Quantifying Ruggedness and Navigability in SARS-CoV-2 Antibody Binding Landscapes

Fitness landscapes provide a framework for studying how viral mutations affect protein-protein binding and potential escape routes, which impact vaccine design strategies. We used empirical binding affinity data and a biophysical fitness model to computationally quantify the ruggedness and navigability of viral fitness landscapes and protein-protein binding affinity landscapes. First, we established null models by simulating random fitness landscapes and computing the number of local maxima. Analysis of the SARS-CoV-2 spike protein's binding landscapes with the ACE receptor and four unique antibodies suggests that the spike-ACE2 binding affinity landscape is smooth compared to spike-antibody landscapes, which show higher ruggedness. We also extend beyond strict peaks to count near-maxima and minima, which allow a small number of escape paths in the landscape. Ongoing work examines concentration-dependent ruggedness. These findings highlight how protein-protein interactions shape evolutionary trajectories and potentially guide future vaccine design.