This plot relates to the tree shown in Fig. 2. The regression is exclusively fitted to genomes sampled between 2014 and 2015. The same colours are used for the data points as in Fig. 2. The dashed yellow lines highlight how the 2021 data points deviate from the relationship between sampling time and sequence divergence. According to this relationship, about 95 substitutions (95% prediction interval: 88–101) are expected on the branch ancestral to the 2021 cluster, whereas only 12 are inferred on this branch.