Tweet
Re: Science paper: The Global Circulation of Seasonal Influenza A (H3N2) Viruses
antigenic data doesn't say much about the evolution.
80% of mutations are synonymous and thus missed by antigenic analysis.
Most of the data is 2002-2006, they don't have 2007 yet, let alone 2008.
This period was disturbed by the appearance of the Fujian strain ,
especially in HA, so, the spreading pattern is even harder to see.
I haven't seen the pictures and tables, but I don't feel so
strongly about their conclusions yet.
It would be useful to have a few genomes at least each year from
South America, China, Africa, Europe.
More suitable to study the evolution seems to be the
H3N2-genome data 1993-2007 from USA which well covers each
year, except a few H1N1-years.
This data is better distributed over time and covers the whole genome,
not only HA1. OTOH it's only USA.
One idea is to examine each polymorphism, each change from
Udorn/72 separately. Where it first appeared before it
became established in a substrain. Then make a statistics
of the timing within the year.
One idea is to examine all sequence-pairs and calculate the
time of their common anchestor. If most such anchestors
of USA-sequences are dated in summer, then they probably were
seeded from the Southern Hemisphere.
I implemented it for 800 US-genomes, see the picture here:
I haven't examined it yet, but I guess, the large peaks in summer 1996,1997
are from the Fujian strain in 2003, whose last common anchestor with
the co-circulating normal,old H3N2 was calculated to have lived in 1996
----ahh, no. 2003-1996 is more than 4 years, so this should have been excluded
-----------------------------
full H3N2-genomes from genbank, November-April isolates 1993-2007 in USA:
average distance between 2 H3N2-viruses in consecutive seasons in USA since 1993 was 179
average distance between 2 H3N2-viruses two seasons apart in USA since 1993 was 217
average accumulation of differences since 1968 was 30 per year.
antigenic data doesn't say much about the evolution.
80% of mutations are synonymous and thus missed by antigenic analysis.
Most of the data is 2002-2006, they don't have 2007 yet, let alone 2008.
This period was disturbed by the appearance of the Fujian strain ,
especially in HA, so, the spreading pattern is even harder to see.
I haven't seen the pictures and tables, but I don't feel so
strongly about their conclusions yet.
It would be useful to have a few genomes at least each year from
South America, China, Africa, Europe.
More suitable to study the evolution seems to be the
H3N2-genome data 1993-2007 from USA which well covers each
year, except a few H1N1-years.
This data is better distributed over time and covers the whole genome,
not only HA1. OTOH it's only USA.
One idea is to examine each polymorphism, each change from
Udorn/72 separately. Where it first appeared before it
became established in a substrain. Then make a statistics
of the timing within the year.
One idea is to examine all sequence-pairs and calculate the
time of their common anchestor. If most such anchestors
of USA-sequences are dated in summer, then they probably were
seeded from the Southern Hemisphere.
I implemented it for 800 US-genomes, see the picture here:
I haven't examined it yet, but I guess, the large peaks in summer 1996,1997
are from the Fujian strain in 2003, whose last common anchestor with
the co-circulating normal,old H3N2 was calculated to have lived in 1996
----ahh, no. 2003-1996 is more than 4 years, so this should have been excluded
-----------------------------
full H3N2-genomes from genbank, November-April isolates 1993-2007 in USA:
average distance between 2 H3N2-viruses in consecutive seasons in USA since 1993 was 179
average distance between 2 H3N2-viruses two seasons apart in USA since 1993 was 217
average accumulation of differences since 1968 was 30 per year.
Comment