Homologous Recombination is Very Rare or Absent in Human Influenza A Virus

[HenryN]

147

RESULTS AND DISCUSSION

148

Two of the ten human influenza A/H3N2 virus data sets (PB2 and NP) analyzed here

149

contained sequences with statistically significant mosaic structure, as determined by 3SEQ, and

150

with putative recombinant sections that were each sufficiently long (>100 nt) that they could be

151

re-analyzed by phylogenetic recombination detection methods. Three of the remaining eight

152

A/H3N2 data sets (PA, NA, MP) and one of the A/H1N1 data sets (NA) also resulted in 3SEQ p153

values that revealed a strong signal of mosaicism, but in all these cases the inferred breakpoints

154

were either close to the gene segment’s endpoints, or very close to each other, making it

155

impossible to infer a credible phylogeny. The remaining five A/H3N2 data sets (PB1, HA,

156

HA413, NS, NA413 – the 413-suffix meaning that it is the HA or NA data set containing 413

157

sequences) and seven of the A/H1N1 data sets (PB2, PB1, PA, HA, NP, MP, NS) did not

158

contain any statistically significant mosaic signals that survived a Dunn-Šidák correction in

159

3SEQ. Recombination analysis results are summarized in Table 1 for A/H3N2 and Table 2 for

160

A/H1N1. The two putative recombinant data sets are discussed in more detail below.

161

The H3N2 PB2 data set assembled here contained 912 distinct sequences, one of which

162

– A/New York/11/2003 – statistically supported a mosaic structure with both mosaic regions

163

longer than 100nt. The two most likely parental sequences, identified as A/Hong Kong/14/1974

164

(major parent) and A/New York/424/1999 (minor parent), revealed a strong mosaic signal

165

(corrected p = 0.013) in relation to A/New York/11/2003. However, while the phylogenies

166

inferred for the minor (positions 202–2189, Figure 1a) and major (positions 1–201 and 2190–

167

2347, Figure 1b) segments revealed topological movement of the putative recombinant

168

sequence relative to the parental sequences, a general lack of phylogenetic resolution, reflected

169

in low levels of bootstrap support (particularly in the major segment), meant that there was

170 insufficient signal to infer phylogenetic incongruence. Since support for phylogenetic

incongruence is necessarily made up of two components – the phylogenetic

171 relationship among

172

the parents and recombinant on the major and minor trees – we call the signal ‘weak’ if one of

173

the components receives only low bootstrap support.

174

For the NP data set of A/H3N2 viruses, a single sequence – A/Christchurch/14/2004 –

175

supported a mosaic structure with both candidate recombinant regions longer than 100 nt. The

176

candidate parental sequences identified by 3SEQ were A/Beijing/1/1968 as the major parent

177

and A/New York/153/1999 as the minor parent (clonality among these three isolates is rejected

178

at corrected p = 0.032). The ML tree for the region 98-1454 is presented in Figure 2a while that

179

for regions 1-97 and 1455-1570 is shown in Figure 2b. In these phylogenies, the putative

180

recombinant sequence was clearly more closely related to a different parent in each sequence

181

region. The phylogenies also revealed sequence A/New York/381/2004 as a better candidate

182

for the minor parent than A/New York/153/1999; the mosaic signal when assuming A/New

183

York/381/2004 as the minor parent in the recombination event was still strong (corrected p =

184

0.052). However, as in the PB2 data set, the lack of bootstrap support in the phylogeny inferred

185

for the major segment indicates that there is in reality an insufficiently strong signal for

186

phylogenetic incongruence to conclude that homologous recombination has occurred.

187

For the two candidate recombinants A/New York/11/2003 (PB2) and

188

A/Christchurch/14/2004 (NP), it is also puzzling that the parental sequences were sampled 25

189

and 31 years apart, respectively. Hence, for one of these recombination events to have

190

occurred, a lineage of viruses closely related to an ‘archaic’ virus (either A/Hong Kong/14/1974

191

or A/Beijing/1/1968) must have circulated until at least 1999 and recombined with A/New

192

York/424/1999 or A/New York/153/1999. Given the rapid rate of influenza A virus mutation

193

through frequent RNA polymerase error, as well as the rapid lineage turnover driven by positive

194

selection on the major antigenic proteins (6, 11, 12, 23), this scenario seems extremely unlikely.

195 Thus, laboratory error, such as template switching during amplification in a mixed or

contaminated sample, is a likely explanation of these apparent homologous

196 recombination

197

events.

198

In sum, our study has revealed that no sequence of human influenza A virus contains a

199

clear signature of phylogenetic incongruence indicative of the action of homologous RNA

200

recombination. Given that more than 10,000 distinct sequences were analyzed, this constitutes

201

strong evidence that homologous recombination plays only a very minor role, if any, in the

202

evolution of human influenza A virus. More generally, the occurrence of phylogenetic

203

incongruence does not in itself constitute conclusive evidence for this process. Specifically,

204

because our analysis is necessarily based on viral consensus sequences rather than the myriad

205

individual viral molecules that characterize any infection, it is equally plausible that the

206

‘recombinants’ detected here in fact represent cases of mixed infection in individual hosts

207

followed by the amplification and sequencing of different viral molecules, thereby producing

208

laboratory-generated artificial recombinants. Hence, to demonstrate conclusively the

209

occurrence of homologous recombination in influenza A virus it will be necessary either to clone

210

(or plaque purify) and sequence multiple viral genomes from an individual host and demonstrate

211

the presence of the recombinant and both parental genotypes within the sample (1), or to show

212

that recombinant sequences form a distinct circulating lineage, with readily identifiable parents,

213

that is transmitted among multiple individuals in a population (30).

214

Finally, although there were 315 sequences in the data analyzed here that carried a

215

strong mosaic signal as identified by 3SEQ, it was impossible to verify the vast majority of these

216

as recombinants since the putative recombinant regions were too short to infer a credible

217

phylogenetic history. It is therefore possible that homologous recombination, should it occur in

218

influenza A virus, more commonly involves the transfer of very short sections of RNA, a process

219

that would be undetectable by the majority of other methods devised to detect recombination. If

220

homologous recombination of short segments is determined to be a relevant process in

221 influenza A virus evolution, the basis of our more frequent observation of mosaicism in A/H3N2

viruses compared to A/H1N1 viruses will need to be investigated further. However,

222 by far the

223

strongest signal in the influenza A virus sequence data analyzed here is that of strict clonality,

224

supporting most models of influenza virus evolution proposed to date.

225
226

ACKNOWLEDGEMENTS

227

The research undertaken in this study was funded in part by Resources for the Future (MFB),

228

NIH/NIGMS grant P50GM071508 (MFB), National Institutes of Health Grant GM28016 (MFB),

229

NIH grant number GM080533-01 (ECH), and the Intramural Research Program of the NIH, and

230

the NIAID (JKT). We thank John Zollweg and Linda Woodard at the Cornell University Center

231

for Advanced Computing for suggesting algorithmic improvements to 3SEQ, as well as two

232 anonymous reviewers for helpful suggestions

[Malcolm]

We need a translation of the post above into easily understood language.

It would be interesting to have a discussion of the paper's stated rarity of recombination as opposed to the reality of recombination's daily occurrence.

[HenryN]

Influenza A viruses are a major cause of respiratory disease in humans,

43 responsible for

44

36,000 annual deaths in United States alone (7, 28) and occasional widespread pandemics

45

associated with much higher levels of mortality and morbidity (27). The viral genome is

46

comprised of eight negative-strand RNA segments, with a combined length of ~13.6 kb, that can

47

evolve through a variety of mechanisms. Most notably, the lack of a proof-reading mechanism

48

during RNA replication results in a high frequency of point mutations which, when combined

49

with large population sizes and short generation times, gives influenza A virus the ability to

50

generate quickly both antigenic variants that can escape host immunity – a process termed

51

antigenic drift (5, 29) – as well as genotypes that provide resistance to anti-viral agents such as

52

the adamantanes (9) and neuraminidase inhibitors (2). In addition to generating genetic

53

diversity by rapid mutation, when multiple viruses co-infect a single cell the eight segments of

54

the influenza virus genome can reassort and yield progeny virions with a novel combination of

55

segments – a process termed antigenic shift. Such reassortment is well documented among

56

those viral strains that differ in their host species, such as humans and birds. Reassortment of

57

this type, involving the acquisition from avian hosts of new polymerase PB1, hemagglutinin

58

(HA), and/or neuraminidase (NA) segments to which there was no prior human immunity,

59

played a major role in the genesis of the human influenza pandemics of 1957 and 1968 (15, 22).

60

More recently, intra-subtype reassortment has also been shown to occur frequently among co61

circulating human H3N2 influenza A viruses (14, 18), which may also impact ongoing antigenic

62

evolution (14). In addition to reassortment among RNA segments, intragenic recombination

63

between different RNA segments, commonly referred to as non-homologous recombination (3,

64

20, 25), as well as intragenic recombination between viral RNA and exogenous RNA (16) have

65

been observed and may possibly play a role in determining pathogenicity (25).

66

More controversial, however, is the occurrence of homologous recombination in

67

influenza viruses, most likely involving copy-choice (template-switching) replication of RNA

68 molecules that co-infect a single cell. Although bioinformatic evidence for homologous

recombination has been suggested (13, 19), these results remain unsubstantiated,

69 with

70

extensive lineage-specific rate variation a likely source of a false-positive signal for at least

71

some putative recombination events (24, 31). Indeed, because the genomic RNA generated

72

during replication is rapidly packaged with ribonucleoprotein, which will act to prevent the

73

occurrence of template-switching that is central to copy-choice replication, homologous RNA

74

recombination is thought to occur rarely, if at all, in both influenza viruses (17), and negative75

strand RNA viruses in general (8). In particular, a comprehensive phylogenetic analysis of

76

recombination in negative-sense RNA viruses found only sporadic evidence for recombination,

77

and not among influenza viruses (8), although the process was recently demonstrated in Zaire

78

ebolavirus, an unsegmented negative-sense single-stranded RNA virus (30). If proven to occur,

79

homologous recombination would facilitate two evolutionary processes in influenza virus: the

80

purging of deleterious mutations and the rapid generation of novel genotypes, potentially

81

including new antigenic and drug-resistant variants.

82

To assess whether homologous recombination has played a role in shaping the genetic

83

diversity of human influenza A virus we compiled a data set of 13,852 sequences representing

84

all eight RNA segments of isolates of A/H1N1 and A/H3N2 subtypes. Using an exhaustive

85

search method (4), we statistically assessed the possibility of every potential two-breakpoint

86

homologous recombination event, considering each sequence as a possible recombinant and

87

searching over all possible parents and all possible breakpoints. In our data set, this translated

88

to considering over seven billion sequence triplets, where two of the sequences in each triplet

89

are posited to have recombined to form the third sequence in the triplet. For those sequences

90

identified by this method to contain putative recombinant sections longer than 100 nucleotides

91

(nt), we used more stringent phylogenetic methods to further verify that they contained an

92

evolutionary signal (i.e. phylogenetic incongruence) compatible with the action of homologous

93 recombination.

[HenryN]

MATERIALS

95 AND METHODS

96

Sequence data. Nucleotide sequences of human influenza A virus were obtained from

97

the Influenza Virus Resource (http://www.ncbi.nlm.nih.gov/genomes/...se/select.cgi)

98

and aligned using MUSCLE (10). Sixteen sets of sequences, two for each RNA segment, were

99

obtained by downloading all of the full-length subtype A/H3N2 and subtype A/H1N1 sequences

100

generated through the NIH/NIAID Influenza Genome Sequencing Project. In addition, two

101

previously published data sets comprising 413 HA and NA segments of human A/H3N2 viruses

102

were also included in the analysis (18). After removing duplicate sequences that were identical

103

at the nucleotide level, a final subset of 10,492 sequences was analyzed.

104

Recombination analysis. As an initial screen for possible recombination, each of the

105

18 data sets was first analyzed using the 3SEQ program (4). 3SEQ tests all possible two106

breakpoint recombination events for each triplet of sequences in the data set, assigns a

p-value

107

(rejecting clonality) to each sequence triplet, and infers breakpoints. Breakpoint pairs are found

108

using a parsimony criterion, with the most likely breakpoint positions being those that minimize

109

the number of mutations between the putative recombinant sequence and a two-breakpoint

110

mosaic of the parental sequences. Breakpoint pairs are reported as ranges of nucleotide sites

111

since there are multiple pairs of breakpoints that can satisfy this parsimony criterion. 3SEQ

112

reports a p-value by calculating the exact probability that this type of recombination signal would

113

be observed under the null hypothesis of clonal (non-recombinant) evolution. Finally, all p114

values are corrected with a Dunn-Šidák correction for the large number of triplets tested. If a

115

particular sequence triplet had a corrected p < 0.05, and if the inferred breakpoints guaranteed

116

that the shortest possible recombinant segment was longer than 100nt (which we deemed

117

suitable for phylogenetic analysis), a secondary phylogenetic analysis of the data was used as

118

an independent verification of putative homologous recombination identified among these

119 sequence triplets. Given that 3SEQ is one of the most powerful methods for detecting

recombination (4) and is the only method available that can scan hundreds

120 of sequences at a

121

time and identify the candidate recombinants with breakpoints and p-values, it is an appropriate

122

method for detecting recombination in large data sets of influenza A virus. However, although

123

simulations show that 3SEQ is generally robust to false-positive results (4), lineage-specific rate

124

variation can generate apparent recombinants that triplet methods (like 3SEQ) detect as real

125

recombinants.

126

To minimize the possibility of false-positive results, we performed a secondary

127

phylogenetic analysis of recombination in our data sets of influenza A virus. For each putative

128

recombinant (or set of recombinants with the same breakpoints), the entire data set alignment

129

was divided at the breakpoint positions established by 3SEQ. If two recombination breakpoints

130

were found in a single sequence, the sequence region between the breakpoints is denoted the

131

‘minor’ region, generated by the minor parent, and the remainder referred to as the ‘major’

132

region, generated by the major parent. Because of the very large size of the data sets in this

133

study, initial neighbor-joining (NJ) phylogenetic trees were inferred using the PAUP* package

134

(26) on either side of the putative breakpoints. If evidence for phylogenetic incongruence was

135

apparent due to a change in topological position of specific sequences, a more detailed analysis

136

using maximum likelihood (ML) phylogenetic trees was undertaken. In this case

137

phylogenetically representative sequences, along with those closely related to the putative

138

recombinants, were selected from the data sets to comprise a final data set of 30-40 sequences

139

on which rigorous phylogenetic analyses could be undertaken using the breakpoints determined

140

using 3SEQ. For these analyses, the best-fit model of nucleotide substitution was determined

141

using MODELTEST (21) (details available from the authors on request) and phylogenetic trees

142

were inferred under this model using the ML method available in PAUP* (26), employing TBR

143

branch-swapping in each case. Finally, to assess the degree of support for the differing

144

phylogenetic positions of each putative recombinant, a bootstrap re-sampling analysis was

145 undertaken using 1000 replicate NJ trees inferred under the best-fit substitution model.

[gsgs]

search over all pairs of sequences and all (double) breakpoints whether
the relative number of differences within the breakpoints differs from
the relative number of differences outside the part enclosed by the breakpoints.

No triples needed, just pairs.
55 million pairs, not 192 billion triplets (how do they get 7 billion ?)

took me and Frenchie 2 days per segment of 5000 sequences for single breakpoint
in 50 nucletide intervals.
(not yet speed optimized)

quite some recombinations in H5N1,H9N2 few in H3N2,H1N1.
several obvious sequence-errors

[HenryN]

Originally Posted by Malcolm

We need a translation of the post above into easily understood language.

It would be interesting to have a discussion of the paper's stated rarity of recombination as opposed to the reality of recombination's daily occurrence.

Actually, the paper says it is hard to find recombination that the author's like and can verify with phyogenetic trees (>100 nt). The paper acknowledges smaller regions of recombination and ignores the possibilty that the smaller regions became smaller because of additional recombination. In addition, the paper limits analysis to a human dataset. Thus, recombination with swine or birds is not analyzed. Similarly, the fact that some obvious examples are missing suggests that the human dataset was also limited, or the program just misses the obvious (like the 2002 South Korean HA sequences which have human sequences in circulation a decade earlier).

[HenryN]

Originally Posted by niman

Actually, the paper says it is hard to find recombination that the author's like and can verify with phyogenetic trees (>100 nt). The paper acknowledges smaller regions of recombination and ignores the possibilty that the smaller regions became smaller because of additional recombination. In addition, the paper limits analysis to a human dataset. Thus, recombination with swine or birds is not analyzed. Similarly, the fact that some obvious examples are missing suggests that the human dataset was also limited, or the program just misses the obvious (like the 2002 South Korean HA sequences which have human sequences in circulation a decade earlier).

The analysis only looks at full sequences, so sequences like the south Korean HA sequences with obvious recombination are excluded.

[HenryN]

Originally Posted by Malcolm

We need a translation of the post above into easily understood language.

It would be interesting to have a discussion of the paper's stated rarity of recombination as opposed to the reality of recombination's daily occurrence.

I am still going through the problems withn this paper, but here are a few general considerations.

First and foremost is the fact that most scientists shy away from publishing negative data because they know that negatives can be created by MANY factors that can creep into the experimental design that lead to false negatives, and this paper is a case in point.

The paper has put a LARGE number of restrictions on the data, which takes a VERY biased database and biases it further.

The limitations include a requirement for a full human sequence for inclusion in the database. This elimnates many sequences from China, Korea, and southeast Asia, where much of the diversity and recombination originates. Similarly, birds and pigs are excluded, which represent signifiant influenza genetic reservoirs. Data analysis imposes further limitations. Short stretches of recombination are not detailed because they do not lend themselves to phylogenetic analysis.

As a result, the analysis fails to find some glaring examples in human influenza, and fails to address glaring examples in swine (other than a quick handwave to suggests positives examples are lab error due to contamination).

Some of the above can be seen in the glaring South Korean H3N2 isolates from 2002. The HA sequences are around 1650 base pairs in length, but it looks like all HA sequences less than 1700 were excluded from this analysis.

The South Korean sequences switch from a contemporary 2002 H3 sequence at about position 575 and through approximately position 1000 switch over to a human H3 sequence from a decade earlier. However, most of the sequneces from a decade earlier are only about 1000 base pairs, so they would also be exlcuded from the database used in the paper.

Consequently, the analysis would miss the recombinants (there are six) and the parents (many, but only 1000 BP).

This series alone would invalidate the major conclusion of the paper. The same general recombination was OBVIOUS in six isolates, which formed two subgroups. The smaller subgroups eliminates the possibility of a 1990 H3 contaminant creating a recombinant during amplification. Moreover, the sequences show that the earlier sequences can be maintained for a decade, which was the same type of result seen in the swine paper in Nature Precedings, which is referenced by this paper, but charaterized as "controversial" (because it has real data conclusively demonstrating influenza homologous recombination in multiple genes in multiple swine isolates).

[gsgs]

Originally Posted by niman

The analysis only looks at full sequences, so sequences like the south Korean HA sequences with obvious recombination are excluded.

they should still have found it.
E.g.:

se7a43.10
1112 : >A/Denmark/18-2/03(H3N2)
1118 : >A/Cheonnam/323/02(H3N2)
--------------------------------------------------------....
............................................................
............................................................
............................................................
............................................................
............................o...............................
............................................................
............................................................
............................................................
............................................................
..............................oo.o........o............o...o
.......o.........................................o.........o
.........o...oo....o.......................................o
............................................................
......................o...................oo................
....o..........................o..o.........................
.......o..........................o.......................o.
...............................o...........o................
............................................................
............................................................
.........................o..................................
............................................................
............................................................
............................................................
...................................................o........
............................................................
............................................................
............................................................
.............................--------------

Full sequences.
I think the problem is that they require triples, both parents

[HenryN]

Originally Posted by gsgs

they should still have found it.
E.g.:

se7a43.10
1112 : >A/Denmark/18-2/03(H3N2)
1118 : >A/Cheonnam/323/02(H3N2)
--------------------------------------------------------....
............................................................
............................................................
............................................................
............................................................
............................o...............................
............................................................
............................................................
............................................................
............................................................
..............................oo.o........o............o...o
.......o.........................................o.........o
.........o...oo....o.......................................o
............................................................
......................o...................oo................
....o..........................o..o.........................
.......o..........................o.......................o.
...............................o...........o................
............................................................
............................................................
.........................o..................................
............................................................
............................................................
............................................................
...................................................o........
............................................................
............................................................
............................................................
.............................--------------

Full sequences.
I think the problem is that they require triples, both parents

No, the Korean sequence is about 1650 BP, so it was excluded. The other parent is from a decade earlier, like A/Seoul/45/91. It will match the 0's in the above figure, but it just goes to position 1000.

Your figure shows obvious diversion in the middle of the gene, but the sequences from the early 90's show where the middle sequence originated.

[gsgs]

ahh yes, genbank gives them as partial.
Although 1650 is pretty long for HA

[HenryN]

For those not used to looking at sequences, GSGS posted a comparison of two HA sequences from 2002/2003. A dash means the sequences are the same at the represented position and the 0's are positions that differ. Thus, the clustering of the 0's alone pretty much eliminates the basic tenet of influenza genetics, which was repeated multiple time in the paper - genetic drift is due to copy errors. Thus, the basic tenet would hold that the polymerase got very stupid in the middle of the gene and made a series of errors, but was generally error free at the beginning and end of the gene.

However, the HA sequences from the early 90's show that the 0's are not due to a stupid polymerase, because the 0's match the sequence of the 1991 isolate. Thus, the 0's were created by swapping the middle of the 1991 sequence for the middle of the 2002 sequence to create a recombinant withe the sequence 2002-1991-2002.

The paper maintains that this rarely or never happens in human influenza, yet the example posted and five other isolates from various locations in South Korea in 2002 show that it does happen, and includes slight variations on the theme.

The six South Korean human HA recombinants (from four distinct locations) are:

A/Cheonnam/323/2002(H3N2)
A/Cheonnam/338/2002(H3N2)
A/Cheonnam/340/2002(H3N2)
A/Kyongbuk/320/2002(H3N2)
A/Daejeon/258/2002(H3N2)
A/Incheon/260/2002(H3N2)

[gsgs]

the paper's method considering triples is too slow, so they unreasonably
limit their dataset to human H1N1 and human H3N2, but most genbank-
recombinations are in H5N1 and H9N2 and in swine or birds.
They apparantly also only find recombinations where all 2 parents are available,
but you can already have strong evidence of recombination when you have only one parent
and the recombinant.They also exclude partial sequences.
So they miss e.g. the pairs

HA : A/TW/4845/99(H1N1) , A/HK/1131/98(H1N1)
NA : A/Ft.Monmouth/1/47(H1N1) , A/Rhodes/47(H1N1)
PB2: A/HK/498/97(H3N2) , A/Albany/1/76(H3N2)
HA : A/Daejeon/258/02(H3N2) , A/Habana/26/05(H3N2)
NP : A/HK/498/97(H3N2) , A/NY/136/02(H3N2)

Their conclusion, that there are only few recombinations in H1N1 and H3N2
(e.g. as compared with reassortments or as compared with H5N1) however looks correct.
I see no evidence for increased frequency of small recombinations.
We should see increased frequency of nearby differences then, which is not observed.
(I had tested this earlier)

The work should be redone with a larger dataset and by looking at pairs
insteat of triples.Also some statistics to test for frequency of small
recombinations. And compare with other viruses or random data.

Another idea is to compare flu-databases with databases of other viruses (which ?),
where recombination is known to be frequent.

[HenryN]

Originally Posted by gsgs

the paper's method considering triples is too slow, so they unreasonably
limit their dataset to human H1N1 and human H3N2, but most genbank-
recombinations are in H5N1 and H9N2 and in swine or birds.
They apparantly also only find recombinations where all 2 parents are available,
but you can already have strong evidence of recombination when you have only one parent
and the recombinant.They also exclude partial sequences.
So they miss e.g. the pairs

HA : A/TW/4845/99(H1N1) , A/HK/1131/98(H1N1)
NA : A/Ft.Monmouth/1/47(H1N1) , A/Rhodes/47(H1N1)
PB2: A/HK/498/97(H3N2) , A/Albany/1/76(H3N2)
HA : A/Daejeon/258/02(H3N2) , A/Habana/26/05(H3N2)
NP : A/HK/498/97(H3N2) , A/NY/136/02(H3N2)

Their conclusion, that there are only few recombinations in H1N1 and H3N2
(e.g. as compared with reassortments or as compared with H5N1) however looks correct.
I see no evidence for increased frequency of small recombinations.
We should see increased frequency of nearby differences then, which is not observed.
(I had tested this earlier)

The work should be redone with a larger dataset and by looking at pairs
insteat of triples.Also some statistics to test for frequency of small
recombinations. And compare with other viruses or random data.

Another idea is to compare flu-databases with databases of other viruses (which ?),
where recombination is known to be frequent.

Their analysis of full human sequences pretty much limits the bulk of sequences to a few locations (US and Australia) and most co-infections will involve closely related sequences and the recombination will look like point mutations. They would call the recombination in Egypt in H5N1 point mutations also, which won't explain the sudden appearance of a silent change on multiple genetic backgrounds at the same time, as was seen for G743A (and the same mechanism is in play for Tamiflu resistance -H274Y in H1N1).

Once homologous recombination is acknowledged, the slope gets VERY slippery, becasue recombination would be most common between closely related sequences (due to frequency of co-infections and ability to switch templates), which is why single nucleotide polymorphisms are really due to homologous recombination.

[Oracle]

That would mean looking at positive, rather than negative-sense ss-RNA viruses.

The stellar case would be the Coronaviridae. Recombination has been shown to occur between coronaviruses for many years.

Caution is advised, however, in comparison of these viruses, as coronavirus has a large, complex genome and it's recombination is driven by a very different mechanism than the one demonstrated by your analysis, GS. Leader sequence misalignment is not an apparent issue in influenza viral replication.

I think you should stay on topic to avoid confusion.

[HenryN]

Commentary

http://www.recombinomics.com/News/03...mbination.html

[gsgs]

Commentary

Recombination Analysis In Human Influenza

Recombinomics Commentary 18:48
March 26, 2008

Using an exhaustive search and a nonparametric test for mosaic structure, we identified 315 sequences (~2%) in five different RNA segments that, after a multiple comparisons correction, had statistically significant mosaic signals compatible with homologous recombination.

More controversial, however, is the occurrence of homologous recombination in influenza viruses, most likely involving copy-choice (template-switching) replication of RNA molecules that co-infect a single cell. Although bioinformatic evidence for homologous recombination has been suggested (13, 19), these results remain unsubstantiated, with extensive lineage-specific rate variation a likely source of a false-positive signal for at least some putative recombination events (24, 31).

The above comments from the upcoming paper, “Homologous Recombination is Very Rare or Absent in Human Influenza A Virus”, describe the detection of small stretches of genetic information consistent with homologous recombination. The comments also note the presence of much longer regions of recombination in Canadian swine, although the possibility of differential evolution within a gene is cited as an explanation for the differential patterns of polymorphisms.

However, the differential evolution was discounted in the swine paper, because much of the divergence exactly matched early swine isolates and the crossover points varied, signaling independent recombination events.

The paper on the human influenza sequences however, failed to find longer stretches of recombination. Only two examples were found, and the authors suggest those two examples may have been due to contamination and the recombinant sequences may have been generated during amplification.

However, the detection of longer examples of recombination was limited by restrictions on the dataset being tested. Only full human sequences were used, which excluded sequences with clear examples of recombination. Moreover, the requirement of full sequences eliminated one set of parental sequences for the recombinants

The clear recombination involved a series of six HA sequences from Korea in 2002. The first and last third of the gene match contemporary H3N2 sequences, but the middle third of the genes match HA from sequences from isolates collected a decade earlier. The earlier origin, and the small differences in the recombined region of the six isolates reduced the likelihood that the recombined region was due to laboratory contamination.

In addition, the analysis did not include swine or avian sequences, which are frequent donor sequences for seasonal flu. Since the study focused on the recombinant as well as both parental sequences, the exclusion of swine and avian sequences would further reduce the number of recombinants identified using the criteria in the paper.

Similarly, the paper did not analyze recombination in avian or swine influenza.

[gsgs]

they don't talk much about those short recombination signals,
how strong they were or if there were more than statistically
expected. (they should have told us)

There is quite some diversity in human flu and the flu travels
around the globe, so the concentration on US and NZ
sequences is not so severe.
And even with the Korean sequences, there is still much
less recombination in human flu than in swine or birds.
Why ?
Is the same true for reassortment ?

[HenryN]

Originally Posted by gsgs

they don't talk much about those short recombination signals,
how strong they were or if there were more than statistically
expected. (they should have told us)

There is quite some diversity in human flu and the flu travels
around the globe, so the concentration on US and NZ
sequences is not so severe.
And even with the Korean sequences, there is still much
less recombination in human flu than in swine or birds.
Why ?
Is the same true for reassortment ?

I posted the P values. For NA in H3N2 there were 240 examples. The p value was 1.2 X 10 to the -10 (the likelihood that the distribution was due to chance was 10 billion to 1).

[HenryN]

Originally Posted by gsgs

they don't talk much about those short recombination signals,
how strong they were or if there were more than statistically
expected. (they should have told us)

There is quite some diversity in human flu and the flu travels
around the globe, so the concentration on US and NZ
sequences is not so severe.
And even with the Korean sequences, there is still much
less recombination in human flu than in swine or birds.
Why ?
Is the same true for reassortment ?

H3N2 recombination frequencies and p values

http://www.flutrackers.com/forum/sho...6&postcount=26

[HenryN]

Originally Posted by gsgs

they don't talk much about those short recombination signals,
how strong they were or if there were more than statistically
expected. (they should have told us)

There is quite some diversity in human flu and the flu travels
around the globe, so the concentration on US and NZ
sequences is not so severe.
And even with the Korean sequences, there is still much
less recombination in human flu than in swine or birds.
Why ?
Is the same true for reassortment ?

I would put the recombination frequency much higher than reported. Almost all polymorphisms are in the database with regions of identity to allow for recombination.

That is what the travel logs are. They take a given polymorphisms and show its distribution in the database.

The name of the game is recombination, regardless of what it is called (mosaics or short regions, etc).

This isn't Kansas. Influenza knows how to rapidly evolve. When the recombiantion is between closely related sequences, the acquistions look like point mutations (or single nucleotide polymorphisms).

Most evolution is recombination and it is well controled, which is why I call it "elegant evolution".

[HenryN]

Originally Posted by gsgs

they don't talk much about those short recombination signals,
how strong they were or if there were more than statistically
expected. (they should have told us)

There is quite some diversity in human flu and the flu travels
around the globe, so the concentration on US and NZ
sequences is not so severe.
And even with the Korean sequences, there is still much
less recombination in human flu than in swine or birds.
Why ?
Is the same true for reassortment ?

http://precedings.nature.com/documents/385/version/1

[gsgs]

I found this:
http://www.3seq.com/sample-output.html
http://www.rff.org/~boni/3seq_manual.pdf
and sent email to M.Boni


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

http://www.rff.org/~boni/boni_etal_genetics_2007.pdf
http://darwin.uvigo.es/people/dposada.html

[HenryN]

Commentary

http://www.recombinomics.com/News/03...ion_Short.html

[gsgs]

the two-breakpoint-recombinations, are they supposed
to happen

during one and the same replication

or

by two consecutive recombinations in different cells ?

[HenryN]

Originally Posted by gsgs

the two-breakpoint-recombinations, are they supposed
to happen

during one and the same replication

or

by two consecutive recombinations in different cells ?

Either is possible. In the same cell, if there is one recombination event, there really is no reason why there cannot be a second event. However, co-infections are common, so a second event can involve another host as well as another donor sequence.

When influenza is isolated, it comes with a history, which can involve multiple recombination events with multiple donor sequences.

Look at PB2 in the Canadian swine

http://www.recombinomics.com/phylo/C...Swine_PB2.html

The North Carolina sequence is at position 755-1594 in three isolates (11112, 57561, 56626)

In remains intact in 11112, but in 57561 and 56626 the Tennessee sequence is nested in the middle. Moreover, in 56626 it also has the 53518 sequence which has the same breakpoint.

Thus, the 2003/2004 Canadian swine isolates reflect a series on well defined (because they exactly match earlier isolates) recombination events.

[gsgs]

how many recombination events ?
how many swine were involved ?
how many different (original,un-recombined) viruses were involved ?
(minimum)


since long recombined sequences are pretty rare we won't expect many
aaa+bbb+ccc ---> abc
(except maybe in Canadian swine,who are special)

[HenryN]

Originally Posted by gsgs

how many recombination events ?
how many swine were involved ?
how many different (original,un-recombined) viruses were involved ?
(minimum)


since long recombined sequences are pretty rare we won't expect many
aaa+bbb+ccc ---> abc
(except maybe in Canadian swine,who are special)

Yes, the Candain swine represnt slow motion recombination. It is infrequent enough, so the long stretches remain, and are easily identified and confirmed.

Most of the time however, recombination is between closely related sequences, so the new acquistions just look like point mutations. However, when they are traced, the parental sequences can be identified.

That is why the travel logs form clear patterns and representing distribution routes. That is also why the same change appears on multiple backgrounds, like G743A on H5N1 in Egypt, Russia, Kuwait, Ghana and Nigeria at the same time.

[camster]

what is the proposed method of recombination for influenza viruses?

[HenryN]

Commentary

http://www.recombinomics.com/News/03...ine_Avian.html