Picking influenza stem strains

Pipeline

I chose 656 H3N2 viruses that largely overlap with the flux analysis. These viruses were selected from the full set of strains from both IRD and GISAID, preferring strains with more HI data, longer sequences and precise dates. I selected at most 50 strains per year from 1994 to 2011. This left 447 strains with the following distribution. Nucleotide sequences are 1715 bases and amino acid sequences are 566 residues. I matched HI data to these strains, giving 10,281 titers, comprising 447 virus isolates and 478 serum isolates. Only sera that were tested against 5 or more different viruses were kept.

Epitopes sites were determined following Munoz and Deem 2005 (“Epitope analysis for influenza vaccine design”). In this case, we have 437 non-epitope residues and 129 epitope residues.