This post is the short form of the now very long post here.  This is just the summary of what I’ve done and what I’m concluding; if you want to see why plus the grimy details, check there.

  1. My question: The regular influenza vaccine this year included an H1N1 virus (A/Brisbane/59/2007 (H1N1)). Will that protect against infection with the new H1N1 strain?
  2. My first question was, How similar is this H1N1 to that in the current vaccine? My answer:  Not very. The vaccine strain is only 79% identical to the present strain, which isn’t terrible but isn’t very good either.  For context, the H3 hemagglutinin in the vaccine is around 45% identical to the H1; we know there’s almost no cross-reactivity between these types.  The HA from the B/Florida/4/2006, a B rather than an A strain and expected to very quite different, is about 30% identical.  See the alignment and the phylogenetic tree in that post.
  3. Are the differences between the viruses especially large at the places where protective antibodies bind?  If so, then we’d expect little cross-protection between the strains.
  4. First, we have to figure out where protective antibodies bind.  A paper from 1982 (Caton et al., (1982). The antigenic structure of the influenza virus A/PR/8/34 hemagglutinin (H1 subtype) Cell, 31 (2), 417-427 DOI: 10.1016/0092-8674(82)90135-0) offered several regions of H1 where protective antibodies might bind.
  5. I ran a script on a few hundred different H1 proteins from different viruses, and determined that the regions Gerhard’s group said bind to protective antibodies tend to be highly variable between viruses.  This is presumably the result of selection at that region to avoid antibody responses. So this is consistent with Gerhard’s conclusion and I’m willing to go with those regions as antibody binding sites.
  6. I compared the amino acids in the regions Gerhard flagged between the vaccine strain, and the new H1N1 strain. If those regions are very similar then I would expect good cross-protection; if very different, I would expect little cross-protection.  The result (shown below) is that there are extensive differences between the putative protective antibody binding sites.

Conclusion (based on lots of assumptions!):

You can see that of the five antibody-binding sites (Sa, Sb, Ca1, Ca2, and Cb), four are really very different, while one is quite similar.  Even in that one,  the single difference, from Ser to Pro, is a drastic change that would probably significantly reduce antibody binding.  So most antibodies wouldn’t bind well to the new H1N1.  However, the most similar region (”Sa”) is the one that Gerhard flagged as the most important for antibody binding, so I’m leaning to the concept that there probably will be a little bit of cross-protection, but not a lot.

But what I would really like to see is an actual experiment, testing cross-protection and cross-reactivity between the epidemic and the vaccine H1N1.