Just to add some detail to the PJ.2.1 picture outlined in the thread below, here's a Venn diagram (modeled on Cov-Spectrum's) of the four major PJ branches' amino acid mutations (excluding all shared, inherited MC.10.1.7 mutations).

Large upload from Ontario yesterday, providing some striking revelations about the evolution of PJ.2.1. First, there is yet another strikingly divergent branch of PJ.2.1. This isn't one variant: it's four. More and more, this resembles the emergence of BA.1/2/3. What is happening? 1/

➡️RD* and SN* stood out: they showed either no such age preference or a much less pronounced one. However, data for these subvariants remain limited.

➡️RE* had the most entries (998) and showed the expected age preference, with children aged 0-9 preferentially infected. Most other subvariants showed a broadly similar pattern.

We demonstrated that BA.3.2 preferentially infects children aged 0-9. However, our initial analyses did not examine BA.3.2 subvariants separately. Here are the additional results, based on all BA.3.2 sequences deposited in GISAID in 2026:

Effects of XFG-adapted mRNA vaccination on SARS-CoV-2 variant neutralization - see our new preprint. Georg Behrens, Markus Hoffmann & teams analyzed neutralizing antibody responses in individuals who received an XFG-adapted mRNA booster, recommended for the 2026/2027 COVID season.

There was also a partial PJ.2.1 sequence recently in German wastewater.

Since Aug 31, 12.8% (10/78) airport surveillance seqs from international travelers to the US have been PJ.2.1. 45% (9/20) from Latin American travelers in that period are PJ.2.1: 4 from Cuba & 1 each from Chile, Colombia, the DR, Ecuador, & Guatemala. PJ.2.1 is the fastest variant in the world now.

Importantly, even in Africa’s much younger population, the age preference of BA.3.2 clearly differs from that of other variants (data 01/2026-09/2026):