Every natural system is rife with interactions. And when only two things interact, the outcome is usually easy to predict. A cue stick strikes a ball, and the ball rolls away. Penicillin encounters the strep throat-causing Streptococcus germ, and the germ dies.
But what happens when you add a third — or fourth, or fifth, or more — component to the mix? The effects of such higher-order interactions can be difficult to forecast.
“You can do a lot of experiments with pairwise interactions, where you take two species at a time, and look at how a population changes over time,” says biological physicist Jacopo Grilli, an Omidyar Postdoctoral Fellow at SFI. “But often when you put more species together, the results are very different from what you would expect from looking at pairwise interactions.
Different regimens of antibiotics, for example, may work together to beat an infection — or promote antibiotic resistance in a pathogen. The antibiotics interact with each other, as well as with the mix of species in a microbiome. SFI External Professor Pamela Yeh has seen these emergent effects firsthand in her research.
“We were surprised to see not only that there were higher-order interactions in antibiotic combinations but that they were so prevalent, and so strong,” she says. “There is a rich literature in fields like ecology, microbiology, and pharmacology looking at pairwise interactions, but higher-order interactions have been mostly ignored.”
Grilli and Yeh, together with SFI External Professor Van Savage, have organized “Higher-order interactions: experiments, inference and models,” a working group to be held at SFI March 3-6. The meeting will merge the expertise of researchers from a range of disciplines — including evolutionary biology, network theory, ecology, social sciences, and physics — to identify general and fundamental questions that drive these effects.
The organizers want to know how and in what ways the whole is different than the sum of the parts, as well as the degree to which the ideas that are relevant to one field are transferrable to others.
Participants will explore experimental designs that can detect the effects of higher-order interactions in a range of disciplines. They’ll also discuss methods for inferring the presence of these interactions in outcomes, and, ultimately, develop theoretical approaches to forecasting.
Savage says he hopes that the diversity of research areas represented at the working group will help them reach a consensus on the best theories and data sets that illuminate these effects. “With pushback from different fields,” he says, “we can work together collectively and push forward new ideas.”