Complex viral evolution as an unintended consequence of social distancing

Compared to influenza viruses, the evolution of the SARS-CoV-2 virus is highly complex. Despite several credible hypotheses revolving around unobserved epidemic reservoirs that obscured the actual evolutionary dynamics, the ultimate reasons remain puzzling. Using a novel evolutionary-epidemiological model combining the classical SIR differential equations with an agent-based approach, we investigate the impact of social distancing interventions on the structural complexity of emergent phylogenetic trees. In our model, interventions are connected to evolutionary complexity through their impact on non-specific immunity in the population. For our analysis, we introduce tree complexity metrics from multiple disciplines that have previously not been used in this context. The novel model can replicate baseline viral evolution models of influenza and SARS-CoV-2. Our results show that interventions can lead to increased viral evolutionary complexity and suggest that this mechanism's salience diminishes over time, as social distancing measures are phased out. An empirical investigation of viral evolution suggests that the evolution of influenza has been affected by the pandemic, supporting our hypothesis. We find that the observed evolutionary complexity is not necessarily the result of features specific to SARS-CoV-2 but may also arise in future pandemics of highly mutable viral pathogens due to social distancing.