Attacking Cancer with Team Science

Reflections is a series featuring Columbia cancer experts, looking back at how far we’ve come in advancing and impacting change in research, treatment, care, and survivorship, and their perspectives on what lies ahead.

December 22, 2022

To cure cancer, one way of thinking is simply not enough. 

That philosophy is driving investigators across medicine and science to approach solving cancer collaboratively. As cancer research continues to evolve, engagement across diverse domains of expertise has become essential for progress. The hardest problems can best be solved through the convergence of mathematics, engineering, biology, chemistry, medicine – and all of their respective subdisciplines.

According to cancer biologist Kenneth Olive, PhD, nothing can be more effective than nucleating large groups of investigators with diverse expertise to tackle, as a team, the hardest problems in cancer. The old archetype of a lone scientist in a lab making a discovery and proclaiming, “Eureka!” to an audience of none, is long gone. 

For example, the recent explosion of large-scale datasets has led to the emergence of innovative computational methods at the intersection of mathematics, computer science, and biology. A scientist can now generate more data in a single experiment than an entire institute would have produced just 10 years ago. As a consequence, it is no longer unconventional for cancer biologists to work closely with computational biologists, systems biologists, physicists, and engineers. But how do these different types of experts work together when so many differences exist, right down to the ways they think and speak? Dr. Olive shares insights from his own career learning to bridge these barriers and build cohesive transdisciplinary teams and how “team science” has become increasingly critical in cancer research and medicine.

What motivated the shift towards team science?

I wouldn’t say there’s a singular moment when collaborations started becoming the norm; it has been an evolution over time. 

The reality is that there are now so many different specialized techniques and approaches to research that are useful, no one researcher or lab can possibly deploy them all. It may be possible to answer a question alone using traditional approaches, but are you answering it as efficiently and as completely as you could? What deeper insight would emerge if you came at the problem from multiple angles? My belief is that if you're not engaging with experts to employ the optimal technique for a question, then you’re probably wasting time and resources. It is also becoming apparent that the hardest of biomedical challenges may only be addressed through the combined efforts of diverse teams. The number of published papers with dozens of authors is increasing over time, as is the availability of funding mechanisms for transdisciplinary research.

What has been challenging about building big research collaborations in cancer?

Different disciplines are siloed from the outset. We are trained in different places, with different techniques, tools, and resources – all of which makes sense in context because each discipline has different professional goals. For example, in medical school, budding physicians are trained to carry out techniques in a proscribed manner with the goal of a 100% success rate; if they deviate from the standard, a patient’s life could be at risk. In graduate school, nascent lab scientists are trained to try new things, often failing repeatedly; if you aren’t failing, you are not pushing the boundaries of knowledge. Similar distinctions define the domains of engineering, math, and computer science. These differences propagate from training, to practice, to mindset. You can even see this reflected in language – each domain and discipline has its own entirely distinct dialect. The words that physicians use are different than those used by scientists, even when we're trying to describe the same thing. This serves as a barrier against productive interactions across disciplines.

Beyond this, there are some very practical challenges. How does a 30-person team allocate appropriate professional credit to all participants? How does one navigate publishing standards and expectations that vary across fields? What constitutes a “publishable unit” in one field may be insufficient in the next. How does one motivate so many people to spend time on one problem?

How do you overcome those challenges?

In my experience, two pieces are imperative. First, you must share your passion for the problem to get everyone excited, focused, and in lock-step. Hard problems like cancer can serve as a center of gravity for the group, but there is no shortcut to motivating your colleagues; it requires spending time sharing your vision through direct engagement. Second, you must communicate your genuine interest in what others are doing. When colleagues recognize that you are interested in more than just what they can do for your personal project, they respond to it. 

Building a team takes building faith. You have to get to know the other players as individuals, as humans, learn to speak their dialect, and get them to want to work together as a team to solve hard problems. Otherwise, you're just stuck with what you can do. 

The fact is that scientists are people – some of them are more social, some less. Figuring out how to engage with all of the different personality types across science is one of the key skillsets of being a translational scientist. For me, it’s fun. I may have been trained as a cancer geneticist, but I find pretty much all of science fascinating. I really like understanding how other people do their stuff.  

So when I go to visit a pathologist, the first thing I do when I walk into their office—which is usually crammed with slide boxes, microscopes, and case files—I don’t start with “I'm Ken, and this is what I need.” Rather, I ask, “What are you looking at?” Next thing you know, there are two of us looking down a microscope, and I’m learning about a disease I’ve never heard of, building my pathology vocabulary, and connecting on a more genuine level.

“Multidisciplinary” and “transdisciplinary” are often used interchangeably. What’s the difference?

Both involve getting groups of people to work together. However, the idea of transdisciplinary research is to merge not only the techniques, but also the mindsets of each discipline so as to build a holistic means of thinking about hard problems. A factory is a multidisciplinary: each station performs one specialized task to a widget and then hands it off to the next station for the next step. This can be productive in terms of throughput. But creativity in the face of complexity requires using all the tools at hand to come up with a new way forward. 

The idea of transdisciplinary research is getting different mindsets to inform one another about how to think, not just how to do.

Why is transdisciplinary the wave of the future?

It’s the most effective way to answer big, hard problems. The problems persist because individuals have been unsuccessful in solving them alone. To do better, we can’t just rely on one smart person to do the heavy lifting but rather work on these challenges as teams. I don’t know how to code in statistical programming languages; I don't treat patients; I can’t build new scientific instruments, and yet my extended teams rely on these techniques daily via collaborators. The transdisciplinary research approach works. Frankly, it's the only way I've ever been successful in answering big hard problems, and I believe this concept is becoming generally recognized in biomedical science.

Large collaborative research efforts are becoming the norm in cancer. Any reflections on what’s in store for the future?

What I hope for the future is that we are integrating both science and medicine into one effort. I think research questions should be built into every aspect of a clinical trial – ahead, during, and even after it. I also love the idea of a post-clinical trial: a research study that's performed after the trial has been completed to better understand the results. There are a few examples of post-clinical trials, but it’s not typical. If we're doing our jobs right, we should be learning new science from every single patient and bringing the results of science to every new patient.