Englander Institute for Precision Medicine

A Physicist’s Perspective on the Future of Precision Medicine

A conversation with the EIPM’s Karsten Suhre, Ph.D. about scientific curiosity, computational biology, artificial intelligence and Cornell’s global research community

A Physicist’s Perspective on the Future of Precision Medicine

A conversation with the EIPM’s Karsten Suhre, Ph.D. about scientific curiosity, computational biology, artificial intelligence and Cornell’s global research community

PROFILE INTERVIEW

Dr. SuhreKarsten Suhre has never been content to stay in one scientific lane. Trained as a theoretical physicist in Germany, he earned his doctorate in France, worked in atmospheric modeling and the automotive industry, and then moved into bioinformatics and systems biology. For the past 15 years, his work has connected computational approaches with questions in human health—from metabolomics and cardiovascular risk to diabetes, infectious disease and precision medicine. Based at Weill Cornell Medicine-Qatar, he is also part of a far-reaching research network linking Cornell teams in New York, Ithaca and the Middle East. In this conversation, he reflects on the experiences that shaped his career, the promise and uncertainties of artificial intelligence, and why he believes scientific curiosity is most rewarding when it can make a difference in the clinic.

Q: Where did you grow up, and when did you first become interested in science?

A: I grew up in Osnabrück, in northern Germany, not far from the Dutch border. I became interested in science in high school. In the German education system, students begin to specialize in 11th grade, and I focused on physics and mathematics. I have always thought of myself as a physicist rather than a physician or a traditional biologist, which I am not.

Q: You have described yourself as “genetically German and scientifically French.” What do you mean by that?

A: I studied physics in Germany through what was then the diploma level, roughly equivalent to a master’s degree. I then moved to Toulouse for my Ph.D. and spent a total of 13 years in France, working with the French National Centre for Scientific Research, or CNRS. It is the major national general research organization, with laboratories working across fields from physics to the social sciences.

That experience gave me a strong connection to the French scientific community and language. I find it easier to talk about science in French than in German. English has become much more natural after 15 years with Cornell, but French was my scientific language for a long time.

Q: Your early research focused on atmospheric chemistry and meteorological modeling. How did that work prepare you for your later research?

A: My doctoral work was rooted in physics and computational modeling. I joined a laboratory in Toulouse that was collaborating with Météo-France to develop a high-resolution model of weather systems—not the global models, but models that could describe cloud systems, storms and other phenomena at a more local scale, on then cutting-edge supercomputers, like the Cray 2 system

My role was to add chemistry to the model. We incorporated substances such as ozone and nitrogen oxides, along with chemical reactions among them, and developed numerical methods to solve the resulting differential equations. It was high-performance computing, with a large number of variables interacting across a three-dimensional grid.

That turned out to be quite relevant to the work I do now. Modeling chemical reactions in the atmosphere has similarities to modeling metabolic pathways and biological systems. The subject matter changed, but many of the computational ideas carried over.

Q: Tell me about working with EIPM Director Olivier Elemento, Ph.D.?

A: What attracted me to working with Olivier Elemento was that we speak the same scientific language.  Olivier brings a strong background in machine learning and computational science, and that creates a natural connection with my own research. Although our scientific backgrounds are different, we share a common way of approaching problems, a curiosity about what the data can reveal, and an enthusiasm for exploring new possibilities. That makes working together both intellectually stimulating and highly productive. 

Q: Your career then took an unexpected turn into the automotive industry. What happened?

A: After my Ph.D., I returned to my hometown for personal reasons and spent two years working as a project engineer for a Karmann, then a major provider of systems to the automotive industry. I was involved in the project management of developing a convertible version of the Mercedes CLK, code-named A209. It was an interesting experience, and I learned a great deal about project management, schedules, controlling and the realities of getting a complex product finished.

But it also clarified something important: I am a scientist, not an engineer. I enjoyed the technical challenges, but I missed research. I was fortunate that my position with CNRS allowed me to return to scientific work.

Q: How did you make the transition from atmospheric science to biology and bioinformatics?

A: It was partly a matter of timing and being open to an unexpected opportunity. I initially planned to return to atmospheric research. Around 2001, however, I was reading about the Human Genome Project and came across an interview with a French researcher, Jean-Michel Claverie. I contacted his group and learned that the human genome project urgently needed bioinformaticians, someone with a background in mathematics and computational modeling.

I put together a proposal to CNRS, it was accepted, and I moved into bioinformatics. The laboratory brought together complementary expertise in computational biology and structural biology. I began working on questions involving protein structure and genomes of microorganisms, including research on large viruses. It was a fascinating transition because I could apply the quantitative training I already had to an entirely new area of science.

Q: What drew you to Cornell, and what has kept you connected to the institution for 15 years?

A: After five years in Marseille, and then another five as a professor of bioinformatics at the University of Munich, I had reached a point in my career when I was again ready for a new challenge. I saw an advertisement for a Cornell position. To be completely honest, I did not know much about Cornell at the time, but I learned that the opportunity was part of an initiative to expand the university’s biomedical research enterprise to the Middle East, and I was curious about the region and culture, and interest shared with my wife, what made the move attractive.

What impressed me was the combination of scientific ambition and the opportunity to build something. Cornell offered a serious research environment, strong collaborators and access to clinical expertise. I also developed close relationships with colleagues in New York, including researchers who were working in Qatar and across the Cornell network. Those relationships have been an important part of my work ever since.

Q: You are based in Qatar. How has that location shaped your work and your view of Cornell’s international mission?

A: One of the things I value most about working in Qatar is the opportunity for intercultural collaboration. Cornell brings together people from many backgrounds, and the Qatar campus adds another dimension to that exchange. I have seen colleagues from different parts of the world work together, build trust and develop research relationships that would be difficult to create otherwise.

I believe education and research can help create common ground. When people have opportunities to collaborate and benefit from one another’s work, that can be a powerful counterweight to polarization. Cornell’s presence in the region is important not only because of the research opportunities, but also because it supports exchange among people and institutions.

There are real challenges to sustaining collaborations across countries and organizations. Relationships take time to build, and staff turnover or differences in institutional structures can make it difficult to move projects forward. But the opportunities are significant, and I hope Cornell continues to invest in them.

Q: Artificial intelligence has become a major focus across science. How long have you been interested in the field?

A: I encountered ideas related to artificial intelligence and machine learning very early in my studies. In the 1980s, we were learning about concepts such as associative memory. By the late 1990s, I was also seeing the use of large computing systems, such as the Thinking Machines CM5 computer and its application to neural networks. The field has gone through periods of intense interest and quieter periods, but the current moment is different in scale.

I have not spent my career as an AI researcher, but I have always been interested in the possibilities. What is striking now is how accessible these tools have become and how quickly they can help researchers overcome practical obstacles.

Q: What has changed most in the current wave of AI?

A: The biggest change for me is that AI can help me do things I could not realistically do before. I work with complex data, including mass spectrometry data, and for years one of the bottlenecks was getting software to work—dealing with versions, dependencies and technical details before you could even begin the analysis.

Now I can describe the problem to an AI tool, provide the software and ask it to make the pieces work together. It can try different approaches and help solve problems that once consumed a great deal of time. Even something as simple as producing a clear scientific graphic used to involve searching through documentation and experimenting with code. Now I can explain what I want and refine the result through a conversation.

The tools can also help reveal patterns or possibilities that I might not have noticed. In that sense, they provide access to a kind of collective knowledge. They are not replacing the time and thought required for a research project; they are expanding what one person, or a small team can accomplish.

Q: Where do you see AI taking science—and what concerns you?

A: I think we are still in the early stages, and the possibilities are extraordinary. AI may help us analyze data in new ways, but it may also teach us something about how the brain works. Researchers are exploring systems that connect living neurons with electronic devices, for example. That raises fascinating questions about the relationship between biological and artificial information processing.

At the same time, we need to understand how these systems reach their conclusions. If an AI system appears to solve a difficult problem, we need to know whether it has genuinely identified something new or is reproducing information it has encountered before. I would like to see more emphasis on explainability and on understanding why a system works, not simply on building a larger or more powerful one.

There is also a human dimension. Voice-based AI, for example, can feel surprisingly personal. When a system speaks with a particular voice or accent, people may respond emotionally in ways they do not expect. We should pay attention to those effects as the technology becomes more integrated into daily life.

My greatest concern is less about the technology itself than about the economic and competitive pressures driving it. Companies and countries may feel they cannot slow down, even when there are important questions to resolve. Scientists could help steer the field toward better understanding and responsible use, rather than treating speed and scale as the only measures of progress.

Q: One of your recent projects uses NMR metabolomics to assess cardiovascular, kidney and metabolic risk. What is the study trying to show?

A: The work builds on a long history of collaboration around the Nightingale Health NMR metabolomics platform, and a longstanding relationship with the lead PI, EIPM Member Dr. Jan Krumsiek. I had worked with this technology in Qatar and with the UK Biobank, and the current study gives us an opportunity to examine how these measurements might be used in a real-world healthcare setting.

Large population biobanks are extremely valuable, but they often involve relatively healthy participants and carefully structured data collected at planned intervals. Our collaboration with Kaiser Permanente allows us to look at people who are already receiving care for cardiometabolic conditions. We can ask a more practical question: What would it look like if a physician had this kind of metabolomics information alongside a patient's actual medical history?

Clinical records are not always neat. A patient may have several measurements close together, then a long gap, then a new diagnosis or a change in a clinical marker. We wanted to understand how metabolomics data could be interpreted in that more complicated, realistic context. The study suggests that these measurements can add useful information to risk assessment, while also highlighting the importance of understanding the patient over time—not just looking at one number or one test result.

Q: What would it take to bring this kind of testing into routine clinical care?

A: One long-term goal is to help move multi-omics approaches from research into clinical practice. Nightingale has made significant progress in Finland, where this type of testing is already much more integrated into the healthcare system. We wanted to create more opportunities to work with the company in New York, and one condition of our collaboration was that it establish a laboratory in the region.

The next question is whether Cornell and its clinical partners can help bring this approach into patient care. It is not just about generating more data. We need to understand what the information means to physicians and patients, how it fits into the medical record, and whether it can support better decisions.

There is also a behavioral dimension. Sometimes a test can make people more aware of their health and motivate them to change their habits. A person may already know that smoking or eating poorly is unhealthy, but seeing information about their own health can make the message more immediate. That effect is difficult to measure, but it is part of the broader potential of personalized medicine.

Q: You have worked across a wide range of diseases and research areas. What connects those interests?

A: For me, the connection is the underlying physics and the opportunity to understand complex systems. As a computational scientist, I often think of a disease as a particular readout or perturbation in a biological system. The methods and questions can be related even when the diseases are very different.

Curiosity is a big part of it. I enjoy learning a new field and working with people who understand the biology or clinical questions in depth. But I also want the work to matter. I am less interested in developing an algorithm simply because it is mathematically elegant than in asking where it might lead and whether it could eventually help address a real problem.

That is one reason I value collaborations with clinicians. They bring the work closer to the questions that matter to patients. Combining different technologies and areas of expertise can open opportunities that no single discipline would reach on its own.

Q: What are the biggest challenges in moving promising research forward?

A: A major challenge is the amount of administrative work involved in research. Ethics reviews, compliance requirements, data-use agreements and legal processes are important, but they can also make it difficult to explore a promising idea quickly. Sometimes a simple preliminary analysis can require weeks of paperwork.

I understand why safeguards are necessary. But we need to find a balance that allows responsible research to move forward. Otherwise, scientists can spend so much time navigating processes that they lose interest to do the science.

Q: What do you enjoy doing outside the lab?

A: I am not sure scientists ever completely turn off their brains. My life is also divided between countries, so the rhythm changes depending on where my wife and I are. When I am in Qatar on my own, I tend to work a lot. When she is here, I try to make more room for other things.

I still enjoy video games. More recently, I have become very interested in 3D printing. We started by printing molecular structures, but I have also explored more personal projects. There is something satisfying about starting a print in the evening and coming back in the morning to find that an object has appeared almost out of nothing.

Q: What continues to excite you about the work ahead?

A: The possibility of connecting ideas and people across disciplines and locations. There are projects that may have been ahead of their time when we first discussed them, and it can be worthwhile to revisit them with new tools and new collaborators. The combination of computational science, emerging technologies and clinical expertise offers many directions to explore.

After 15 years with Cornell, I still find that environment stimulating. I want to keep learning, keep asking new questions and, most importantly, keep working on research that has the potential to make a difference.

Q: What has it been like to work in the Middle East during a time of conflict? 

A: The past year has reinforced for me how important it is to maintain scientific collaboration and human connections between cultures, even under the most difficult circumstances. Science and specifically scientific education can bring people together across borders and backgrounds, and that is what Weill Cornell Medicine does in Qatar. Despite the challenges and uncertainty, our commitment to research, knowledge-sharing, and improving human health remains constant. In many ways, that commitment is more important than ever. 

# # #

Weill Cornell Medicine Englander Institute for Precision Medicine 413 E 69th Street
Belfer Research Building
New York, NY 10021