Biophysical Fitness Landscape Design: Reshaping Protein Evolution

In evolution, populations climb toward the top of fitness landscapes in the way that particles are drawn toward the bottom of potential energy landscapes in physics. While methods exist to optimize protein structure and function along the local fitness landscape, we lack protocols for globally reshaping the fitness peaks and valleys that shape protein evolution. Here, we establish foundations of fitness landscape design (FLD), introducing computational algorithms which allow antibodies to be tuned to quantitatively control and trap protein evolution akin to how holographic optical tweezers can be tuned to trap ions. First, we derive an analytically tractable, in silico-, in vitro-, and epidemiologically validated biophysical model of viral protein fitness from microscopic chemical reactions. We compute theoretical bounds on FLD and validate them using a recently published experimental dataset of binding affinities between thousands of mutated antibody-antigen pairs. We then apply FLD to suppress the fitnesses of SARS-CoV-2 genotype networks and subsequently discover proactive vaccines which restrict the fitness trajectories of viral escape mutations before they emerge. FLD opens the door to improved biosecurity and pandemic preparedness via proactive antibody and vaccine design.