Scaling of SARS-CoV-2 Genomic Divergence with Cumulative COVID-19 Infections

The rate at which viral lineages accumulate genetic divergence is central to understanding pathogen evolution, yet direct empirical measurements are limited. The COVID-19 pandemic, through the combination of pandemic-scale replication and dense genomic surveillance, offers an opportunity to calibrate this relationship directly. Unlike previous analyses that use time as the independent variable, this study uses cumulative infections, providing an epidemiological clock calibrated in units of replication opportunity that is not confounded by variation in transmission intensity. A total of 1279 SARS-CoV-2 genomes from the Nextstrain open dataset, sampled between February 2020 and February 2023, were paired with contemporaneous estimates of cumulative global infections from the Institute for Health Metrics and Evaluation (IHME) to model genomic divergence from Wuhan-Hu-1 as a function of cumulative infections. Divergence increased sublinearly across roughly 14 billion infections over three years, with the most diverged circulating genomes accumulating only about 120 substitutions relative to Wuhan-Hu-1. Extrapolating the fitted relationships, generating the roughly 1000 substitutions that separate SARS-CoV-2 from its closest known sarbecovirus relatives would require on the order of 100-300 billion human infection-equivalents. These requirements would increase substantially in wildlife reservoir hosts, where viral replication per infection is orders of magnitude lower. Conservative estimates of the replication opportunity available in Rhinolophus bat populations imply that the common ancestor of SARS-CoV-2 and its nearest sequenced relatives predates modern virological surveillance by roughly a century. More realistic estimates raise this estimate to millennia, suggesting the closest known relatives of SARS-CoV-1 and SARS-CoV-2 cannot have served as their progenitors.