Analyzes 103 early SARS-CoV-2 genomes and identifies two linked, defining positions (position 8782 and 28144) that split the earliest sampled diversity into two named “types,” L (~70% of the sample, more common in Wuhan before 7 Jan 2020) and S (more similar to bat/pangolin outgroup coronaviruses, interpreted as ancestral). Argues L was more prevalent in the earliest phase and increased in frequency over the following weeks, and speculates this reflects stronger selection/transmission for L. This is the earliest influential attempt to split the early genomic diversity into two lineages and argue about their relative timing/ancestry and fitness — direct ancestor of the later, much more rigorous lineage-A/B two-introduction framework (S-1) though the L/S split does not map cleanly onto the A/B split used later. The “aggressive L type” framing was quickly and heavily criticized on methodological grounds (see S-6).
relevance_note: origin of the “two founding lineages, one more successful” framing that Pekar et al. later formalize; also a cautionary tale about premature functional claims from early phylogenetic snapshots.