A Phylogenetic Method Identifies Candidate Drivers of the Evolution of the SARS-CoV-2 Mutation Spectrum

The molecular processes that generate new mutations evolve, but the causal mechanisms are largely unknown. In particular, the relative rates of mutation types (e.g. C > T), the mutation spectrum, sometimes vary among closely related species and populations. I present an algorithm for subdividing a phylogeny into distinct mutation spectra. By applying this approach to a SARS-CoV-2 phylogeny comprising approximately 8 million genome sequences, I identify ten shifts in the mutation spectrum. I find strong enrichment consistent with candidate causal amino-acid substitutions in the SARS-CoV-2 polymerase, and strikingly three appearances of the same homoplasious substitution are each associated with decreased C > T relative mutation rates. With rapidly growing genomic datasets, this approach and future extensions promise new insights into the mechanisms of the evolution of mutational processes. Keywords: Mutation Spectrum; Phylogenetic Analysis; SARS-CoV-2 Evolution.