Interview with Dr. Jason George

Dr. Jason T. George is an Assistant Professor in Biomedical Engineering at Texas A&M University, with a joint appointment at the School of Engineering Medicine in Houston. Holding an M.D. from Baylor College of Medicine and a Ph.D. in Bioengineering from Rice University, his research uses mathematical models and machine learning to explore tumor behavior, immune interactions, and treatment resistance. Dr. George also shares insights for students interested in the future of mathematical oncology.

Isaac

Hello and welcome everyone to this special episode of High School Mathematical Oncology. I'm Isaac, and today we’re diving into a fascinating conversation about how math, biology, and medicine come together to take on one of the biggest challenges in science: cancer. It’s an absolute honor to introduce our guest today, Dr. Jason George.

Dr. George is an assistant professor in the Department of Biomedical Engineering at Texas A&M University. He leads a research group focused on using mathematical models to better understand cancer, the immune system, and how tumors resist treatment. His work blends advanced math, biology, and machine learning—helping shape the future of personalized cancer therapy.

Dr. George holds an MD from Baylor College of Medicine and a PhD in bioengineering from Rice University. He’s trained with some of the world’s leading cancer researchers, and today we’re lucky to hear about his journey, his scientific research, and where this exciting field is headed. Dr. George, thank you so much for joining us.

Dr. George

Thanks for that great introduction. I also want to compliment this initiative. Especially for students who are early in their careers, I think it’s really exciting to have a platform that shares the kind of work going on in this field and helps spark early interest. I’m happy to be here and excited to talk a little bit about what we do.

Isaac

Thank you! So starting off with your background as an MD/PhD, what inspired you to combine medicine with research? And how did your journey lead you to study cancer using mathematical models?

Dr. George

I’ll try to gear this toward high school students and undergraduates. My pathway wasn’t predetermined. At that stage, I had no idea what I wanted to do. I just followed the things I was passionate about. From an early point, I knew I wanted to do research—that was always an aspirational goal. I also wanted my work to improve people’s lives. And I loved math.

So I followed math through undergrad, got research experience, tried different topics, and eventually found one that really clicked: cancer modeling. That first experience was during my senior year of undergrad, at Baylor College of Medicine. It was transformative—it solidified my desire to study cancer for the rest of my career.

At that point, I also realized that to do something impactful and quantitative, I needed expertise in two areas. I needed to understand the math—developing and applying sophisticated tools—and I needed to understand the biology and clinical aspects of cancer, especially tumor resistance and treatment.

That’s where the MD and PhD training came in. The MD helped me understand the real-world clinical challenges, and the PhD gave me the technical training in math and engineering. It’s a long program—8 to 10 years—but over time, I kept falling in love with cancer research and mathematical modeling. Eventually, I realized that marrying these two skills could really drive impact in my research program.

Isaac

That’s really cool—how those two skills came together and led to your current work.

So in your research today, how do collaborations with researchers and clinicians ensure the accuracy of your models? And how can these models be used to optimize cancer treatment for individual patients?

Dr. George

That’s a great question. Collaboration is key. Many of our most exciting projects involve close collaboration with other researchers, clinicians, and lab scientists.

One benefit is that while you can focus on developing new theoretical tools, those tools need careful application, evaluation, and validation. That means applying them to real data—and for that, you need strong collaborations with clinicians and experimental scientists.

Selecting the right data and applying the tools correctly is a non-trivial task. It takes a full-time effort in its own right. Collaborations make this possible.

Another benefit is the generation of new questions. When people from different backgrounds come together, you see problems from different perspectives. The vocabulary is different, the problem-solving approach is different, and the creativity that emerges in that back-and-forth is incredibly valuable. That’s how some of the most impactful research ideas get born—through active discussion and cross-disciplinary thinking.

Isaac

Yeah, I’ve seen some of those team meetings—it’s always so cool to see ideas just bounce back and forth.

Dr. George

Exactly. And it’s also helpful to have that back-and-forth process where each person goes off to refine their part, then comes back to share progress. A clinician might not solve the math problem, and I wouldn’t help much during surgery—but together, we can identify the right problems and work toward a common goal.

Isaac

So what are your thoughts on pursuing an MD/PhD for a career in cancer research? Would you recommend that path? Or are there other good options?

Dr. George

Good question! And no—I definitely wouldn’t say the MD/PhD is the only path.

In fact, I’d say there are many paths to success. If you look at successful researchers, each one probably took a slightly different route. It’s less about the exact path and more about the common factors that successful people share.

First, you have to be passionate about the problem you’re working on. You need to really fall in love with it. Second, you need to have—or develop—the skills to contribute meaningfully to the field. For example, I love art, but I’m not a skilled painter, so I don’t pursue that professionally.

It’s also important to be curious early on. Explore different fields. If you’re good at many things, try closing some doors—it helps you focus. High school and undergrad are great times to explore different interests and figure out where your passions and skills align.

Now, about the MD/PhD specifically: I think it’s a very useful path, especially for people interested in translational research. It teaches you how to think, speak, and work like both a physician and a scientist. That’s helped me communicate with clinicians much more naturally, because I understand their day-to-day challenges and their way of thinking.

On the PhD side, it’s important to choose a field that aligns with your long-term interests. For me, combining the MD and PhD was the right choice—it ended up working out very well. But at the beginning, it was not obvious that it would be. And looking back, I definitely wouldn’t say it was the only way to do the kind of work I’m doing now.

Isaac

Yeah, I guess that’s great news for high schoolers who aren’t sure what path to follow—it’s more about curiosity and passion.

Dr. George

Exactly. And one more thing—if you’re unsure what to do, formal education is rarely a bad bet. Programs like MD-PhDs often come with a stipend, making long-term training more feasible.

Isaac

Switching gears a bit—what is a stochastic model, and why is randomness important when studying cancer’s interactions with the immune system?

Dr. George

Great question. Stochastic models account for uncertainty or incomplete information, unlike deterministic models, which assume everything follows exact rules.

Randomness is everywhere, especially in biology. For example, we can’t know exactly what the weather will be a year from now—that’s uncertainty. Same with cancer progression. Stochastic models help assign probabilities to future outcomes.

Take a coin flip: If we knew all the physics behind a flip, we might predict the outcome. But tracking everything would be too complex. So we model the outcome as a 50/50 chance. In cancer modeling, it’s similar—many small unknowns build up into large-scale unpredictability.

Isaac

Yeah, that makes sense. You can't know everything, so you have to model what you can with some wiggle room.

Dr. George

Exactly. Some things are unknowable in theory, and others are just practically hard to measure. In both cases, stochastic models help.

Isaac

How do your models track cancer–immune system interactions and predict resistance? Does randomness play a role there too?

Dr. George

Yes, absolutely. Say you're modeling tumor and immune cell populations—resistance or immune evasion would be certain events of interest. The math defines those events, and parameters control how strong those interactions are.

In another example, if you're predicting whether a T cell receptor will bind a tumor antigen, you're modeling a probability of recognition. That's another kind of randomness—whether that interaction happens or not.

And in time-dependent models, you can calculate the expected time until a tumor goes extinct or escapes. But these times are random variables—they vary with each simulation or patient. So we compute averages, variances, and confidence intervals.

Isaac

That’s really cool—understanding the likelihood of outcomes rather than trying to force a single answer.

Dr. George

Exactly. And there’s a strong parallel to physical chemistry—many molecules interacting at once. With tumors and immune cells, it’s similar: many moving parts, best described with probabilities.

Isaac

In your paper, you studied T cell receptors using machine learning. How does that help in designing better cancer vaccines or therapies?

Dr. George

Machine learning is a powerful tool—like anything else, it’s about using it where it makes sense.

In one project, we used it to identify key genes in the epithelial-to-mesenchymal transition. In another, we used it to predict which T cells recognize specific tumor antigens.

It helps optimize complex models, fit parameters better, and classify behaviors like recognition vs. non-recognition. But machine learning isn't magic—everything still needs to be validated rigorously with independent data.

Isaac

Yeah, that’s really cool. It’s like there’s a new tool, and I guess in mathematical oncology, there’ve been so many changes in the last 15–20 years. And this is just another tool that can help.

Dr. George

Yeah, and it’s an exciting opportunity for students who are now learning. I think this is probably no longer a secret, but students coming through the math-onco pathway now—especially those your age who haven’t fully committed yet—are going to be seeing models that didn’t exist for people like me when we were training. I see that as a really exciting opportunity for the next generation.

Isaac

So you’ve also used AlphaFold 3 and REST together. How do those work, and how does clustering with TCRdist help?

Dr. George

That’s a very specific example, so I’ll try to frame it in light of what I just said. AlphaFold 3 is a cutting-edge deep learning tool—a large language model architecture—that predicts how proteins fold into their three-dimensional structures. We’ve used AlphaFold and other similar tools to make structural predictions.

This is useful because our models track T cell receptor (TCR) interactions with tumor antigens—proteins presented on the surface of tumor cells. We’re also interested in viral infections that involve similar antigenic recognition.

Traditionally, structural data came from experimental techniques like X-ray crystallography, which is painstaking, time-consuming, and expensive. As a result, we have very few known structures in the immune context. Tools like AlphaFold 3 help expand the number of available templates, allowing us to model more interactions computationally.

As for TCRdist, that’s a separate tool developed by Paul Thomas’s group at St. Jude. It tackles the problem in a completely different way, clustering TCRs based on sequence similarity rather than structure. It provides orthogonal information—complementary to what we get from structural tools.

Integrating these tools—using AlphaFold’s structure predictions and TCRdist’s clustering—helped improve our ability to predict TCR recognition of tumor and viral antigens. That hybrid approach proved valuable for our research.

Isaac

Yeah, I think the first time I heard about AlphaFold 3 was from the YouTube channel Veritasium.

Dr. George

Yes.

Isaac

He really emphasized how important this could be for research and how many opportunities it opens up.

Dr. George

Absolutely. AlphaFold 3 is just the latest step—they’ve been iterating on this for some time. These models are only going to get better. This is a very active research area with major implications for protein design and identifying clinically relevant protein-protein interactions. So stay tuned—there’s a lot happening here.

Isaac

We’re running a bit short on time, so for students interested in mathematical oncology, how can they get started?

Dr. George

That’s a great question, especially for high school and undergrad students. I’d say a solid foundation in math is a great way to begin. “Mathematical oncology” involves both math and oncology, and people come into it from both ends.

Math is especially important because it builds on itself—basic concepts are prerequisites for advanced ones. So learning math is never time wasted. Even if you don’t end up in mathematical oncology, those skills are widely applicable.

In terms of formal coursework, look at fields at the interface—bioengineering, biological physics, or mathematical biology. These show how math is applied to biomedical problems, often involving cancer. If you’re more comfortable with math, those courses can help ease you into biology. If you come from a biology background, they’re a good entry point into quantitative thinking.

Beyond coursework, get research experience. As an undergrad—or even a high schooler—find a project that interests you and try to contribute or learn something from it. It doesn’t need to be your lifelong focus. Just understanding how research works and how to think like a scientist is incredibly valuable.

For more advanced students in math, I’d recommend proof-based courses in mathematical analysis. These build logical reasoning and problem-solving skills that carry over directly into research.

Ultimately, don’t worry about having it all figured out right away. Explore different areas, build your skills, and stay open to new interests. Even if you change directions, that experience will still serve you well.

Isaac

Thank you for that advice. And one last quick question—where do you see mathematical oncology in 10 years?

Dr. George

That’s a big question, but I’ll try to summarize.

One exciting direction is AI-driven innovation. Mathematical oncology models are going to help advance areas where traditional modeling has struggled, especially in complex systems where we now have large volumes of data.

That’s another point—the data. We’re gathering increasingly detailed and complex datasets. This opens up enormous opportunities for collaboration between experimentalists and modelers. You can now ask really rich, meaningful questions that weren’t accessible before.

A specific application would be digital twin technologies—personalized virtual models of patients that can help optimize treatment in the clinic based on individual characteristics.

And finally, something that’s close to my heart: the development of new mathematical frameworks. This is about using contemporary math to tackle modern questions in oncology. That requires deep expertise in both advanced math and biology. It’s challenging, but it’s where some of the most foundational work can happen.

Isaac

Dr. George, thank you so much for your time today. It’s been great hearing from you and learning where mathematical oncology is now, where it’s headed, and how students can get involved.

To all our viewers—whether you love math, biology, computer science, or just want to make a difference—there’s a place for you in this field.

Thank you so much for watching today. Stay curious, stay passionate, and we’ll see you next time.

Dr. George

The pleasure was mine. Thanks so much for having me. Take care.