It would be helpful to have phylogenetic trees for the four clades as supporting figures, since as best I understand Figure 2 and 3 do not represent the phylogenetic relationships, but only show the influence of those relationships on the antigenic map. Also, are the phylogenetic trees built on just HA1 or the whole HA protein?
The virus phylogenies can be seen to some extent in Figures 2 and 3, i.e. phylogenetically distinct clades are apparent. However, we agree that it would be helpful to include a cleaner picture of the virus phylogenies. We've included a new figure (current figure 4) that shows more standard time-resolved phylogenies. We've also included discussion of the negative correlation between antigenic drift and phylogenetic diversity.
The trees were inferred from the whole HA gene when available, but for many viruses only HA1 was available. The phylogenetic methods available in BEAST are able to properly deal with this sort of missing data. We've made note of this in the text.