Announcement

**super_flu** · July 20, 2009, 03:51 AM

Re: Don't Blame Birds for 1918 Flu

Originally posted by gsgs View Post

I'm not finding anything that indicates that data other than
mere difference-counts of virus-pairs in the segments are being
used. So, presumably just the data which I gave in the tables above.

But apparantly more sophisticated analysis of trends and timings

however, not influenza-specific

they estimated August 2008 - Jan 2009 as the likely start of ******
leading to headlines in the press that it may have been
circulating since August

Quite different. Give it another read. The paper is freely available at the link I posted.

In this case rates that determine the branch lengths are drawn from a parametric distribution where the mean of which can be determined apriori or estimated from the sampled trees.

By relaxing the constraints poorly sampled data or even increased rates due to interspecies transmission events can be accounted for in the date estimates.

Since you're a math guy I think you'll appreciate the methods section in the paper. Lots of fun symbols and hyperparameters of hyperparameters, but, in my opinion, elegant.

While not influenza specific it certainly was designed with analysing fast evolving viruses in mind.

**gsgs** · July 20, 2009, 04:21 AM

Re: Don't Blame Birds for 1918 Flu

long paper, no good summary.
I had a brief look and still didn't find anything
to indicate they use something else than just
(nucleotide-)differences.
But then we just have the data, shown above.
However sophisticated the method, it is hard to imagine
that they can conclude more than what you can just
see or picture/graph when comparing how other strains
evolve, e.g. H3N2 or H5N1.
Test it to predict their (known) reassortments by ignoring
one decade or just some years !

Maybe a better model looks (also) at individual polymorphisms ?

it's not specific to segmented viruses which reassort and have
different rates in the segments, depending on immunity
and thus host-species. Which have codon-bias and more
A,T nucleotides in mammals.
Which preferrably reassort HA and NA, which have seasonality,..

**gsgs** · July 26, 2009, 08:31 AM

Re: Don't Blame Birds for 1918 Flu

most obvious reassortment events in H3N2 :

new segments,year,virus,comment

24,1968, A/HK/1/68, H2N2 acquired avian H3-HA, causing a pandemic
1245,1969, A/HK/3/69
1345,1971, A/HK/46/71
278,1972, A/Udorn/72
146,1972, A/HK/50/72
26,1993, A/NY/758/93
17,1998, A/NY/521/98
356,1999, A/NY/332/99
38,2000, A/NY/187/00,
4,2003, A/NY/213/03, Fujian strain acquired a new HA, escaping the
vaccine and causing a severe season in USA
1,2004, A/Queensland/47/04, Fujian virus acquired PB2 from
non-Fujian H3N2 but didn't spread very much

**HenryN** · July 26, 2009, 09:44 AM

Re: Don't Blame Birds for 1918 Flu

Originally posted by gsgs View Post

most obvious reassortment events in H3N2 :

new segments,year,virus,comment

4,2003, A/NY/213/03, Fujian strain acquired a new HA, escaping the
vaccine and causing a severe season in USA
1,2004, A/Queensland/47/04, Fujian virus acquired PB2 from
non-Fujian H3N2 but didn't spread very much

What did 2002 Korean isolates acquire?

**gsgs** · July 27, 2009, 03:08 AM

Re: Don't Blame Birds for 1918 Flu

that was just copied from this thread:

http://www.setbb.com/fluwiki2/viewtopic.php?t=339&mforum=fluwiki2

database from 2008, only complete genomes, searching for
reassortments, not recombinations

**gsgs** · August 19, 2009, 11:18 AM

Re: Don't Blame Birds for 1918 Flu

see here, how current 1918-related viruses are closer to
human viruses from the 1930s in segments 1,2,7,8 but
closer to swine viruses from the 1930s in segments 4,5,6 ,
while segment 3 is almost a tie.

(well, it mutates more in HA in humans than in swine)

Code:

  2 >A/Index/early human/~1935(H1N1)            
  1: 77,159, 62,478, 73,518, 37, 60   A/Index/1968(H3N2)
  2:  0,  0,  0,  0,  0,  0,  0,  0   A/Index/early human/~1935(H1N1)
  3: 32, 21, 26, 51, 24, 38, 15, 17   A/Brevig Mission/1/1918(H1N1)
  4: 58, 57, 58, 76, 49, 70, 37, 42   A/Index/early swine/~1935(H1N1)
  5:  0,  0,  0,213,  0,187,  0,  0   A/Anas crecca/Spain/1384/2007(H1N1)
  6: 97, 84,114,  0,135,500, 51, 66   A/Index/7 birds similar to BM18/~1975
  7:117,155,147,439,132,436, 57, 67   A/pintail/Alaska/20/2005(H12N5)
  8:123,154,141,488,132,494, 61,281   A/mallard/Interior Alaska/2/2007(H4N6)
  9:119,154,138,481,130,420, 62,289   A/pintail/Alaska/279/2005(H3N8)
 10:117,153,138,483,130,422, 62,289   A/mallard/Alaska/715/2005(H3N8)
 11:128,106,139,  0,153,420, 89,119   A/duck/LA/17G/1987(H3N8)
 12:125,158,145,482,160,427, 85, 91   A/swine/Chibi/01/2005/12/15(H3N8)
 13:103,151,132,480,146,430, 65, 76   A/equine/Uruguay/1/1963(H3N8)
 14:112,151,133,481,154,479,145, 83   A/equine/Kentucky/1a/1975(H7N7)
 15:172,182,180,481,164,480,143,171   A/equine/Lexington/1/1966(H7N7)

  4 >A/Index/early swine/~1935(H1N1)            
  1:110,163,102,482, 94,510, 66,100   A/Index/1968(H3N2)
  2: 58, 57, 58, 76, 49, 70, 37, 42   A/Index/early human/~1935(H1N1)
  3: 35, 38, 36, 32, 31, 45, 32, 32   A/Brevig Mission/1/1918(H1N1)
  4:  0,  0,  0,  0,  0,  0,  0,  0   A/Index/early swine/~1935(H1N1)
  5:  0,  0,  0,193,  0,172,  0,  0   A/Anas crecca/Spain/1384/2007(H1N1)
  6:106, 94,111,  0,127,500, 61, 78   A/Index/7 birds similar to BM18/~1975
  7:123,159,147,439,128,450, 69, 79   A/pintail/Alaska/20/2005(H12N5)
  8:127,161,139,492,127,499, 74,284   A/mallard/Interior Alaska/2/2007(H4N6)
  9:124,156,139,482,124,422, 65,290   A/pintail/Alaska/279/2005(H3N8)
 10:123,155,139,484,126,422, 65,290   A/mallard/Alaska/715/2005(H3N8)
 11:128,111,136,  0,144,418, 94,135   A/duck/LA/17G/1987(H3N8)
 12:132,164,144,479,160,421, 88,103   A/swine/Chibi/01/2005/12/15(H3N8)
 13:112,156,130,477,140,430, 75, 83   A/equine/Uruguay/1/1963(H3N8)
 14:120,160,130,490,152,475,136, 92   A/equine/Kentucky/1a/1975(H7N7)
 15:178,179,182,489,156,478,134,184   A/equine/Lexington/1/1966(H7N7)

**gsgs** · October 4, 2009, 02:46 AM

Re: Don't Blame Birds for 1918 Flu

segment 1 is not clear, some bird-viruses are closer to the swine
origin and some are closer to the human origin in segment 1

differences of some selected viruses to early swine - early human

so (-) is closer to swine and (+) is closer to human

these differences are (more) evidence that BM1918 is not
the direct ancestor of both, the early human and the early
swine viruses

(differences in promille of different nucleotides)

Code:

 -9  +2  -5      -8      +8 +14 >A/Dk/UKR/1/1963(H3N8)
 -5  +3  +8      -9      +4 +15 >A/duck/Czechoslovakia/1956(H4N6)
 -6  +5  +3      -8      +6 +17 >A/duck/England/1956(H11N6)
 -5  -2  -5      +2      +7     >A/Ck/Germany/n/1949(H10N7)
 +1  +5  +1      -7     +11  +7 >A/fowl/Dobson/1927(H7N7)
 +1  +8  +2      -3  -8  +2 +13 >A/fowl/Rostock/45/1934(H7N1)
 -5  +3  +6      -5  -5  +3 +17 >A/Ck/Scotland/1959(H5N1)
-11  +7  +1      -5 -14 +12 +13 >A/human/China/GD02/06(H5N1)
-12  +8  -2      -6 -13 +10 +13 >A/Bar-h Gs/Qinghai/62/05(H5N1)
 -3  +3  +0 -13  -1 -14  +7 +16 >A/swine/Germany/2/1981(H1N1)
 -1  -1             -21 +10     >A/******/index/2009-02-01
 +9  +5  -2      -6     +10  +7 >A/equine/Uruguay/1/1963(H3N8)              
 +6  +2  +1      -4      +3  +1 >A/mallard/Alaska/715/2005(H3N8)            
 +5  +2  +1      -6      +3  +1 >A/pintail/Alaska/279/2005(H3N8)            
 +4  +7  -2      -5     +13  +3 >A/mallard/Interior Alaska/2/2007(H4N6)     
 +8  +8  -2      -7     +13     >A/duck/Memphis/546/1974(H11N9)             
            -20     -15         >A/Anas crecca/Spain/1384/2007(H1N1)        
 +9 +10  -3      -8     +10 +12 >A/Index/7 birds similar to BM18/~1975      
 +6  +4  +0      -4     +12 +12 >A/pintail/Alaska/20/2005(H12N5)            
 +0  +5  -3      -9      +5 +16 >A/duck/LA/17G/1987(H3N8)                   
------------------------------------------------------------
      +       -   -   -   +   +

the Smith et.al paper says on page 2:
> preliminary phylogenetic analysis showed that the BM/1918
> virus H1,N1,PB1,PA,NP genes clustered with human H1N1
> influenza A viruses, whereas in PB2,M and NS genes
> clustered with swine.

**gsgs** · October 4, 2009, 02:57 AM

Re: Don't Blame Birds for 1918 Flu

----------------------------

does someone understand table 1 ?
what's "nodes"
do they claim all human H1N1 descended directly from BM18
in 2,3,5,6 but from human flu in the other segments with
reassortments in ~1910,1913,1911,1915 (expectation value) ?
the years in brackets give their 95% confidence interval ? ---edit: yes, 95%,
they call it Bayesian credible interval(BCI)----------

do they claim classical swine (e.g. SW/31) descendd directly
from BM18 in 1,7,8 but from another virus(es) in 2,3,4,5,6
with reassortment between 1918 and 1931 ?

what's TMRCA of one virus (BM18), node 2
what's swine/human in node 1, they didn't mix after 1918, did they

-------------------------------------

OK, the nodes must be the numbers in the phylo-charts
of figure 1, each segment has its own chart and the numbers are included

Just a moment...

http://www.pnas.org/content/early/2009/07/10/0904991106.full.pdf

node 1 = common ancestor of early classical swine(1931) and early human H1N1 (1935)
node 2 = common ancestor of classical swine and BM/1918
node 3 = common ancestor of early human and BM/1918

(?)

----------------------------------

keywords:

Zitation
Discussion
Dating the emergence of pandemic influenza viruses.
Smith GJ, Bahl J, Vijaykrishna D, Zhang J, Poon LL, Chen H, Webster RG, Peiris JS, and Guan Y

Our results indicate that genetic components of the 1918 H1N1 pandemic virus circulated in mammalian hosts

1918

A/Brevig Mission/1/1918(H1N1)

origin

-------------------------------------------

Smith et al get for the TMRCA

sw-hu:1881,1906,1904,1885,1884,1907,1884,1899
BM-hu:1903,1914,1914,1916,1909,1913,1896,1908
BM-sw:1910,bm18,bm18,1913,bm18,bm18,1911,1915

I get:
Sw-BM : 1911,1913,1910,1914,1915,1916,1912,1911
Hu-BM : 1912,1918,1914,1916,1916,1916,1916,1917
Hu-Sw : 1908,1908,1908,----,1908

Announcement

Don't Blame Birds for 1918 Flu

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment