Engineering Collisions: How NYU Is Remaking Health Research

This sponsored article is brought to you by the NYU Tandon School of Engineering.

The traditional approach to academic research goes something like this: gather experts in a discipline, put them in a building, and hope that something useful emerges. Biology departments do biology. Engineering departments do engineering. Medical schools treat patients.

NYU is disrupting this model. To his new Institute for Health Engineeringthe organizing principle centers on disease states rather than traditional disciplines. Instead of asking “what can electrical engineers bring to medicine?” “, they ask themselves “what would it take to cure allergic asthma?” ”, then bring together everyone who can answer this question, whether they are immunologists, computational biologists, materials scientists, AI researchers or wireless communications engineers.

Jeffrey Hubbell, NYU vice president for bioengineering strategy and professor of chemical and biomolecular engineering at NYU’s Tandon School of Engineering.New York University

The first results suggest that they are on something. A chemical engineer and an electrical engineer collaborated to build a device that can detect airborne threats, including pathogens, which is now a startup. A visually impaired doctor teamed up with mechanical engineers to create navigation technology for blind subway riders. And Jeffrey Hubbell, director of the Institute, proposes “reverse vaccines” that could reprogram the immune system to treat conditions ranging from celiac disease to allergies — work that requires equal mastery of immunology, molecular engineering and materials science.

The underlying problem these collaborations address is as much conceptual as it is organizational. In his field, Hubbell says modern medicine has optimized around a single strategy: developing drugs that block specific molecules or suppress targeted immune responses. Antibody technology has been the workhorse of this approach. “It’s really suited to blocking one thing at a time,” he says. The pharmaceutical industry has become extraordinarily good at creating these inhibitors, each designed to block a particular pathway.

But Hubbell asks a different question: Rather than inhibiting one bad thing at a time, what if you could promote one good thing and generate a cascade that contravenes multiple bad pathways simultaneously? In the event of inflammation, could we bias the system towards immunological tolerance instead of blocking inflammatory molecules one by one? In cancer, could you drive pro-inflammatory pathways in the tumor microenvironment that would simultaneously overcome multiple immunosuppressive features?

This shift from inhibition to activation requires a fundamentally different toolbox – and a different type of researcher. “We use biological molecules such as proteins or material structures (soluble polymers, supramolecular structures of nanomaterials) to drive these more fundamental characteristics,” says Hubbell. You can’t develop these approaches if you only understand biology, or if you only understand materials science, or if you only understand immunology. You need an understanding and mastery of all three.

“There will be people doing AI, data science, computer theory, people doing immunoengineering and other biological engineering work, people doing materials science and quantum engineering, all very close to each other.” —Jeffrey Hubbell, NYU Tandon

Which logically leads us to the next question: how do we create researchers with such interdisciplinary depth?

The answer is not what one might expect. “There may have been a time when the goal was to make the bioengineer understand the language of biology,” says Hubbell. “But those days are long gone. Now the engineer has to become a biologist, or become an immunologist, or become a neuroscientist.”

Hubbell isn’t talking about engineers learning enough biology to collaborate with biologists. It describes something more radical: training people whose disciplinary identity is truly ambiguous. “It’s very difficult for neuroengineering students to know whether they are engineers or neuroscientists,” says Hubbell. “That’s the whole idea.”

His own students are an example. They publish in immunology journals and appear at immunology conferences. “No one knows they are engineers,” he says. But they bring technical approaches – computer modeling, materials design, systems thinking – to immunological problems in ways that traditional immunologists would not.

The mechanism for creating these hybrid researchers is what Hubbell calls a “milieu.” “Learning everything on your own is hopeless,” he admits, “but learning it in an environment becomes very, very effective.”

NYU is expanding its facilities to include a science and technology center designed to force encounters between people from various schools and disciplines who would not naturally cross paths.Tracey Friedman/NYU

NYU makes this environment physical. The university acquired a tall building in Manhattan which will serve as a science and technology center – a deliberate co-location strategy designed to force encounters between people from diverse schools and disciplines who would not naturally cross paths.

Juan de Pablo is Anne and Joel Ehrenkranz’s executive vice president for global science and technology and executive dean of the NYU Tandon School of Engineering.Steve Myaskovsky, courtesy of NYU Photo Office

“There will be people doing AI, data science, computer theory, people doing immunoengineering and other biological engineering work, people doing materials science and quantum engineering, all very close to each other,” says Hubbell.

The strategy reflects what Juan de Pablo, Anne and Joel Ehrenkranz executive vice president for global science and technology and executive dean of the NYU Tandon School of Engineering, describes as organizing around “grand challenges” rather than traditional disciplines. “What drives recruiting, the spaces and people we recruit, are the problems we’re trying to solve,” he says. “Great minds want to leave a legacy, and we make that possible here. »

But physical proximity is not enough. The Institute also cultivates what Hubbell calls an “explicit” rather than “tacit” approach to translation – thinking about clinical and commercial pathways from day one.

“It’s a terrible thing to solve a problem that no one cares about,” Hubbell told his students. To avoid this, the Institute organizes “translation exercises”: group sessions in which researchers trace the entire path from discovery to deployment before launching multi-year research programs. Where could this fail? What experiments would quickly prove this idea wrong? If it is a drug, how long would the clinical trial last? If it is an IT method, how can it be deployed securely?

The new cross-institutional initiative represents a major investment in science and technology and includes the addition of new faculty, state-of-the-art facilities and innovative programs.NYU Tandon

This approach stands in stark contrast to typical academic practice. “Sometimes academics tend to think about something for 20 minutes and start a 5-year PhD program,” says Hubbell. “That’s probably not a good way to go.” Instead, the Institute brings together people who have actually developed drugs, built algorithms, or commercialized devices – importing their hard-won experience into the planning phase before a single experiment is launched.

The timing may be fortuitous. De Pablo notes that AI significantly reduces lead times. “What we thought would take 10 years, we might be able to do in 5 years,” he says.

But he quickly sees the limits of AI. While tools like AlphaFold can predict how a single protein folds – a breakthrough of the last five years – biology operates on much larger scales. “What we really need to do now is design not a single protein, but collections of proteins that work together to solve a specific problem,” says de Pablo.

Hubbell agrees: “Biology is much larger: it has many, many systems. » The liver and kidneys are in different locations but interact. The gut and brain are neurologically connected in ways that researchers are just beginning to map. “AI isn’t here yet, but it will be one day. And that’s our job: to develop the data sets, the computing frameworks, the systems frameworks to take the next steps.”

It’s a moment of unusual ambition. “At a time when we are seeing certain research institutes withdraw a little and limit their ambitions,” says de Pablo, “we are doing exactly the opposite. the big challenges which we want and must tackle.

The bet is that the advances worth making cannot emerge from the work of a single discipline. They require collisions – sometimes planned, sometimes accidental – between people speaking different technical languages ​​and wanting to develop a common one. NYU designs these collisions on a large scale.