Early evolution of BA.2.86 sheds light on the origins of highly divergent SARS-CoV-2 lineages

Since the emergence of SARS-CoV-2 in late 2019, highly divergent variants with novel constellations of mutations have periodically emerged to displace previously co-circulating virus lineages. The evolutionary mechanisms behind this process remain unclear, but the prevailing hypothesis supports emergence through single events followed by a rapid accumulation of mutations in individual long-term infections. In July 2023, the Omicron-descending BA.2.86 Variant of Interest (VOI) emerged to outcompete other virus lineages, with descending sublineages still dominating as of December 2024. We use BA.2.86 as a case study to test the hypothesis that highly divergent variants emerge during a single chronic infection. By applying a comprehensive suite of evolutionary analyses to independent high-resolution datasets, we identify a cluster of approximately 100 BA.2 genomes that fall directly ancestral and shorten the branch leading to BA.2.86, carrying some BA.2.86-specific lineage defining mutations. These genomes, sampled from multiple countries and at different timepoints, share a single ancestor dating back to December 2021. Additionally, we detect over 2600 earlier BA.2, BA.2.75, and XBB.1.5 genomes with multiple mutations later associated with BA.2.86. These genomes represent a standing genetic variation that likely facilitated a gradual evolution in the direct ancestry of this variant. We further uncover a complex evolutionary process involving recombination between the BA.2.75 and XBB.1.5 parental lineages. Together, our results challenge the notion of a single emergence event, and instead support a more complex scenario, with BA.2.86 likely having evolved over multiple transmission chains.