Figure 1.
Antigenic tree for influenza A (H3N2) viruses.
Branch lengths represent antigenic distances (maximum of up- and down-weights for each branch) inferred from a maximum likelihood tree of 258 hemagglutinin sequences of seasonal influenza A (H3N2) virus isolates and serological data. (A) Colored edges show antigenic type transitions, with internal branches with high average antigenic weights (≥1.0 antigenic units). Gray-blue edges represent high weight branches leading to a subtree with three isolates or less, representing low abundance types. (B) Isolates are color-coded by antigenic clusters according to Smith et al. (2004). Three isolates (A/Christchurch/4/85, A/Hong Kong/34/90 and A/Netherlands/172/96) are only present as antisera and were not assigned a cluster label.
Table 1.
Internal branches with high average antigenic weights (≥1.0 antigenic units) and according antigenic types in comparison to antigenic clusters identified by antigenic cartography (branches leading to three or less isolates are excluded).
Figure 2.
Schematic drawing demonstrating the up/down tree concept.
For the two taxa t2 and t4, no antiserum is present, and thus, b3 and b6 only have up-weights. A path from t1 to t3 would use the up-weights of branch b1 and b2, and the down-weights of branch b4 and b5. Similarly, the path from t2 to t1 would use the up-weight of branch b3 and the down-weight of branch b2. Notably, the path from t1 to t1, namely the antigenic distance from antigen t1 to the antiserum raised against strain t1, would use the up-weight and the down-weight of branch b1.