As I was making the graph from the output on those 200 recent BA.5.2.1 genomes from Georgia, the number of changes to the Orf9b protein stuck out to me.  For such a small protein (97 amino acids versus the 1,273 in Spike) it sure has a lot of mutations.  That could mean something (that its under a lot of selection pressure) or it could be that this isn’t truly area of expertise and I’ve done something wrong, but I don’t think so.  Below I charted for each viral gene the total number of changes in the protein due to mutations divided by the length of the protein – an incredibly simplistic (and perhaps foolish) way to normalize for protein length. Still, kind of interesting. 

Oh, and PS, I should have taken Orf1a and 1b out of this graph because each of those makes lots of smaller proteins, and I didn’t have time to look up the lengths of each… so the fact that those are zero means… zero.