealth Cancer Institute, but we collaborated to design a study around bladder cancer survival using data from the National Cancer Institute, more specifically The Cancer Genome Atlas or TCGA.
TCGA is really important. If you look at the Human Genome Project where we sequenced DNA start to finish, TCGA is the number two achievement of our country in the space of genomics where they looked at 30-some odd cancer types, and they gathered biomedical data from hospitals across the nation about clinical patient data, and then biomedical patient data.
Sal Daher: Basically, to match up the genomic information with the expression of those genotypes of that set of genes in a patient's population's health.
Layne Sadler: Right. You mentioned there's an explosion of different types of data, and TCGA really, early on, gathered a lot of those, DNA, RNA, proteomics, epigenomics. All different types of -omics is the catchphrase there.
Sal Daher: Very good. What is the significance of that?
Layne Sadler: Yes. The bladder cancer study that we did, it's important to understand that this is not diagnostic. We weren't going in looking for what causes bladder cancer. We were looking at patients that already have cancer, and we were trying to find the drivers of survival. It's a very different mindset shift. You have to think about not about diagnosing, but what causes the outcomes.
Sal Daher: What particular genetic differences are causing these patients to survive longer in the standard treatments?
Layne Sadler: Not tumor patients and not tumor genetics. We had to zoom out and look at--
Sal Daher: You're looking at the patients, not the genetics of the tumor but the genetics of the patient population and then saying, "These people who have these genetic traits, these variations tend to have greater survival when the bladder cancer is treated with standard of care, right?
"... The DNA of tumors, over time, it mutates and it becomes very different from the DNA of the host..."
Layne Sadler: That's right. The DNA of tumors, over time, it mutates and it becomes very different from the DNA of the host. What we found was that host genetics, germline genetics is the technical term, outperforms tumor genetics and tumor expression side by side in the same algorithm by a factor of 10 to 1.
Sal Daher: Wow. Wow, 10 to 1. Yes, that's tremendous. Germline refers to, basically, the genes you're born with. You're talking about the tumor genomics is the genetic makeup of the tumor, which are always a transition of changing and extremely noisy.
Layne Sadler: Right.
Sal Daher: Excellent. What is it that you need for the next step?
Layne Sadler: Yes. Right now I'm in the phase of the commercial on-ramp, is what I would call it. I'm evaluating partnerships and conducting case studies with different well-established biotech companies out there today that are working with biopharma. The other side of the coin is validating the research. Finding other bladder cancer data sets. Right now we're in discussion with Emory University, some of the people that help stand up TCGA to look at their data.
Sal Daher: Emory University. If you're in possession of an unusual bladder cancer database, get in touch with Layne Sadler.
[laughter]
Layne Sadler: We could really use your help.
Sal Daher: Yes. Yes, with focusing on bladder cancers, very good. At this point, maybe we could just do a brief promo for the podcast and then step back and delve a little bit into your journey into becoming an entrepreneur, which is always, to me, it's a little miracle that somebody has the courage to go out and do this extremely difficult and brave thing of starting a new business, a new company.
In terms of a promo for the podcast, I just want to say that, listeners, if you are enjoying this conversation with this brilliant product person who is tremendously informed in data analytics and the life sciences, if your mind is being expanded by this conversation, do us a favor. Help us get found in the podcast world by, first, following us on your podcast app. Every week, Angel Invest Boston shows up. More specifically, if you like the stuff that Layne is saying, you can rate the podcast. Please, give us five-star ratings, and also, leave a written review. It doesn't have to be much. Just a brief written review is very helpful. It tells the algorithm that somebody has cared enough to write up something. Do that for us.
Anyway, so Layne, do you have entrepreneurship in your family? How did you decide to do this unbelievably hard thing? You're a brainy guy whose work's hard, why are you doing this hard thing of starting your own company?
Why Layne Chose Entrepreneurship
Layne Sadler: I guess, I haven't really thought about the why so much. I just see an opportunity and want to chase it. It just started out on--
Sal Daher: Is the dog chasing a car?
Layne Sadler: Is the dog chasing a car? Yes, I see a hard problem and I want to solve it. Making money is just so boring. I went to Bentley which is a finance school and I realized very quickly that I didn't want to do finance. I knew I wanted to be on the business side of technology. I didn't really know what that meant, but I ended up graduating early and working at EMC, and from a middle-class America standpoint I had made it. It was as good as it's going to get, and I was just--
Sal Daher: EMC is a computer storage company.
Layne Sadler: Yes, but it's a big enterprise company.
Sal Daher: Big enterprise, enterprise computer storage. What you mean is you can have a very comfortable existence, being a well-compensated cog in a large machine that has been built by these-- Actually, I knew someone who's an early, early employee of EMC because he was a very generous donor to a nonprofit that I was involved with, a little school. I was a treasurer. We had some EMC shares that he donated in early, early days of EMC.
In this huge enterprise, you had a very comfortable life but you were seeing this massive opportunity which is, my gosh, they're doing clinical trials with just their keyhold. The way they're looking at selecting and designing these studies, it's just they're not getting the full scope of information that can be gotten if I build this platform and so forth. You had this dream and then you went off to do it.
Layne Sadler: Well, that was maybe five companies later. I worked at various startups.
Sal Daher: Okay, okay. I've heard this before. It's like, oh, maybe the next startup or the next company will be doing the thing that really scratches my itch, so to speak. Is there in your family anybody who's an entrepreneur?
Layne Sadler: Not really, no. They were all on the business side and that's why I went to business school. I have an uncle who worked at HP pretty early on.
Sal Daher: Okay, so large, large corporate enterprise companies working for enterprise companies as executives.
Layne Sadler: Right.
Sal Daher: Right, but not anybody creating their own enterprise. I interviewed Caleb Wursten, who's a Babson undergrad, and the guy when he was in high school, he had a bicycle rental business. He made enough money in that business to pay for the part of his Babson education that wasn't paid for by scholarships [laughs] because his parents are entrepreneurs. Both the mother and the father are entrepreneurs. That happens frequently, but this is interesting. Very interesting, very interesting. Where did you grow up?
Layne Sadler: Lancaster, Mass, so it's Nashoba Regional School District.
Sal Daher: Excellent, excellent. Very good. Layne, we've talked about the opportunity that you're addressing. We've talked about your economic model, your business model for making money, royalties. We've talked about your reasons for starting the company. I want to mention something here. This conversation that we're having really stands on the shoulder of giants, I should say, because prior to this conversation we had another conversation with Armon Sharei founder of SQZ, was kind enough to participate, Joe Caruso a very well-regarded angel investor here in Boston and an advisor to KeyBio, Layne's company, and also my friend, Gil Syswerda, who is an AI expert.
We had these three very knowledgeable people talking and trying to get to understand clearly the value proposition that is at play here. I just want to extend my thanks to Armon, to Gil, and to Joe for the time that they put into this project so that we could have a much more informed conversation so that I would be coherent speaking to Layne who has a tremendous depth of knowledge in this field and a lot of experience. I just wanted to express my gratitude. Any further thoughts Layne?
Layne Sadler: Yes, I just want to say thank you so much for having me, and helping me tease the value out of this in a way that's understandable. Pulling in the other guests to help do so as well. I really appreciate it.
Sal Daher: When you see pitch decks hundreds, sometimes thousands in a year as I do, as Joe does, Joe Caruso, sometimes you see a founder early in the race, and then you see the founder late in the race, and you are astonished by how much sharper the message becomes, just the conversation that you have about the proposition, the venture. You know what your business is, you have this idea, but you cannot articulate it in a way that people understand it fully until you have this back-and-forth, this dialogue. Presenting pitching is really important. In a way, this podcast has served as that.
Joe Caruso has a very good nose for this process of refining and improving the messaging which is really the important thing to explain to investors and to explain to potential customers. Of course, you're going to have different messages for different audiences because if you spoke to the angels in this audience the way that you speak to people in the pharmaceutical companies, it would go right over our heads and we wouldn't understand what you're saying. This has been an effort to put it in language that's understandable to us which may be accessible to people in the pharmaceutical, researchers and the executives of the pharmaceutical industry, but not to us.
Layne, I'm very grateful to you for having the patience to go through this multi-date effort so that we can come up with a valuable interview that will explain the propositions of KeyBio.
Layne Sadler: Thank you. Yes, it's been a lot of fun. Like you said, I think you pick up a nugget in pretty much every conversation you have. You get one thing that clicks and sticks with you with each person that you talk to, and then you develop that well-crafted and honed message like you said.
Sal Daher: Exactly. Thank you very much. This is Angel Invest Boston. I'm Sal Daher. Thanks for listening.
[music]
Sal Daher: I'm glad you were able to join us. Our engineer is Raul Rosa. Our theme was composed by John McKusick. Our graphic design is by Katharine Woodman-Maynard. Our host is coached by Grace Daher.