You can find the concise version featured on the Longitude SoundBytes podcast here: 126 Tips on Science Funding, Communication, and AI | Rowland Pettit

Transcript

Louis: You studied biophysics as an undergraduate at Duke University, then went on to Baylor College of Medicine, where you earned an MD and a PhD in bioinformatics and artificial intelligence. You also received an MBA from Rice University while at BCM. Why did you choose to pursue an advanced degree in business while already pursuing advanced degrees in science and medicine?

Rowland: Well, thanks, Louis, and thanks for having me on. I mean, this is a really great question, and it was one that I debated at the time. I mean, I definitely do believe in the value of formal education. You don't know what you don't know. And I have tremendously benefited from having incredibly smart people take the time to frame things and explain things to me. So, the blanket answer is, I was just curious. But the formal answer would be, I really wanted to understand the commercialization process. I had at that point, this was 2019 when I was considering it, I started in 2020. But I had done a good amount of medical school and PhD graduate school, and I had seen a lot of interesting innovation potential both in science and medicine during training, both certainly in med device coming to market, when I started the PhD, frontier science and its clear applications to both human health and agricultural science, etc. And I didn't really understand how people thought about bringing those to market. I was certainly getting to see how people were reacting to that, how they were, you know, performing clinical trials to test it or think about rolling it out with informed consent and bringing it to patients or otherwise, but I just wanted to understand that.

So, I, you know, took the next step and tried to think about if I did want to understand the business side of science and medicine, what would that look like? I was in Houston, which has the Texas Medical Center, a huge medical center. I was very fortunate to train at Baylor College of Medicine, which is a top medical school and also pursued my PhD there. Right across the street is Rice University, which is an enormous resource. They have an excellent entrepreneurship-focused MBA program. So, the long story short of it is, I was able to get a little bit of a discount as a current student and apply for their evening MBA course, which is still an MBA. It's just in person at night, which gave me good flexibility to do my PhD during the day and then from like six to nine at night, learn about commercialization and just the overall business practices with some focus in medicine, but of course, also just the general MBA training as well. So, the long story short is, I really wanted to understand how you get from A to B, right? Not just say, hey, we have a product and here's how we're going to implement it responsibly for, you know, in science or for patients. And so, that's what drove me to pursue the MBA during my MD PhD.

Louis: That's an excellent overview, Rowland. We are sure to get into some of your work in those fields. I'd like to start with a high-level question. Could you briefly explain how research in science, technology, engineering, and math gets funded for both public and private sector projects?

Rowland: Louis, I love this question. And when you sent it to me, I was really excited to get to talk about it. This is one area that I do think I've had a front row seat in order to see all the areas of funding throughout all stages of the process. And there are several different ways that we could try to think about this. But in the interest of time, I'll focus more on the practicalities of it. But I would, you know, if we need to, we can talk more about it. So, let's start with what I, you know, maybe would consider more, you know, blue sky, primary research for the sake of expanding, you know, human knowledge of the world. And that oftentimes the traditional heuristic would be that occurs in academia.

That doesn't perfectly hold true. There's plenty of big biotechs and others that are doing great primary science, as well as startups. But the way the majority of the world works, I would still think, would be in funding academia, which oftentimes comes in the forms of grants. Then there is the commercialization stage funding, which is kind of the bridge where you've got a partnership between academia and industry for some sort of commercialization, with the two working together. Then there's the return on investment models. So, this would be kind of debt-based financing, where you might get a grant that has some requirement to pay back the capital plus a little bit of interest. And then, of course, there's another area that I'm particularly interested in, which is venture capital and private equity, where you're actually going to commercialize a product fully and sell a piece of a company in order to realize its value. So, let's briefly touch on each of these if I can.

Louis: Sure.

Rowland: So, starting in academia, right, this in theory provides one of the avenues for the greatest degree of research freedom, where academics could do primary research on fundamental problems without having to be focused on some near-term milestone of translation of that science or commercialization of that science. This is just science for the sake of science. And that, if there's no necessarily immediate material realized gain, is primarily funded through the government, honestly, or individual institutions like the American Heart Association or others, right. And the way that works is through grants. These are federal grants delivered through individual agencies. So, as part of the congressional budget, Congress will pass certain amounts of taxpayer-funded dollars that can go to the National Institute of Health, the NIH, the National Science Foundation, NSF, or defense-specific organizations. And based off of where they put money allocations is where those grants can fund research in those areas, right.

The NIH is divided up into several institutes. The biggest would be the National Cancer Institute (NCI), but there are all sorts of institutes for pretty much any amalgam of interesting processes. There's one for biomedical engineering, one for the National Heart, Lung, and Blood Institute (NHLBI), and so on. The idea is taxpayer funding goes to these institutes, goes to the NIH or NSF, goes to one of the institutes. And then those institutes receive grant applications from academia saying, "Hey, I would like to pursue research in this space." A review process is done to assess the merit of the grant and what it's proposing. The grant is scored, and in this context, lower is better, similar to golf. And then if your grant scores lower than a certain threshold, let's say in the 20s or so, then it is considered acceptable or suitable for funding. It meets the funding payline, and then funding is distributed to that academic institution. That's a long process.

Usually, grants are reviewed three times a year, and it takes almost a year for them to get executed. So, this is a long lead time. It is a significant amount of money. For instance, a primary grant, an R01 grant, could be a multimillion-dollar grant to a Principal Investigator (PI) for three to five years for a specific research initiative. The last point I want to make before we move on is that this involves a large sum of money with relatively little gatekeeping in terms of outcomes. Anything accomplished research-wise must be attributed in publications to the funding source, but there's no obligation to repay the funds. It's purely an investment in primary research, funding and driving the research institutes we see today. When you receive an R01 grant, for example, a $3 million grant supports your lab. What is less discussed but true is that a portion of that grant, on top of the awarded funds, goes to fund the institute itself for providing access and administration for researchers. So, if you bring in a $3 million grant, about 60% to 70% of that, depending on the institution, goes as a direct payment to the institution.

Baylor College of Medicine, my former institution, which I love, brings in over $400 million in federal funding grants every year, which funds their labs. But then a certain percentage of that goes directly to the institution itself. This budget item allows it to run and enables our academic institutions to be interested in primary research, blue-sky research that doesn't have a material gain immediately accessible at the end of it. That's part one.

That would be funding in academia with grants. Next would be commercialization stage fundings. There's a bit of a bridge here where we're still in the world of grants that don't require repayment. They're still primary investments in science without any debt or equity-based commitment. These are what you might know as STTR and SBIR grants, which are technology transfer grants, etc. This is usually where industry and academia have partnered up. Typically, industry leads these grants, and I've been part of several of these. We've raised a couple million dollars through this avenue. The idea is you're saying, there's some technology that's worth pursuing and has clear market potential. So, you pitch to the same organizations—still taxpayer-funded dollars, still the NIH, the NSF, or others. You say, we've got this technology that is possibly housed in a university based on their IP portfolio, and we want to take it to market. So, we write a grant with a similar structure, you still have your six-page research strategy and aims, but you also add a six-page market commercialization strategy saying, here's how we're going to bring it to market.

And if you win one of these grants, they have phase one grants, which are smaller grants, typically around $250k, for prototyping or initially proving out your thesis. These can enable phase two grants, which would allow you to fully commercialize the product. This avenue is exciting because it provides opportunities for small startup companies to pay big academic research institutions for access to either their technology or some of their researchers on a part-time, grant-funded basis to commercialize this together. We've leveraged this several times. Notably, there's no return necessarily expected for the government. When you receive one of these grants, it's not free money; there's definitely reporting necessary, and you have to meet your milestones and do what you said you would do. But there's no interest on this payment; you don't have to pay it back. So, it's a very effective vehicle for startups when they're just getting started to work with or out-license technology from universities. That's a really cool commercialization stage funding through STTR, SBIR grants.

Then one level up from that, in my mind, I think of it as return on investment models. This can be thought of as almost like debt-based financing. So, debt-based financing would be, you go to a bank, get a loan, and you have to pay back that loan one day with some interest on top of it. That's the model we're working with. I've worked in industry and consulted for a couple of companies, specifically for a company called InformAI, and we've used this before through the Cancer Prevention Research Institute of Texas, or CPRIT. CPRIT is a kind of state local government-based organization. They provide commercialization grants or what they call product development research grants in a similar vein. In this case, it's still a partnership between industry and academia, with industry leading, and CPRIT or whatever organization will fund that grant with some sort of return expectation on the back end. These are still very favorable to the operators in the sense that the expectation is revenue-based. So, the investment you get, let's say it's a couple million dollars, is an investment for you to take technology to the market. And then, when you achieve product market fit and start to get sales, there is a percentage of your sales revenue that you'll return to the institution until you meet some threshold. Let's say you'd give 2% of your revenue indefinitely or however you want to frame it, until you pay back your initial capital, that $2 million investment or whatever, plus some bonus, maybe it's a multiple of two.

So, you end up repaying the principal and an additional amount back to the institute. This financing method is still favorable because as a company, you're not taking on material debt in the sense that you have to repay regardless of your financial success; repayment is contingent upon generating revenue. Additionally, it's fixed in time; you repay a certain principal plus a bonus, and then you're done. The institution does not appear on your capitalization table, they don't own a portion of your company, nor do they have rights to your revenue indefinitely. It's an effective vehicle to encourage investment in an early startup, positioned in the debt-based financing category. It's one step up from a grant with no commercial return expectation, yet it remains favorable by not involving equity-based investing.

And then finally, and this is the one everyone likes to talk about, and it's one that I'm very interested in and enjoy participating in, is the world of venture capital and private equity. This is a very specific mechanism for funding science and comes with certain constraints. The idea here is if you have a product that you believe is making a substantial change, one that can drive serious market returns, not just a return of principal but potentially 10 times the principal return, then you would attract venture capital investors who are interested in partnering with you on product development. The concept is, if you want to commercialize quickly or if you want a larger amount of resources upfront to develop your product, you might partner with a VC firm.

As I briefly described with the debt-based financing approach, taking on venture capital dollars means you're not selling a portion of your future revenue. It's not a buyout where you give them 2% until you've returned their capital plus some interest. When you partner with a VC firm, what you're essentially doing is forming a company that owns the IP or similar assets, and you're selling a piece of that company, along with all its future revenue potential, as equity to an institutional investor for a price. You might say, "I've got this innovative technology that's going to enable XYZ. We are a pre-seed or seed-stage company, meaning we're early in the process. So, we'd be willing to give you 20% of our company for $2 to $5 million." The idea then is that you are partnering. This usually involves the institutional investor taking a board seat in your company or participating in other ways. The main idea is that you have entered into a partnership that will last until there is some liquidation event, AKA when another company buys you out, buying their ownership percentage, or you have an initial public offering (IPO), and the public buys out the shares of your company. This structure is very attractive because it enables significant change and allows for rapid and efficient scaling. However, it can comes with substantial constraints: you're giving away a piece of your company and its future success in exchange for capital now, while also gaining the partnership of the institutional investor, with all they can offer in terms of their network, experience, or ability to help shape your commercialization strategy.

So, that was maybe a longer answer to your question, but I do think it's important to think through. You've got blue sky research or commercialization research in academia or private biotechs or others that get funding in the form of grants. There's a kind of bridge in between where you've got commercialization funding without constraints in STTR, SBIRs. You've got debt-based financing with product development research that has some return on the initial capital, plus maybe a multiple. And then you've got, on the final other end of the spectrum, venture capital or private equity investment where you sell a piece of your company and its future potential for that capital allocation now. So, that is how I think of the world in terms of how research is funded in science, technology, engineering, etc. But we'd love to see if I've missed anything. So, feel free to reach out, but that's the way I look at it.

Louis: That was fascinating. I think you distilled it down well. There is a lot packed into that question, but the way you distilled it down, and I think you allocated an appropriate amount to each of those topics. And if you missed anything, I wouldn't know. You're clearly very knowledgeable about this. I thoroughly enjoyed learning about it, and I'm sure the audience will as well. Let's take the venture capital side, since that's what you primarily focus on now. We'll do a little example here. So, let's say there's a private company developing technology considered to be the next frontier of medicine. The company is pitching to investors in the hopes of raising money for its next stage of development. What are investors' thought processes to determine whether or not to participate in the fundraising round?

Rowland: This is a great question, and it's one that I think my time at business school helped me learn initially. You could certainly learn this on your own, but one of the things that I did think came out of my MBA was having a structured thought process and getting to learn from other VCs not only in the life sciences space but also in consumer, tech, and other different areas of investment. How do they structure their thought when dealing with companies at different stages with different layers of risk? That's kind of the answer there.

So, I think, not to bleed into the next question, but the general idea is the way you think about this. I'm still a bit green to the process, having had a couple of years of experience now, for which I'm very thankful. But from what I've learned so far, the approach you take is to try to think about the seed. What is the stage of the company? Is this just the first pitch deck, where you've got a scientist who's been a postdoc for a couple of years and they're literally presenting to you a science-based pitch deck? Or is this a company that already has product-market fit, with initial customers and a balance sheet showing cash flow that you can evaluate? In general, the way we think about the world is, what stage? And there are different investors who like to play at different stages. There are people who feel like they're really good at, or that they prefer to think about, primarily calling the shots early, trying to understand what is the incremental change directly just from the science itself, the early, the pre-seed companies.

And then there are other institutional investors who might have a bit more business savvy. They know about the market and prefer to think about companies that are more mature and de-risked, at later stages of development. But coming back to this question you're asking about how you would evaluate a company that's pitching to you as an investor, I think the best model to consider is something I learned during my time at KDT Ventures, an excellent VC firm based out of Austin, Texas. Particularly from my friend Patrick Malone there, he really likes to think about layers of risk. The idea is that several different risks can present themselves when looking at a company. It's okay to have a couple of different layers of risk, but if the layers of risk stack up to the point where there are many, then at that point, it might not be worth investing in them. Does that make sense? I can go into the different layers of risk we like to think about, but in general, you try to see which area of risk you're willing to tolerate. If there are too many, then you might not pursue the company at that time.

Louis: Right. The layer of risk determines the price you're willing to attribute to the company and the terms. This does lead into our next question, which considers the differences between early-stage companies and late-stage companies. You explained it well enough; early on, there's a lot of risk involved with pre-seed and seed companies. Later on, it turns into more of a business approach, analyzing the balance sheet and looking at maybe more long-term outcomes of the business, as opposed to taking a calculated chance on an early-stage company with their technology. I would say, just maybe briefly, have some concluding thoughts on the layers of risk you see in an early-stage versus a late-stage company, and then we can move on to the next question.

Rowland: Yeah, sure. That's a great point. So, let's identify the layers of risk very briefly. There are certainly more than what I'm going to list here. Then, you could think about how an early-stage investor might consider those versus a later-stage investor. As a physician-scientist, the first thing that I like to think about, whether it's the best thing to think about first or not is up for debate. But the first thing I like to consider is the frontier science itself. Is this technology going to create an incremental change or some sort of transformative change?

I think one of the first books a lot of people read when they're interested in investing is Zero to One. This is Peter Thiel's book that offers some pretty good framing for what you might be looking at. Is this science going to just build upon additional technologies, or is it going to open up a new avenue for technology application? If it's the latter, that's where venture capitalists really get excited because there's plenty of white space to work with. Outside of that, is this an incremental or a transformative change? There are, of course, the layers of risk. Some people like to think about Porter's Five Forces. You can look that up if you're interested. It's a business school concept where you could think about competitors, new entrants, etc. But peeling it back, let's think about market viability and potential. Is there a market for this?

There's plenty of great science and many projects that could lead to small companies making founders a lot of money. But is there a big enough market to sustain venture-style returns? People think of this differently, but if you're going to be a seed-stage investor and you're going to put in a $2 million check, are you going to get out seven, eight, 10 times that amount when this company liquidates down the line in terms of either an acquisition or an IPO? If you can show that you're going to get these enormous returns, and some people think even bigger. Some argue that if you're going to invest capital into a fund, that investment should be able to return the entire amount of your fund itself. That's a bit extreme, I think, but some really feel strongly about it. The idea is whether there's a big enough market to have venture-style returns. Because if you can't sustain venture-style returns, even if it's really interesting technology, it's not going to be a VC-based investment. It might be a good candidate for debt financing or a bootstrapped investment, where you can build it yourself and still have a very successful career as a founder. But market size is a thresholding risk factor. Number two, so in the way I think about it, is incremental change, then just thresholding, is there really a market here? And then three, and I weigh this pretty highly, is the team.

Louis: Definitely.

Rowland: Is this the team? Even if it's a killer idea, is this the team that's going to execute on it? Sometimes I'm still growing in my experience in the venture world, but learning from the experiences of others and some I've seen myself, you might have a really good idea and maybe not quite the team you want behind it. Conversely, you might have a fantastic team with a good idea that could be refined into a great one. Having a killer team, where repeat founders or those able to demonstrate strong networking skills or other abilities, can really help mitigate your layers of risk. Is this going to be a winning team? This is crucial if we focus on the frontier science-based approach and medical applications, where there's a lot of emphasis on product risk and regulatory risk. If you're building a drug, is this change worth going through the 10-year product development pipeline with multiple phases of clinical trials? Is that really what we anticipate coming out of this, or is it not going to work? Is this a product that must go through the FDA without a predicate, potentially facing difficulties in approval? Technology and regulatory risk is another consideration. Lastly, the competitive landscape is worth mentioning, assessing how crowded the field is and the company's potential to stand out.

I think there's certainly a fear of missing out, and even if you're aware of it, it's hard not to let it influence you when thinking about a company for investment. But what is the competitive landscape? Who's investing in these areas, and what do their companies look like? I personally like to think of, I don't know if you've seen the image from one of the late 2000s Olympics where Michael Phelps and his competitor are swimming forward, and just because the competitor looked straight next to him, he took a second to look to the side, and Phelps ended up winning. I'm sure there are plenty of examples of the opposite, but I do think about that. Just because there are competitors in your field doesn't mean that this can't be the winning take, especially if sometimes being second to market may have an advantage in terms of not having to do all the legwork of regulatory compliance and otherwise, and they can just get to work.

There are problems with that, but it's certainly worth appreciating the competitive landscape. Who has what? What is the funding stage of these companies? What are they working on? Where do we fit in, and where can we win? Just taking a step back, is this an incremental change? Is the market there? Is this the winning team? What are the layers of risk for technology development, regulatory development, and then who else is out there in the competition? That's a pretty good initial take for the layers of risk. Now, very briefly, in terms of the stage of development, as we mentioned, if this is a... And just to clarify for all readers or listeners here, when we think about the stages of development, when you first start a company, if it's just you and your friends or your co-founders, and you put a pitch deck together, that's often what's termed as a pre-seed investment, a pre-seed company, a pitch deck, and an idea, right? Once you maybe have a bit of technology or you've de-risked it in some specific way, then you might have a seed-stage company where you're saying, "Hey, invest in this technology, this IP, this team, etc." The stages build to where people think of Series A, depending on what kind of field you operate in.

A Series A is where you have a product and some initial product-market fit that you need funding to fuel, driving forward to gain additional customer traction or to enable whatever next stage studies you want to go through. It builds from there, according to the alphabet, with Series A, B, C, D, and those later stages, Series D, E, and F, are seen most in tech-based companies or some consumer-based companies where you have, for example, Salesforce-type companies with a B2B SaaS model. That's where you end up having these indefinite private financings into the Series D, E, and F later before you get to an IPO because you're waiting for the right time to take it to the public market and liquidate all those previous investments.

So with that framework in mind, from seed to Series whatever before a public offering, the early-stage investors pretty much operate in the pre-seed to Series A, some can reach into the Series B, but mainly pre-seed to Series A. These are hundred-thousand-dollar checks to maybe a couple million, to maybe tens of million-dollar checks. That's kind of the seed stage investors. And then the late-stage investors, the Series B and on, these are usually multiple millions of dollars in checks that you're investing in later-stage companies to really just fuel their actual business execution. So, with that in mind, the appetite of an early-stage investor is going to be totally different from the appetite of a late-stage investor. So, at the early stage, you're really thinking about the team. You've got some tech that you think is enabling, but really all you're betting on is that this is the right team that's going to kick down the brick wall and bring this to something.

So, in early-stage investment, you're really thinking about the team and more about the market potential versus the current market fit. There's no way that you can conduct a discounted cash flow or try to estimate potential earnings. That's all somewhat nebulous at this stage. So, you're just trying to scope the market and say, "Hey, this is the market for radiation oncology delivery in the US. Here are the product codes. Here's what we can potentially see happening in terms of if you brought this to market and captured X percent of that market, there would be enough money to value the company high enough that we would get the returns we want down the line." It's kind of back-of-the-envelope. You can use market comparables. You could say, "Hey, in biotech, that's a really good example. These stage companies with these single assets are able to generate X upon whatever." And then you can gauge or estimate or even do some modeling, but often you're thinking about the overall market rather than market fit at the early stage. And then the other things would be early-stage investors really have to be able to sit with their investments for a longer period of time, especially in the life sciences, especially in biotech, where you're going to have to go through multiple stages of clinical trials, etc. Early-stage investors, especially in the life sciences, usually think in terms of seven, eight, or even ten-year time horizons to get a return on their investment, while later-stage investors may have the opposite perspective.

Then there's the regulatory pathway uncertainty. Certainly, at the early stage, there's a lot you have to figure out. You may need to figure out how to license and take the IP out of a university. You have to figure out how to get it regulatory approved. These are layers of risk that an early-stage investor might be more willing to take on. Flip that to later-stage investment. Later-stage investors don't really want any of that. At that stage, if you're a Series B or C company, you've probably seen several iterations of that individual team, and as the company is worth maybe a couple hundred million dollars on paper, they've been able to attract and recruit top senior executives. Perhaps the CEO has made it the whole time, and as a founder-friendly investor, that's my goal for as long as that seems fit. But certainly, some of the other executive roles, there's really room for key strategic hires. You've hired that Chief Scientific Officer (CSO) who's worked at an X stage company. You've basically de-risked the team to where anybody would back that team when you're at a later-stage company. At that point, later-stage investors are also looking for detailed financials and projections.

This is where the MBA-type learnings really kicks in in terms of detailed financial analysis. Looking at revenue, examining burn rate, and identifying areas for optimization are concerns of the later-stage investors, as I've seen it. I've never been a later-stage investor. This is just my understanding of it from business school and from being around the industry, trying to prepare companies to go through those later-stage investments. But later-stage investors are looking for a shorter time horizon. They want a clearer exit strategy, so if they're going to put in tens of millions of dollars, they might not be looking for a 10X return at that point. But if in two to three years, you could get a two to three X, that would still be very appreciable and meaningful for their firms. Later-stage investment is traditionally where you see some of the big private equity firms starting to participate more so than just the early-stage institutional VC investors.

If I had to break it down very briefly, early-stage investors are really focused on the team, market potential, and the feasibility of the tech. They have a higher risk tolerance and are looking at longer time horizons. Whereas later-stage investments involve putting in a lot more money. They want to see a top-tier team with proven product-market fit, products in the market generating appreciable and predictable revenue that can be clearly built from and could scale. They want to see detailed financial projections and clear strategies for exiting, clear plans for an IPO in the coming years to months, or a clear ability for acquisitions or partnerships, etc. So, that's how I think about it in terms of the late versus early-stage investor base and how they might approach layers of risk. I've already mentioned the disclaimer, but I've not yet been a late-stage investor. I think that'd be a lot of fun one day, but that's just the way I see it as of now.

Louis: I'm glad we dwelled on that question because you had a wealth of insight to share on what would perhaps be considered a nuanced question. It is very broad, and there's a lot to speak to. Some of the points you made are intrinsically important to the field. There's just a lot to understand as a whole about it, and I think you very accurately summed it up. That's a challenging task to do for such a complicated field with so many different variables and considerations. It's very exciting to listen to that.

Rowland: Yeah, thanks. I mean, it's a fascinating one, and it's one of the reasons why I'm so interested in focusing on continuing to work in ventures because it's exciting. Just as a quick plug there, I mean, the whole process is important, but probably and arguably, when done well and within the bounds of appropriate development, venture-based investing can be one of the quickest and most meaningful ways to bring meaningful technologies to market, to patients quickly, and at scale.

As a physician-scientist, what really drives me is the desire to develop drugs to help patients, and the way that can be done quickly and safely is through venture-based investing. Then, through venture-based investing and subsequent pharma acquisitions, et cetera, you can bring innovative new machine learning, deep learning technologies for diagnostics, or any of the hosts to market. And we'll talk about that later, so I don't want to preview that too much, but…

Louis: It is the perfect segue.

Rowland: Right. So that's why I'm excited about this space and love getting to be a part of it, even in a small way.

Louis: So there are two sides to this coin of innovation, and like I said, I think this is a perfect segue. Let's shift from the business side and the finance powering the innovation to the science behind it. As a physician-scientist, you are involved in cutting-edge research, particularly in bioinformatics and artificial intelligence. Could you share an overview of those fields and your current work within them?

Rowland: Yeah, absolutely. This is something I'm very passionate about. For context, as you mentioned, I did my PhD in Quantitative and Computational Biosciences. This was focused on several different subfields. Right. So I got to touch on and be meaningfully a part of bioinformatics, computational biology, machine learning, and deep learning. And particularly, my PhD thesis focus was on the world of statistical genetics, genetics, etc. But maybe just to take a step aside here, this is, I think, the cutting edge in terms of what will meaningfully drive change in the life science and biotech ecosystem for the near future. And anybody that's listening to this who is interested in a STEM field, I think, has to absolutely take serious consideration of acquiring this skill set. I'll take a quick aside here.

When I was a medical student, I initially joined as an MD-only student and then applied and was able to join the PhD part to do the MD-PhD training. But when I was considering doing this, it was honestly kind of scary. At the time, computational biology and bioinformatics were big, bad math and coding-based skills, which I hadn't really touched in a long time. I studied biophysics at Duke and had taken an intro to programming ‘Python 101’ type course at Duke. I also had a bit of a math background; I had done linear algebra and all that through my biophysics. But that was, I guess, five years ago at that point when I was applying to the PhD program. So, it was kind of scary. And I'm just trying to think about my reflections there. And if I could encourage anybody looking at that field, it was unbelievably worth it. My journey was that I had to put in a bit of elbow grease, learn how to code, learn statistics, learn these bioinformatics pipelines, and physics-based approaches to understanding protein folding or whatever. All of that was fascinating and a bit of an uphill battle, but very exciting and totally worth the time spent. It has enabled me now to sit at the seat of being able to utilize the top technological advances for anything I want.

I'm currently getting to look at, certainly through some of my informatics work, applying deep learning for all sorts of very interesting health predictive metrics. So, optimizing in the field of radiation oncology, trying to be able to look at CT scans to maximize radiation to a tumor while fully treating it and minimizing radiation to off-target effects. This involves 3D image-based processing that I get to do. And at the same time, I can think in the world of oligonucleotide therapeutics and about protein folding and protein folding dynamics, using deep learning and transformer models to predict outcomes there, particularly in understanding protein stability, protein folding conformations, protein binding, etc. These are areas that are so exciting and meaningful in terms of building applications for patient care that if anybody's interested in science or medicine, I have to encourage it. The thing I'll add is that it is more accessible than ever.

And this is kind of a double-edged sword, I will say. I spent a lot of time throughout my PhD and medical school learning how to be an excellent technical writer, being able to convey my thoughts very clearly for diverse audiences, etc. I also spent a lot of time learning how to code and understanding the depth behind statistical analyses and deep learning-based fundamentals from a math perspective. That is totally accessible to you now. And I have to stress this: if you are interested in one of these fields, you've got your own personal tutors. I mean, you can go on ChatGPT or Perplexity and just say, "teach me to code," "teach me to implement this biostats package." Anything you're interested in doing, you've got your own personal tutor to where this is a much more accessible field. And I would encourage anybody, even without a math or physics-based background like I had, to learn about it. Of course, you'll be responsible and need to understand that, but you can learn it in a much easier way. Several examples of that I still use: if there's a biostats piece of code that I'm trying to understand, now in R you can use the `getAnywhere` kind of code prompt and get the whole package, put it into ChatGPT and say, "line by line, teach me what's going on here." It's very effective for education and understanding the statistics, the approach.

And so I'll step off my soapbox here, but I do want to say that being involved in cutting-edge science research and venture companies is very satisfying, especially with the idea of ultimately bringing some of these products to patients. But I must emphasize that it is totally accessible to really anyone at this point. And anybody interested in STEM should consider computational biology and bioinformatics as very exciting fields. They are not some big, bad, scary thing. You should learn about them, and it will be important for future growth and development.

Louis: If I could jump in there real quick, I completely agree. I think one of the superpowers of these technologies is not just the science and outcomes it brings but also how it will empower and democratize this previously higher institution technology to all sorts of people, as you mentioned. So, I think it's perfectly reasonable and our duty as scientists and engineers to talk about the positive implications that it's going to have for all sorts of people. So, I'll give it back to you to finish your thought there.

Rowland: No, yeah. I mean, we feel the exact same way, and this is why I love being your friend. But yeah, so I have this background. I want to say that double-edged sword, now my skill set is slightly less unique because anyone can do it, but I'm still very proud of it and understand the foundation behind it. But I also want to encourage the next wave of scientists to go out and try to do the same thing because I do view that this is going to be where a lot of the most exciting innovations are currently happening and will happen in the next 10 years for innovation. So, just a point of encouragement there.

But the specific question was an overview of those fields and the current work within them. So, I think we touched on it a little bit, but briefly, those words that I put out there, computational biology is really focused on a physics-based approach, but certainly with deep learning-based approaches as well, trying to understand, oftentimes, the cellular world and how it interacts. How do proteins go from a string of amino acids and fold up into a meaningful little entity that can enact change in its environment? So, that's kind of the world of computational biology. And there's a structural component with imaging like bioelectron microscopy and others. That's very interesting, but there are also predictive components with deep learning or physics-based models, etc. Bioinformatics is really the coding side of that where you try to understand how you can computationally enact things. There's certainly a pipeline-based aspect there where you could take raw genetic data you might be sequencing or have other sorts of genetic input and figure out how to meaningfully process and quality control that data and then do informatics on the back end. That's the world of informatics. That's one of the aspects that I really, really love. I love the coding to outcome or prediction-based approaches. That's a lot of fun.

There's the AI/ML piece. I've got a couple of papers out on that, but machine learning, deep learning, these would be the ideas of... That's probably too big for me to even summarize here, but a lot of it can be predictive. Trying to use deep learning-based approaches to predict outcomes from tabular data, like Excel spreadsheet-type data, to predict certain variables' outcomes. It can be image-based, where you're trying to understand diagnostics or focus on certain points or understand relationships within an image, or it can be way more interesting and abstract than that. There are really cool architectures out there called transformer-based models. There are several good YouTube series that I could maybe link or post here, but just understanding how you can take multiple areas of unstructured data and encode it and decode it and then be able to learn from representations. Its fascinating.

Then finally, there's statistical genetics. This involves applying statistics to robustly test genetic data for association or causal inference, etc., to understand how genetics, epigenetics, protein folding, etc., is associated with, or even can be statistically causally linked with outcomes of interest. I went through that kind of fast, but the basic idea of computational biosciences is the focus. It's really the computational side of understanding how physics, chemistry, and biology apply to health, human disease, agricultural science, etc. That's how I would define those fields.

Louis: Okay. All right. You clearly have a lot of ideas about this space, and your formal education certainly has powered that. For example, you recently gave a fascinating TED-style talk about organ transplant decision-making processes. Could you share your process for contemplating ideas and preparing talking points that resonate with diverse audiences?

Rowland: Yeah, sure. Thanks for looking at my LinkedIn and finding that talk. It was a lot of fun to give. For context, one of the projects I've worked on pretty extensively, and one of the ideas that came out of medical school that I pursued and initially pitched, and got the STTR commercialization grants for, is for improving informatics within transplantation. The one-liner there is that there's a lot of really high-quality data that clinicians are trying to synthesize on their own as individual data points, and they could be meaningfully integrated into robust, accurate outcomes-based variables for consideration at the time of transplant.

If we could predict waitlist mortality more accurately, predict transplantation graft survival outcomes, predict length of stay, and predict the risk or the likelihood that an organ will be discarded, then that information could perhaps be helpful, or serve as a high-quality granular decision metric. Knowing these expected values (and the confidence surrounding these estimates) would be valuable for a clinician to appreciate when they're trying to consider organs offers and potential pairing with recipients.

That's a really fun problem. It's a very complicated problem that we've been working on for several years, but I love getting to work on it. The basic idea here is I was invited to give a TED-style talk, not TED itself but a TED-style talk, to the main transplantation conference this past May. Part of that involved hiring a coach to help me prepare for this talk, which was an unbelievable experience. They hired a coach who specializes in TED-style coaching to help guide me through what that process might look like. There are a few pieces of feedback that I'd love to share here, just for anybody trying to prepare a talk.

The first thing was I had to think about who my audience was. I've delved into the technical details of this problem, talked about it with friends, family, and neighbors, and so I've kind of, over the years, gotten a sense of what resonates with different people and what they are interested in. But the first thing whenever you're going to talk is to try to think about who your audience is. For this podcast, I was trying to think about this probably being a lot of individuals interested in or already in STEM fields, thinking about commercialization, etc. I've hopefully tried to tailor that to this particular audience. That's the first piece of advice: just try to think about who you are talking to, what they understand, and what they don't understand.

I truly believe that there's not that big of a difference in anyone's intelligence whom you're really talking to, so it's really just about getting people up to speed and trying to help them quickly get through the key points of information so that they can be at the same understanding and then think through rationally what might come next. So when I am contemplating ideas and preparing talking points that would resonate, you try to think who your audience is. Second, you've got to start with a story, right? If you know, I didn't here I guess I told you a little bit about my story, but if you're going to try to draw somebody in, you want them to relate, so either a personal story about you, in this case, you know, I think what I focused on, you can watch that talk if you want, was just trying to understand what my background is, why I'm particularly interested in the problem of transplant informatics, why I think that could drive incremental change, and why I'm personally invested in it.

You know, these pieces of information should be intentionally thought about and conveyed very simply, right? The other thing would be to use analogies, right? A really good analogy can totally drive your point home. In the case of transplant informatics, we settled on the analogy of Google Maps. So the idea would be saying, hey, you know, we didn't really know we needed Maps or Google Maps or whatever you want to use, but as soon as we had it on our iPhones, you know, people love using it, right? It didn't stop us from charting out our own course in our heads, but it gave us real-time information about what different courses might look like in terms of time to get there, traffic, problems along the way, and it would be updated in real-time, right? If new information came to the table, it could give you a new route that you might not have thought of before because that might be the most appropriate route given the different considerations, right?

The other piece of information, or analogy, that was helpful, was that you know you still stay in the driver's seat. Map applications are note driving you there or picking your route; it is just giving you the most up-to-date real-time information for your consideration in your decision-making to get from A to B, right? This was the present analogy that we thought would resonate really well with the clinicians in the room because they would be able to think, okay, providing an information dashboard with high-quality, granular decision metrics that integrate all the data available would be helpful to understand what's the best path from A to B, which organ goes to which recipient, while it still keeps them in the driver's seat of making the ultimate decisions and provides insight into how those decisions might be made.

So, that might be delving a little too far into the specifics of that particular presentation, but the idea holds true: a really good analogy can help bring a diverse audience with different technical backgrounds to the same place in terms of understanding your problem and why you're interested in it. Making it clear why you are motivated is a key differentiator, also in venture. You want people to understand that you're a missionary, not a mercenary. You would work on this problem for free, even if it was just you, because you want to see it come to market, whether there's a commercial incentive or not. That's a much more powerful, and hopefully true—I mean, in these cases, everything I've worked on, that's how I feel—but it's an important selling point, really, to be able to say, "Hey, I'm not being paid to do this. I'm not just collecting my paycheck because I get to work on this problem that someone else thinks is good. I am driving this. I want to see this come to market. I want to see this happen for XYZ, right?" That's the second thing you want to convey when you're doing public speaking.

The last thing, and this is something I haven't followed at all today, is you've got to speak slowly. You know, I've had a lot I wanted to condense into this, and so hopefully, I didn't speak too fast today. But when you're giving a formal presentation to an audience, right, it is never a problem to have a pregnant pause, to speak slowly, and to let people think through what you're going through and what you're presenting. So, speak slowly, think about your audience, have a story or analogy—those are the things I think about when I'm doing public speaking, and that's what this coach was able to imbue in me—so I really have to give credit to her. That was excellent.

Louis: Thanks for sharing those, Rowland. Those are certainly tips I think we all can implement – I certainly will. I really liked the analogy, how that can drive home, and the Google Maps one was really good. Like when you said that in the talk, I immediately grasped it, and I think that's an excellent way of helping a diverse audience understand the point you're talking about. And you seem to be talking at a fine speed for me. I mean, this is complicated material that I may be more familiar with than my audience, but you don't seem to have much trouble with the words coming out. But is there ever a time when you experience difficulty putting your ideas into words, and is there a structured or creative process you follow to break through writer's block?

Rowland: Yeah, I think this problem has been solved again in 2020 — I mean 2022 — with ChatGPT. I'm not going to lie. So, what I do view, that writer's block, at least for me, in my experience, it's not so much that I don't have ideas. The problem is when I write down an idea and then immediately start to try to edit it, then I forget the next ideas, right? And so, what I like to do, whenever I have to really do anything, right — applications to med school, PhD, MBA school, right — applicable, you know, residency applications. When I am thinking about writing a grant, particularly for grant writing, right, or when I'm trying to do like an investment memo for a company, or if I'm trying to think through, you know, friendly but polite criticisms, right — if you, if you of companies, right — if you just try to start writing, you're not going to like your writing tone. You're not going to like your style. You're going to be thinking of ways that you could say things more politely, or more friendly, or more warmly, right? And you're going to get stuck.

And so, the one-liner is like — and this is kind of cheesy, but this is what I do — I put on dictate, you know, either on my phone or, you know, I have a little hotkey on my computer. I turn on the dictation, open up a Word document, I just close my eyes and I just try to answer the question, right? I try to write it all out, and I just kind of word vomit. I don't care about grammar. I don't care about structure. I just do it. And then, ever since 2022, I just copy that into ChatGPT and say, 'Structure my thoughts,' right? I just literally say, 'Structure these thoughts,' or 'Edit for grammar, edit minimally, edit for clarity and content,' you know, whatever it is. ChatGPT, Perplexity — those are the two that I kind of like. Bard is getting good now, too. So, just copy it in and edit it, and then boom, it comes back with your raw output, now in some structured way.

And sometimes the way they structure it, I like, and I think, okay, that is good. I can mention this point and this point and this point. Sometimes I hate it, and I have to iterate. But you have your iteration tool. Previously, I relied on friends, family, and parents to do this, where I would send people texts and just bother the heck out of them, saying like, 'Hey, can you edit this email? Can you edit this paragraph? You know, I've got this grant, can you look at this, whatever?' And that was high-quality feedback, but it took time, right? I could sit there with Bard, Perplexity, ChatGPT, and just edit for hours and just say, 'Iterate, iterate, iterate.' And so, that's what I think is the key to writer's block: close your eyes, hit dictate, word vomit, right? And then say, 'Structure my thoughts,' and then go from there. And then you've got, you've got stuff on the page. It's much easier to write when you've got stuff on the page because then you're editing, you're not creating new content. So that's my thing. I would hope to share, and I think that it works for me, and I hope it can work for some other people too.

Louis: I love it. I'm personally a huge fan of dictation, as you know, and at Rice Ignite last year, I probably mentioned it to you. I really think there's a power of dictation that we haven't unlocked yet, you know. The idea of talking to yourself, I think, is very powerful, and I encourage everyone to try it, as you just did. Moreover, I remember last year you said that you use ChatGPT, or you know, other, they have now come to market, but other sorts of transformer-based large language models, you used it every day, and this is a very powerful tool that I think everyone could at least benefit from trying out. I wholeheartedly agree, I think that's an excellent answer.

Rowland: Yeah, the last thing I'd say there is, I think maintaining a healthy and active network is incredibly important. People and social capital are the best things that you can maintain, and they should be protected and also intentionally maintained. Every interaction you have with people is kind of building that, and so I don’t see why every email I write, no matter how small, I put in ChatGPT, or Perplexity, or Bard, and I just say, 'Edit for grammar and clarity,' or I say, 'Make it direct, polite, and warm,' just to be succinct but to have that little bit of sophistication. It's free. It's like, why not? It takes an extra 10 seconds, and all my emails are really polished, I think, and it just conveys a sense of sophistication and caring with who you're talking to, right? So, I do use it every day. I mean, I max out the limits on GPT-4 all the time just because I'm like, 'Edit this, edit this, edit this.' I just want to be constantly investing in that—not just social capital, but professionalism in my environment with other people. I want to convey that and give that respect. So, I do think that more and more people will incorporate that into their workflows very substantially.

Louis: Gut instinct or intuition often contrasts with calculated, well-thought-out decisions, yet both can hold equal value. How do you balance the use of gut instincts with analytical data in your decision-making process? Can you share an example where your intuition played a key role in a significant business decision?

Rowland: Yeah, this is a great question, and I think, you know, there are some very public figures who have made commentary on this effect. I did get my MBA, and I think it's a very valuable degree, but there are several notable people who kind of don't favor MBAs because they think they bring too much structure or analytical processes to the table where it might not be justified. Not everything can have a discounted cash flow, and I see that, and I think that's valid, but the point is maintained, right?

When you are looking at tech that isn't well fleshed out, it's kind of inappropriate. It might bias your decision-making to try to think too granularly about it. This is kind of why I like early-stage investing, as opposed to, so far—I mean, I haven't done later-stage investing—but I like early-stage investing because you get to be a bit of an optimist. You get to believe, see, try to dream the glimmer of greatness for a new technology, and try to think through, 'Can this hold innovative, or incremental, more than incremental, change to be transformative?' That's fun, and sometimes that's more gut than anything else. You have to bring data to the table and defend your position, but some of that is gut-driven, and so I think those can be held in balance, and there's an appropriate use for that, to be an optimist and to weigh data manually rather than just objectively weigh data as it's coming in.

In terms of your own personal kind of higher organization, an example of that was with InformAI. We get a lot of different pitches in terms of companies or tech that people want to commercialize. The idea is, we're looking at informatics in healthcare, trying to use deep learning, machine learning to enable high-quality efficiency, quality, safety, and standardization practices in clinical care. That's a very niche product, and unless there's a market to drive it, there are so many different interesting technologies that aren't necessarily venture backable, venture sustainable, and we have to be a little judicious with our time in terms of what we have time to work on and what we don’t.

So again, as I thought about it, the other side, as I thought about as an investor with the layers of risk, so too it applies as an operator trying to think about, 'Okay, early-stage company, you don't want to lose focus. You have to think about what actually has a market size to develop, and then of course, what's going to bring a transformative change in terms of patient care and quality of delivery of patient care.' So those are things we think about, but there are times where, even if the objective metrics aren't on the table, you do want to pursue something.

One of those was with some really early and interesting technology coming out of UT Southwestern in the form of deep learning in radiation oncology, and there were some great data points behind it. So, I'm not trying to say it was totally gut instinct, but there was a high-quality team that had put together some really good medical technology, as I've already previewed previously, for optimizing radiation delivery to a tumor while minimizing constraints to healthy tissue nearby. It just struck me in the gut as, 'This is really exciting.' So honestly, even if they had a lot of good data points at their disposal in terms of their implementation in the clinic, etc., but even if they hadn't, I was really wanting to chase that down either way, and it was truly more of a confirmation as I was learning about the market and learning about the ability to product develop in this space that was driving me. Of course, I was taking them on balance, but I have to admit that it was an exciting technology. I think it was good, and at any point along the way, if there had been a killer issue, we would have politely said we can't push this forward, but it is exciting technology and worth pursuing. But that was the initial gut feeling that kind of drove me to chase that down, at least in the initial stages.

Louis: Rowland, that was a fascinating answer and example of how you use gut instincts to power some of your business decisions. We're going to keep along with the philosophical, meta side of contemplation and decision-making here, and it's going to be our last question. Inspiration is often a guiding force in finding answers to questions. When do you seek out inspiration as opposed to charting your own course?

Rowland: This is a great question, Louis, and thanks so much for asking it. I mean, if we just take a step back, you know, there is really good importance to—there's definitely attention that should be paid to staying humble, right? Like, you can very quickly think an idea is good and chase it down in your head, and unless you're sampling your environment or soliciting feedback, you could be pursuing meaningless efforts. With regards to inspiration, I don't think in any way that it is necessarily fully self-obtained. I think ways that I try to sample my environment and seek information are several-fold. On the personal side, I do have a personal faith that I maintain, but outside of that, I really try to seek mentorship and maintain mentorship with people who are smarter than me. I think there are so many people—pretty much most people I meet—I think there is value to be had and learnings to be obtained. People have such rich experiences that it's kind of hard-hearted not to think that everybody you meet has something to teach you. So, going with that perspective to try to learn from people in your interactions is one that I try to hold, and I try to focus on.

In terms of seeking out inspiration versus charting your own course, I do try to maintain regular and intentional mentorship with people that I very much look up to, through following up on phone calls. I have a contact management system that I built in Notion just to try to—at least where I downloaded my LinkedIn contacts and imported them—you can do this; it's free. And then I at least tried to just filter and sort to where I've got a core set of individuals that I really want to keep in touch with, in all the different aspects of areas that I'm interested in, just to continue to learn from them and not be so focused on charting my own course that I miss those regular touchpoints.

On the flip side, I do think there is something to be said about being contrarian—a concept often discussed in the VC world. If everyone is interested in, knows about, or is doing something, then the ship has probably already sailed. There might be enough people working on it that it's not necessarily as interesting anymore. So, there is also something to be said for following your internal compass and having your own personal conviction about what you think might be interesting to pursue, even if you're not receiving positive signals from the outside.

So, a bit of a hand-wavy answer—I don't think there's going to be a definitive playbook there—but I would say that it's important to actively maintain a strong personal network of people you respect, so that you can solicit feedback quickly and efficiently and learn. But then also, try to think deeply and have some personal convictions because that seems to be where some of the best innovation happens as well. So, that's how I think about it, anyway.

Louis: Well done. I completely agree. I think a good takeaway is it's a healthy balance. We're recording this in January, so it's not too late to have a New Year's resolution. Exactly. I think it's a really great summary, Rowland. It was a pleasure to have you on. I will be listening to this again, once, maybe twice, just because of all the valuable insights you provided. So, thank you again.

Rowland: Absolutely, Louis, and I can't wait to hear yours. You're one of those guys I think is smarter than me, so I want to hear and learn from you here soon on the podcast. So you're next up!

Louis: Absolutely.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit rowlandpettit.substack.com