Olivier Elemento, PhD

“I grew up in a village in the South of France, surrounded by vineyards in a very rural area. There were no scientists, no doctors in my family. When I was a kid, I had severe allergies and asthma. My parents took me from doctor to doctor: that essentially exposed me to medicine and led me to want to participate in scientific discovery. My parents wanted to cultivate that interest and got me a microscope so that I could look at algae, microbes in the water, things like that. I absolutely loved it. Around the same time my grandfather bought me a computer and I got hooked. [...]

“Here at Cornell, I’m working on applying this idea of computational biology to very specific medical problems. Trying to solve, for example, how to go from sequencing somebody’s genome to treating that person. [...] The first time that we actually sequenced somebody’s cancer genome I was pretty puzzled to see so many mutations.

"That was around 2010, and it led to our whole program of personalized medicine: trying to understand what makes every disease unique and how we can use that information to treat patients. Let’s say two people get prostate cancer. If we look at the DNA of these two individuals, we’ll see lots and lots of differences. In fact, we’ll see mostly differences. So it doesn’t make any sense to treat those two individuals in the same way, right?

“We’ve gone beyond just genomes now. With cancers, for example, we look at normal cells, immune cells, cancer cells—they’re all intertwined in the tumor. The tumor cells require the immune cells and the normal cells—it’s like an ecosystem. We want to be able to build a virtual model of each patient’s cancer so that we can simulate what would happen if we were to treat a given patient with two or three different drugs at the same time. That’s very much driven by AI. [...]

“Ideally, we want to use models to design specialized drugs that you can make yourself with a machine, essentially 3D-printing the drug. In the process of designing those drugs the machine will learn from the body, making sure that the drug isn’t toxic for normal cells. We’re not there yet, but that’s the holy grail. We know that every patient has a different disease, due to the diversity of their genetic background and to the heterogeneity of the disease itself. We should be able to produce a drug or treatment that’s completely specific to each patient. I don’t know if we’ll be able to do this in my lifetime, but all these technologies kind of exist. It’s a matter of engineering, of putting it all together.”

As told to Parker Menzimer.