Researchers from institutions across the United States are contributing to innovative projects at the NSF-Simons National Institute for Theory and Mathematics in Biology. These research projects focus on developing mathematical frameworks that illuminate emergent capabilities of biological systems. The NSF-Simons NITMB is developing the theory and mathematics needed to highlight the fundamental roles of physical, chemical, and biological constraints as organizing principles for understanding biological mechanisms. In order to highlight the diversity of experts developing NITMB Supported Research and the significance of their contributions, we will share insight into our growing community of researchers as part of the NITMB Spotlight series.
Andrea Liu is the Hepburn Professor of Physics at the University of Pennsylvania, where she is also the director of the Center for Soft and Living Matter. Liu’s research group uses a combination of analytical theory and computation to study soft and living matter. Andrea Liu is also a contributor to the NITMB-supported research project, ‘Exploring Global Epistasis through the Lens of Physical Learning.’
We spoke with Professor Liu to learn more about her work, involvement in the Biological Systems that Learn workshop, and how the NITMB is supporting Liu’s exploration of biological systems.
What is a big question you’ve been asking throughout your research?
“What I’ve been thinking about is how can we think about biological functions, properties that organisms need to survive. Just as a neural network is trained to have certain properties that we ask for, like the ability to classify cats and dogs, survival required organisms to develop biological functions. Can we think about that as having come from individually adjustable interactions as in neural networks? And if we think about it that way, how are they adjusted if the adjustments occur on time scales short compared to biological evolution? What are the consequences when you have a physical system like an organism or part of an organism, rather than a digital neural network that just lives on the computer? How are physical systems with individually adjustable interactions different from conventional condensed matter systems?”
What disciplines does your research integrate?
“Quite a few. Physics certainly, and biology certainly. But I’m also talking to mathematicians, computer scientists, electrical engineers and mechanical engineers. It’s a very wide range.”
Where do you find inspiration?
“I find it from trying to ask very simple questions about systems that I know about or come across.”
What aspects of your research could be interesting to mathematicians or applied to biology?
“The big idea of the NITMB workshop, that it is useful to bring people together across fields to study how individually-adjustable interactions can give rise to complex collective behaviors, is interesting to both math and biology. Also, the seed project that we were funded by the NITMB to do, which funds my terrific collaborator Dr. Farshid Mohammad-Rafiee to work on a joint project with my Penn colleague Josh Plotkin and me, was to think about global epistasis in proteins. Suppose you have some protein, and you want to introduce changes, I mean mutations, into the amino acid sequence. I can mutate any one amino acid and look at the effect, but what happens if I mutate two of them together? That’s not the same as mutating one and then mutating the other and adding the effects together. Global epistasis says they’re not additive in a certain underlying way, and Josh has developed a machine learning approach to actually pull out what that nonlinearity is. We’re thinking that by changing the amino acid sequence, you’re changing effective interactions within the protein. Can we think about this property of global epistasis coming about because the protein is a physical object with interactions that have been adjusted by evolution in order to have that function?”
What about the NITMB do you find exciting?
“First of all, the funding to study this question of global epistasis is very exciting to have. It’s useful for biotechnology to mutate a protein, so it still has the same function. Global epistasis is identified currently by machine learning, but it’d be nice to have physical insight into where it comes from--some quantitative understanding of its origin and why it has the form it has. We have over 100 years of studying systems in which the interactions are fixed, but the mathematical tools and the theoretical tools to study systems in interactions that are individually adjustable still need to be developed. It has consequences for biological function. This could potentially help us understand how non-biological physical systems that behave as neural networks work, as well as a whole host of other non-biological systems being developed. At the Biological Systems that Learn workshop, we got a whole bunch of people together from mathematics, biology, physics, material science, engineering, electrical engineering, and mechanical engineering, and we started forming the connections, common language, and common outlook for tackling these systems.”
What career achievement are you most proud of?
“I’m quite proud of the idea of unifying systems with individually-adjustable effective interactions as “tunable matter.” I call this idea ‘many more is more different.’ One of the properties that comes from having individually adjustable interactions is that the larger the system gets, the more things it can potentially do. That’s why neural networks are so useful now—they got big enough to learn tasks that are complex enough to be useful to us. Usually in condensed matter physics, properties don’t change much as systems get bigger, so this is a really different kind of system. I’m also proud of work on jamming, which is about understanding how systems that are disordered can become rigid. That work was done with Sid Nagel at the University of Chicago--from jamming we plunged into tunable matter together so are still collaborating after more than a quarter century--I’m proud of that!”
Outside of your research, what other interests do you have?
“I have two daughters, one of my daughters is a graduate student at the University of Chicago with NITMB member Stephanie Palmer. My other daughter is in her last year of college studying biology. I am vice-chair of the Committee for Human Rights of the National Academies of Science, Engineering and Medicine. I am very invested in upholding human and civil rights, especially now, and am proud to be part of the committee. I also like to play chamber music when I can.”
What are you hoping to work on in the future?
“I really hope to be working with people across fields to establish how to handle the collective behavior of tunable matter. That’s what I can see for the foreseeable future. I want to work with people.”
More information about Andrea Liu’s work is available in her Oppenheimer Lecture given at the University of California, Berkeley. Liu’s publications are also a valuable resource for learning more about this work.