Evolutionary Dynamics of Infections on Networks with Superspreaders

Diseases spread through networks of contact between individuals, whether people, animals, or other organisms. Some individuals have many more contacts than others, so-called "superspreaders”. These individuals are known to speed up the spread of disease. This project investigates a surprising and previously overlooked consequence of superspreaders: they also make it much harder for natural selection to weed out harmful mutations in disease-causing organisms. Preliminary work by the research team shows that in networks with superspreaders, harmful mutations can persist and spread almost as if they carried no disadvantage at all. This matters because such mutations can later become dangerous, for example, if they help a pathogen resist a vaccine or medication, or if they are stepping stones toward more dangerous forms of a disease. The team will build new mathematical and computational tools to understand this effect in general terms, and then apply these tools to real social network data collected over twelve years from wild Asian elephants in Sri Lanka, a population threatened by a lethal virus. The findings will help inform conservation strategies for endangered elephants and provide broader insight relevant to problems such as antibiotic resistance, cancer development, and the makeup of microbial communities living in and on humans. The project also creates hands-on research training for undergraduate students, including a new course module in which students perform their own simulations and contribute original findings, broadening participation in quantitative biological research and helping prepare the next generation of scientists. Network structure fundamentally shapes infectious disease transmission, but its consequences for pathogen evolution, particularly against deleterious mutations, remain poorly understood. Preliminary simulations reveal that degree-heterogeneous networks containing superspreaders dramatically suppress selection against disadvantageous mutants, rendering their fixation probabilities nearly indistinguishable from neutral expectations. This proposal develops rigorous mathematical theory to explain and quantify this phenomenon, with three integrated aims. Task 1 addresses the core mathematical challenge that infection dynamics inherently involve three interacting classes (susceptible, wild-type-infected, mutant-infected), precluding standard two-state Moran process analysis. We construct an approximation that eliminates the susceptible class by weighting nodes according to steady-state infection probabilities, yielding a tractable node-weighted Moran process on a graph. This framework produces a suppression ratio, Z, quantifying how degree heterogeneity rescales effective selection strength, with analytical predictions for fixation probability and fixation time derived via weak-selection perturbation theory and validated across star, scale-free, modular, and fragmented network topologies. Task 2 extends this framework to describe behavior of selection-mutation balance, multi-strain genetic heterogeneity (using e.g. Shannon entropy), and fitness-valley crossing dynamics (to characterize sequential-fixation versus stochastic-tunneling regimes). Task 3 applies these frameworks to twelve years of empirically derived Asian elephant contact networks (Udawalawe National Park, Sri Lanka; 800+ identified individuals). We will compute the Z score and associated evolutionary observables longitudinally, revealing changes in evolutionary force. Further we will computationally perturb empirical networks to model conservation interventions, such as corridor restoration and altered group connectivity, testing predicted effects on pathogen evolutionary trajectories in a real-world, fragmented host population. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.