There were a number of questions on Part 4 of this thread that I’d like to address with some sequence examples.
Consider 3 sequences:
1. TATAGT
2. TACAGT
3. TATAGC
Let’s say each of these is found in a different H5N1 virus. How to interpret them?
If you believe that mutation is extremely rare and recombination very common (as in the Theory of Recombinomics) one might assume sequence 1 is a recombinant with sequences 2 and 3 as the parental strains. The TAT came from sequence 3 and the AGT came from sequence 2 after co-infection of a cell with these two strains.
If, on the other hand, you believe mutations are common and recombination uncommon (conventional science) your interpretation would be very different. Sequence 1 could be a parental strain that infected two ducks, A and B. Once in Duck A a mutation occured in the third postion, T to C. In duck B, a mutation occured in the sixth position, T to C.
The point here is that the same data can be logically interpreted in two different ways. The phylogenetic trees imply very different things to Dr. Niman than to conventional scientists because his assumptions regarding mutation and recombination are so different.
One further point. In vitro studies have been cited on previous threads demonstrating that recombination is rare and mutation is common, supporting conventional science. It has been suggested that the virus may behave differently in animals than in cell culture. No mechanism has been proposed for this or any experimental data cited. Nonetheless, I do also think that in vivo veritas. However, in vivo studies *have* been done. Here’s one I have cited several times:
L and H2 mutants of the A/NJ/11/76 H1N1 strain of swine influenza virus differ by having either a lysine or a glutamic acid at position 153 of the hemagglutinin glycoprotein of the virus. In two separate experiments, experimental infection of swine with various doses of the H2 mutant resulted in the emergence in 11 of 20 animals of virus with the L phenotype. All evidence indicates that the H2----L mutation, selection, and evolution to predominance occurred within the 7-day span of individual infections. L and H2 mutations appear to act as alleles in the adaptation of virus, respectively, to natural and laboratory hosts. Although the gradual evolution of mutants during sequential infections is commonplace, the present recognition of rapid and predictable evolution of mutants of increased replication efficiency and specific phenotype in the natural host, to our knowledge, is unprecedented.
This paper was published in 1988, but it is very relevant to the current discussion. It demonstrates that mutation can occur at a very high rate in animals. Further, I must say I found reading this paper and thinking about what has happened in the Karo cluster quite sobering.
Non-random evolution
Much has been made of the non-random nature of the polymorphisms observed in various strains of H5N1. The argument has been made that if mutations are random then non-random polymorphisms must imply some other mechanism, namely recombination.
Well, not really. What’s missing from this idea is the role of selection in evolution. Once a virus mutates, it becomes subject to selection. So, although mutations are random (mostly), selection is definitely not random and will lead to non-random distribution of polymorphisms.
Consider again the 3 sequences from the example above:
1. TATAGT
2. TACAGT
3. TATAGC
These represent non-random sequences. Why? Because the genetic code is degenerate, that is, the same amino acid can be code for by different sequences of nucleotides, the three different sequences above code for the same amino acids: Tyrosine (TAT and TAC) and Serine (AGT and AGC). Thus, at the protein level, where selection acts (mostly), the three sequences are identical. Thus, these mutations can occur without any effect on viral fitness. However, if a mutation in sequence 1 changed the second nucleotide from an A to T, a different amino acid would be coded for (phenylalanine). Statistically, changes in the first two postions of a codon, the three nucleotides that code for one amino acid, are more like to result in a change in the amino acid coded for than a change in the third position. Unsurprisingly, more polymorphisms are observed in the third postion of codons than in the first 2 postions. This is evidence of negative (purifying) selection. Although mutations may occur as often in positions 1 or 2, they are less likely to be observed in viral sequences because these changes result in changes at the protein level. Usually such changes are negative, ie, the mutant protein doesn’t work as well so these mutants quickly lose out in the struggle for survival. Therefore, the pattern of polymorphisms observed will be non-random. This is why it is difficult to calculate mutation rates by observing what the virus does in the wild. You are looking at the results of selection as well as the results of mutation. This is why the in vitro studies and the carefully designed in vivo study cited above are so important.
the Kilbourne et.al study says nothing about recombination.
There were no dual infections.
On the other hand there were some studies (e.g. Palese et.al,1992)
which seem to show that recombination is quite frequent,
if there is double-infection. Now you can argue, how likely double
infection is. Reassortment requires it too.
Those synonymous mutations which don’t change the amino-acid are
not random either. There are papers and the available sequences
clearly show this too. I don’t know, whether there are explanations
or theories to explain this.
Also the mutation-rate itself could be subject of mutation and selection…
from part 4 of this thread, re-posting for ease of discussion:
Name,
“Could statistical analysis of past studies shed some light on the argument over whether polymorphisms in the wild are aquired more often by random mutation or recombination, as someone alluded to above? What is the statistical probability of random mutation producing the observed polymorphisms that Niman cites as evidence of recombination?”
I don’t know if this answers your question exactly but the best review I have come across is still this paper originally posted by Theresa42 Phylogenetic analysis reveals a low rate of homologous recombination in negative-sense RNA viruses, Chare et al, Oxford UK
The whole paper is very instructive and I would recommend it to everyone interested in the subject.
The authors acknowledged “the potential importance of recombination in RNA virus evolution” and used as sample “all those negative-sense RNA viruses for which population samples are available currently in GenBank; this comprised 35 viruses from six families and totalled 2154 individual gene sequences.”
This was a very large sample, in fact, as large as you’re ever going to get.
They analyzed the sequence alignments in 3 different ways:
1) Phylogenetic incongruence: Basically they looked for portions where the sequence alignments caused changes in phylogenetic position, then they determined if this was more than what could happen by chance alone and also determined the most likely breakpoints using a maximum likelihood (ML, Holmes et al, 1999) method. Then they estimated the ML (maximum likelihood) phylogenetic trees of the sequences on either side of the breakpoints and analyzed those (for ‘bootstrap support’) to see if they support the notion that this was indeed a location of phylogenetic incongruence. “Strong evidence for phylogenetic incongruence, and hence recombination, is deemed to be present if the conflicting tree positions are each supported by >75% of bootstrap replicates.”
2) Sawyer’s runs test. In this test, a search is made for unusually long fragments within an alignment over which a pair of sequences are identical, or nearly identical, even though these sequences do not share common ancestry. Whether these runs of similarity are longer than expected by chance is assessed using randomly permuted data sets derived from the same data. ie comparison with other regions of the same set of sequences.
3) Informative sites test. This test looks for sequence variation or polymorphism. (Could this be what Niman refers to frequently?) It detects recombination by distinguishing the ‘apparent rate heterogeneity’ among nucleotide sites caused by recombination from the ‘real rate heterogeneity’ among sites caused by mutation. Such a distinction is possible because mutation and recombination affect the pattern of variability at polymorphic sites in different ways, The values of this variability q computed for the real data are then compared to values obtained by simulating 200 data sets along a ML tree assuming no recombination. This test is useful for detecting recombination in more closely related sequences which may not show up in the other 2 tests.
The first 2 methods gave similar results: they found evidence of recombination in some strains of 5 of the 35 viruses: the G2 gene from Hantaan virus, the F gene from Mumps virus, and the HNgene from Mumps virus, the HN gene from Newcastle disease virus, and the N gene from Newcastle disease virus.
The informative site test, however, showed 10 other viruses in addition to these 5 (ie total 15) as showing evidence of recombination, including one for the HA gene of influenza A H1N1/South Carolina/1918.
To explain this discrepancy between the informative sites test and the other two tests, the authors suggested 2 possibilities. First, it could be that recombination has occurred but “that the effects on tree topology are so slight that they have not resulted in clear phylogenetic incongruence.”
Alternatively, it is possible that this test produces frequent false-positives because of positive natural selection,eg in immune escape. This positive selection “can cause unrelated strains to show similar sequences, but with changes that are too localized to show up as phylogenetic incongruence.” The fact that there are more surface antigens that show positive only on informative sites test may support this explanation, as these antigens are under strong selection pressure.
In the case of the HA gene from 1918 H1N1, suggestions that this was a result of recombination (Gibb 2001) had been previously refuted by Worobey 2002 as due to an error of ‘midpoint rooting’ in constructing the phylogenetic tree (ie starting with human and swine strains instead of the parental avian lineage) which masked the uneven rate of evolution of HA1 and HA2 in the human lineage as compared to the swine, so that Gibb’s alignment showed an apparent (false) recombination event using the informative sites test but which when properly aligned with rooting from avian lineage, showed the supposed recombinant regions to be ‘almost identical’.
The authors of this current study re-examined Gibb’s alignment and showed that even with this alignment it did not fulfill the criteria for the first 2 tests discussed here ie phylogenetic incongruence and Sawyer’s runs test. Thus they concluded that “At present, therefore, the case for RNA recombination in Influenza A virus remains unproven, with extensive rate variation producing a similar phylogenetic signal in this case.”
anon at 08:51 “On the other hand there were some studies (e.g. Palese et.al,1992) which seem to show that recombination is quite frequent, if there is double-infection.”
Can you give us a link or at least the title of the Palese paper that you refer to? The only ones I found had to do with experimentally generated viruses, such as this study Transfection-mediated recombination of influenza A virus.
Monotreme-
Your H2-L paper seems to support the idea of attractant or rules-based recombination which Henry has frequently noted can be rapid.
I don’t really think that the rate of genetic acquisition is the foundation of Niman’s recombination. Progenitor A and Progenitor B interact to form a daughter. Daughter and progenitors are sequenced with matching strings. Following the examples that Niman has given takes you several generations further to strings that still match and are moving between daughter strains. Too many consistent changes to be random. Too many examples of multiple generation passes to be random.
Uh oh.
NS1 – at 16:07
Monotreme- Your H2-L paper seems to support the idea of attractant or rules-based recombination which Henry has frequently noted can be rapid.
The paper I cite does *not* support any sort of recombination at all. Quite the opposite in fact. The paper is freely available. It does not require a subscription. Please read the Materials and Methods section carefully. Animals were first tested to make sure that were not infected with influenza at the start of the experiment. They were then infected a single type of virus. They were housed under controlled conditions, ie, they could not have become infected from another source. Thus, since co-infections were not possible, both reassortment and recombination were removed as sources of variation. This leaves in vivo mutation as the only source for the variation.
This work seems to have been carefully done, as near as I can tell. This paper alone would appear to demolish a central tenet of Recombinomics, that mutations occur very rarely in vivo. When combined with the many other papers, in fact all the other papers on the subject that I could find, the fact that influenza mutates rapidly seems inescapable.
Note, I am not denying that recombination occurs or that it is potentially very important. What I am denying is that mutations are rare.
anonymous – at 08:51
the Kilbourne et.al study says nothing about recombination. There were no dual infections.
Agree completely. See my post at 21:27.
On the other hand there were some studies (e.g. Palese et.al,1992) which seem to show that recombination is quite frequent, if there is double-infection
Which Palese publication are you referring to? This one, published in 1992, concluded that:
It should be noted that the frequency of these recombination events must be low since only a few recombinants have been identified during our studies and since we also suggest that the frequency of recombination is increased by our artificial transfection system.
Not only is this not an endorsement of Recombinomics, it is strong evidence against it. One of the central pillars of Recombinomics is that recombination is frequent. I have read many papers written by Dr. Palese. He clearly thinks that mutations are frequent and recombination is rare - the exact opposite of the central tenets of Recombinomics. If you believe otherwise, please provide quotes and links.
Also the mutation-rate itself could be subject of mutation and selection…
Yes, that’s possible. There has been some speculation regarding so-called mutator strains of virus.
“Nature has now obtained more detail on the genetic changes, which suggest that although the WHO statement was not incorrect, plenty more could have been said. Viruses from five of the cases had between one and four mutations each compared with the sequence shared by most of the strains. In the case of the father who is thought to have caught the virus from his son — a second-generation spread — there were twenty-one mutations across seven of the eight flu genes. This suggests that the virus was evolving rapidly as it spread from person to person.” (Revere at Effect Measure July 13/2006)
I am just one opinion of many here at flu wiki…but this statement seems to confirm the hypothesis that Monotreme has been patiently explaining for some time.
One of ways that the WHO could regain my trust (a………………. long-shot at this point) would be to convince Monotreme to give up his alias, and ‘get to work saving the world’.
Bull-**** does baffle brains up to the point when you realilze that a pandemic is really going to happen…that time was about three months ago.
By the way, no wonder the Tamiflu was sent, other measures were taken and the tone got very serious, very fast!!
Monotreme – at 21:27
I should have placed more emphasis on rules-based and less on the word recombination.
Of course, traditional recombination does not occur without additional genetic material. I’m certain that we cannot test successfully for most of the applicable donors of genetic material (small concentration bacteria, small substrings via plasmids, et al). So we cannot fully exclude the possibility of non-influenza donors in this test.
Regardless of that I’ll go with the assumption that the individual hosts had exactly zero genetic material donors.
Given: One influenza strain where a SNP occurs in 7 days of reproduction with no genetic donor for that SNP, no transposon, et al. Random? Not a chance.
Please let me try to restate my feelings more precisely . . .
Rules-based genetic acquisition seems to be very clearly at play in the paper you cited. One for one polymorphisms, almost predicatable, seem to revolve around a rule.
We are still not looking at a single paper concerning wild viral genetic acquisition. My denotation and connotation is strongly bent toward variance of genetic acquisition set in nature.
Rinse, repeat.
After 6 months, we should have some strong signals.
Do we have any viable data?
Tom DVM,
Thanks for the vote of confidence, but I have way more influence on pandemic preparedness as Monotreme than I would under my own name ;-) I have very few credentials in virology and none in public health.
I think the Karo cluster mutations are pretty strong evidence for a high rate of mutations. But I would like to actually see the sequences deposited in GenBank before I come to any firm conclusions.
NS1 – at 00:15
I have no doubt that there is a rule that affects which polymorphisms we observe in nature - it is called Natural Selection.
Recombinomics flatly denies that mutation occurs frequently. The paper I cited contradicts this pretty clearly. As do all other papers on the subject.
As regards your proposed experiment, it is sadly impossible because any tag would diluted shortly after the tagged virus was released due to many rounds of replication in a short amount of time.
The many other approaches all come to the same conclusion. Although any one paper can always be criticised for not excluding a possible source of error, what are the odds that the many, many papers on the incidence of mutations done in different ways are all wrong?
‘treme,
And if you are going to argue with those papers, you’d better show up with some evidence, n’est pas?
Monotreme. You’re specific credentials and the specific credentials of others of my colleagues on flu wiki are of no concern to me. I have been around a lot of ‘know nothings’ with credentials and a well developed God Complex in the past as have probably you as well.
I’m not all that good with the English language but what was the British Navy term when they would grab you in a bar and you would be in the Navy for four years.
If I was in a position of authority at the WHO today, I would be going around the world doing just that to many here and I couldn’t care less with qualifications but rather ‘track record’…as the Math teacher wrote on my brother’s report card…”Watch it Charlie, your brains are showing.”
Tom,
The word is “impressment.”
Melanie – at 00:28
And if you are going to argue with those papers, you’d better show up with some evidence, n’est pas?
Oui! In fact, you better do some experiments of your own and publish them. The greater the claims, the greater the required evidence.
Forget the experiments…let’s hop a plane and go over there and find out where the virus is hiding, straighten out a few god-like authorities into releasing the sequences and then come home…should take Oh probably ten days.
Melanie Thanks!! Sounds suitably British.
Tom,
The same thing happened in the US during both the Revolution and the Civil wars. It ain’t an exclusively British concept.
Melanie. Thanks!! It still sounds suitably British. /:0)
I can say that because I am from the colonies. hehehehehehe(I think that is how Bronco Bill does it)
Here’s some more from the War of 1812, which was an American expansionist war. Google the War of 1812 and get quite the education.
monotreme at 22:15, yes, that Palese paper. I had found it from a link here. I didn’t read that comment about low recombination rate. But “low” is relative. It continues : …nevertheless…contributes. I’d like “low” to be replaced with a number-estimate. Maybe in 10% of double-infections recombination occurs ? Some say that’s low, some say that’s high.
I also noticed some events with recombination in one virus in several genes. This should be very rare if recombination is rare in coinfection cases.
I agree, that point mutations are frequent and I can’t see how most of these could be recombinations. But recombinations are dramatic steps and when they occur, they usually result in more significant changes than point-mutations.
Monotreme – at 00:18 Tom DVM,
“Thanks for the vote of confidence, but I have way more influence on pandemic preparedness as Monotreme than I would under my own name ;-) I have very few credentials in virology and none in public health.”
I agree and I feel that I am in exactly the same position. Not as Monotreme, of course, :-)
anonymous – at 02:15
“But recombinations are dramatic steps and when they occur, they usually result in more significant changes than point-mutations.”
Yes, and the best examples are the non-homologous recombinations (ie acquisition of nucleotide from a different gene segment of its own genome without co-infection) with insertion of nucleotides resulting in the addition of basic amino acid residues in the HA cleavage site, causing the viruses to turn from LPAI to HPAI in one jump.
Recombination Resulting in Virulence Shift in Avian Influenza Outbreak, Chile 2002
Novel Avian Influenza H7N3 Strain Outbreak, British Columbia 2004.
I find these intriguing and worrying as a mechanism that can potentially create a pandemic strain literally overnight. However, these are the only 2 instances, and the fact that they appear to be so similar eg both to do with basic aa in cleavage site, both happening in chickens, both turning LPAI to HPAI, etc may point to some unique circumstances with regards to this particular part of the HA segment, and not a reflection of a general tendency to recombination.
“But “low” is relative. It continues : …nevertheless…contributes. I’d like “low” to be replaced with a number-estimate. Maybe in 10% of double-infections recombination occurs ? Some say that’s low, some say that’s high. I also noticed some events with recombination in one virus in several genes. This should be very rare if recombination is rare in coinfection cases.”
I think the most important point Monotreme is trying to make, and which he has repeated many times, is that recombination can happen, but are far rarer than mutations.
Therefore, IMHO, if we are interested in educating ourselves with regards to pandemic influenza, with the exception of the above specific instances of LPAI to HPAI, our time would be far more profitably engaged in studying mutations, selection, reassortment, host response, possibility of mammalian host, source and mechanism of evolution of 1918 virus, systematic surveillance of whole genomes and matching them to different host species, etc.
if recombination is potentially “not inconsequential,” why should it not be studied as well as the rest, even if it is a relatively rare occurance? Makes no sense to limit study in such a way.
anonymous – at 02:15
I agree, that point mutations are frequent and I can’t see how most of these could be recombinations. But recombinations are dramatic steps and when they occur, they usually result in more significant changes than point-mutations.
Yes. No disagreement there.
if recombination is potentially “not inconsequential,” why should it not be studied as well as the rest, even if it is a relatively rare occurance?
I agree completely.
Makes no sense to limit study in such a way.
As far as I know, there are no limitations on the study of recombination, other than financial. In fact, recombination is being studied as the paper from Palese and the ones cited by anon_22 indicate. You can argue that more studies are in order, and I agree. But that is true of almost all of biomedical science. You should understand that funding for biomedical research is extremely limited given the costs of the necessary experiments and is getting worse.
anonymous,
In my post at 08:11, I should have emphasized the words IMHO our time would be far more profitably engaged.
There are many things that one can do, more things that one would like to do. But the fact is there are only so much resources, each of us have only one lifetime and 24 hours in a day, and there are far too many things that require our attention that each person must make some judgment as to how they want to spend their time and efforts.
What is being discussed here is not whether one should study it, but rather its relative importance in our efforts to understand pandemic flu. Which means that if you are interested in recombination, by all means go ahead and study it.
bump
So it has been decided that recombination is relatively unimportant in pandemic flu? Don’t want to reread all these threads, but I think I recall monotreme and others saying recombination, even if rare, could be very important.
anonymous, everyone has their own opinion on the importance of studying the sequences and how variation occurs. I happen to think this subject is very important and that recombination could be very important. Other opinions may differ.
Pure logic would seem to dictate that recombination of any degree would be important. And even if recombination is a minor factor, it could be the most telling factor in the development of panflus. I say the last only because we seem to know next to nothing about the development of panflus, and especially avian ones like H1N1 (1918).
I don’t want to put words in Monotreme’s mouth but if you didn’t read the whole thread then let me just say that both Monotreme and myself think that recombination may be important if it happens, but that it appears to be very rare, and so the importance may be less because of that.
anon_22: Thanks for the Cliff’s Notes version. : )
You’re welcome, MM.
anon 22--“but that [recombination] appears to be very rare, and so the importance may be less because of that.”
with emphasis on the “may.”
I think frequencey of occurence and importance are not necessarily related. Mutations are frequent and recombination is infrequent, IMO. However, most mutations are not important, except as markers of lineage. Recombination may have caused the 1918 pandemic. Severe pandemics are extremely rare but very important.
Monotreme:
Though it’s almost an aside at this point, what you’ve just presented could be an example of the sort of “historical event” I was talking about in our lengthy “Neutralism versus Selectionism” discussion. It’s not so hard to see how a single micro-event could be very significant despite being very rare. When Niman says that recombination is important, however, part of what he’s saying is that it’s important because it happens frequently.
Racter – at 15:43
what you’ve just presented could be an example of the sort of “historical event” I was talking about in our lengthy “Neutralism versus Selectionism” discussion
I never doubted the historical nature of specific mutations. The issue for me was whether, in a global sense, selection is predictable - which I think it is.
When Niman says that recombination is important, however, part of what he’s saying is that it’s important because it happens frequently.
Yes, and this is the difference between his Theory of Recombinomics and conventional science. Conventional science has no problem with the idea that recombination can be very important. But this is insufficient for Dr. Niman. His Theory insists that it is very common as well. And that mutation is very rare.
If a pattern is clearly observable from the sequences, does it really matter what we call it?
Does it matter what the frequency of random mutation is? Does it matter how “selection” effects random mutation rates?
The primay concern is the outcome and if we can prepare some sort of pro forma based on the patterns?
We stare at the pixels and miss the Picasso.
The sequences do show a pattern of SNPs and longer substrings that recur.
If these are genetic donation, attractant rules magnetism or random mutation, let’s use them to define the pattern.
Ask me sometime about attractant rules magnetism or the persuasiveness of conductors.
Not now, just remind me later, please. Maybe you, Racter.
NS1 involuntarily impersonating a bystander.
Hi everyone. Just thought I would throw my two cents in. I think pandemic type viruses have special capabilities not present in seasonal flus. Which of the three processes provides the unique ability, I’m not sure…
…but I have a sneaky suspicion that we may see all required mutations at once and a sudden or instant transformation to a pandemic virus.
That last jump in the Karo cluster is pretty amazing for a random set of mutations.
But is it significant at p < 0.05 ? Please show your work.
Can’t calculate with significance because we still don’t have the sequences. All data being reported is in question due to potential error.
The count of genetic changes being reported is unusual for a single generation jump.
NS1 – at 03:54
Can’t calculate with significance because we still don’t have the sequences. All data being reported is in question due to potential error.
This is my position as well. I’d hate to overinterpret sequences that turn out to be due to errors.
The count of genetic changes being reported is unusual for a single generation jump.
I’m not so sure about this. I have been looking at other sequences from human isolates and there are some very strange patterns. Need more time to organise the data before I post on this.
Monotreme – at 09:44 Don’t forget one of the Karo cluster nurses came down with the flu during the incident. The father may have had contact with her during hospital visits before he got sick himself. But of course, dual infections are supposed to be rare, and recombination even more so.
that won’t explain the 8 mutations in HA
anonymous – at 16:07
Don’t forget one of the Karo cluster nurses came down with the flu during the incident. The father may have had contact with her during hospital visits before he got sick himself. But of course, dual infections are supposed to be rare, and recombination even more so.
Dual infections of H5N1 should be extremely rare given the low incidence of the disease and the seroprevalence studies. Dual infection with H5N1 and H3N2 is quite possible and quite scary to many of us. Dual infection could result in reassortment as well as recombination.
anonymous – at 16:58, I can’t tell if you are arguing with anonymous – at 16:07 or if you are anonymous – at 16:07.
This is why those of us who have handles have are so frustrated by those of you without handles. If you are concerned about cookies, block them with your browser - I do. I type Monotreme in the Author field each and every time I post.
I am cookie-free and I have a handle. Try it, we’d all appreciate it.
Monotreme. With an influenza that appears to me to have a higher mutation rate then seasonal influenzas, would infection each unique infection not become a dual infection or greater, in short order anyway, due to mutations.
Tom DVM – at 18:45
With an influenza that appears to me to have a higher mutation rate then seasonal influenzas, would infection each unique infection not become a dual infection or greater, in short order anyway, due to mutations.
That’s possible. It would depend on how long the infection continued before the virus was cleared.
Monotreme. How long do you think it would take to have an active infection with two strains by mutation?
Tom DVM,
It also depends on the mutation rate and chance. Mutation doesn’t occur with a regular periodicity like a metronome. Also, it’s possible that a mutation in a polymerase gene could trigger a rapid increase in the mutation rate.
So, bottom-line, I can’t answer your question.
Monotreme. It a mutation that favoured the virus occured within the first say twelve hours of an active infection, it seems to me that given replication rates of a slightly favoured virus, you would have a dual infection in short order…
…which means that natural selection could be occuring simultaenously and naturally in every person infected?
NS1 admits that there really isn’t enough evidence to say that the changes are statistically significant, yet calls the changes “amazing” in one post then “unusual” in his next. These mutations could simply be part of the random flow of the virus under selective pressure of the human host.
The recent cluster is interesting mostly for the fact that we have samples from all but the first case. Has any other sequence of flu infections been so carefully watched ?
a’Akova – at 19:56
Has any other sequence of flu infections been so carefully watched ?
You are forcing my hand before I am ready! I have been quietly working on this page in an attempt to address that very question.
My preliminary answer is no and yes. No because most clusters have not been explicitly acknowledged. Yes, because if you dig hard enough, and trust me you really have to dig, there is sequence information from other clusters.
More on this later, but I’ve still got a ways to go before all the information is ready.
Thank you Monotreme, for your hard work and contribution and everything. We all appreciate it. :-)
anon_22, you’re welcome :-)
Monotreme,
Great work.
I adjusted some formatting and added a few commas to align dates for readability.
No content changes made.
You make it go Montreme…We are not worthy!
mono at 18:44, no matter who I were, the argument is exactly the same. A dual infection of H5N1 and H1N1 or H3N2 won’t create mutations in the HA-gene by reassortment or homologuous recombination. Too different is H5 from H1 and H3.
Thanks NS1 and Dude.
anonymous – at 02:36
A dual infection of H5N1 and H1N1 or H3N2 won’t create mutations in the HA-gene by reassortment or homologuous recombination.
That’s true.
Monotreme – at 08:05
This one really needs a bit of explanation. I take it that only the HA or NA segments of the H5N1 will remain intact, but that the other segments can reassort with the H1N1 or H3N2. Correct?
From the thread entitled “Sequences from Maharashtra India” posted by Theresa42 at 02:49, and the article she quotes from “Current Science, Vol.91, No.1, 10 July 2006″, Phylogenetic analysis revealed genetic similarity of the H5N1 avian influenza viruses isolated from HPAI outbreaks in chickens in Maharashtra, India with those isolated from swan in Italy and Iran in 2006 and From the study it is clear that the H5N1 virus that caused outbreak in chickens in Jalgaon did not originate from the neighbouring district of Nandurbar where the first outbreak occurred; instead both the HPAI outbreaks in Maharashtra were due to two different populations of the virus introduced at two different times.
Would it be a far-fetched to imagine it possible for some unfortunate animal, whether avian or mammal, to catch both different strains of H5N1 in this State, and therefore fulfilling the dual infection discussed previously? Just in theory?
birdman – at 08:12
I take it that only the HA or NA segments of the H5N1 will remain intact, but that the other segments can reassort with the H1N1 or H3N2.
AFAIK, any of the genes can reassort. I think the point is that reassortment doesn’t change sequences within genes, it just results in trading of entire genomic segments.
FrenchieGirl – at 08:40
Would it be a far-fetched to imagine it possible for some unfortunate animal, whether avian or mammal, to catch both different strains of H5N1 in this State, and therefore fulfilling the dual infection discussed previously?
I haven’t had a chance to look at the whole paper, yet, so any comments from me would be premature.
Monotreme – at 09:05
Thanks a bunch. I think I have it now.
The subject father from the Karo cluster indicated an H5N1 HA with an unusually high number of nucleotid changes. Even for a virus whose HA mutates frequently, the father’s changes seem high. So, reassortment seems to be a possible culprit, but only if the father was dually infected with H5 viruses. So, if the nurse in the hospital had a seasonal flu bug (say H3N2), and infected the father, that still would not explain the high HA changes, since the H3 from the nurse cannot reassort with an H5 in the father.
Question remains: Where and how did the father pick up two H5N1 viruses? Or, can H5 really mutate that often and within such a very short period of time? Heaven help us, if true.
Clarification. An H3 can reassort with an H5, but the resulting virus would no longer be an H5. Since the father’s HA changes were in an H5, had to have been another H5 if reassortment had taken place.
birdman – at 09:05 and 09:36 - Hence my question at 08:40:
If 2 H5N1 viruses are co-circulating in the same region, can an animal (mammal/avian) catch both? In the last few days, we’ve had report of two different strains circulating in nearby farms in Niger, and today, we read the same in India. And I’ve just now read Dr. Jeremijenko on another thread, about the strange Karo cluster, the cats and the pigs… I’m sure Monotreme will be able to clarify this later if we give him some time to unravel these mysteries.
birdman – at 09:31
Question remains: Where and how did the father pick up two H5N1 viruses? Or, can H5 really mutate that often and within such a very short period of time? Heaven help us, if true.
I am lookiing into this question right now. You might want to look at the sequences in the Dong Thap cluster from VietNam.
FrenchieGirl
I’m sure Monotreme will be able to clarify this later if we give him some time to unravel these mysteries
Thanks. I’m working on it. When I first started looking at the H5N1 sequences closely, I felt like I fell down the rabbit hole. Subsequent events have become curiouser and curiouser.
yes, they can reassort in theory. This has happened in southern China. But in birds, of course. Human infections are still rare.
birdman, yes when you change H5 with H3 or even recombine them in a large area (if they would recombine at all) then you get hundredth of mutations. Some other genes might be quite similar, though. But I guess this is also unlikely.
It was reported that recombination and reassortment of H5N1 happened in southern China but not elsewhere. Once the strains left China, they didn’t mutate so much.
There is also much larger diversity in Chinese viruses than e.g. in Indonesia. And these viruses from India,Europe, Turkey,Iran,Siberia,Africa,Qinghai are all very similar.
“It was reported that recombination and reassortment of H5N1 happened in southern China but not elsewhere. Once the strains left China, they didn’t mutate so much.”
Yes, I have been thinking about that. I wonder whether this is due to some unique host species in which such reassortment is more likely and that species live in limited habitats only in southern China. I am not just thinking of H5N1 but pandemic flu in general, and the theory that that area is the epicenter of panflu viruses.
my estimate: 60% that the pandemic starts in China, 15% Indonesia, 25% elsewhere (provided there is one)
bump
here is my takeaway from this discussion:
Reassortment/Recombination may happen rarely, but when it does, it has the potential to immediately transfer genetic changes which may take many many generations of random mutation and selection to develop. Environment may have a lot to do with the frequency and effectiveness of such reassortment or recombination.
Mutation happens far more frequently, but is less dramatic in the changes that result from each mutation.
Is that the consensus? IS there a consensus?
LMWatBullRun – at 17:59 What you are watching is science, and it is a bit messy under the hood. It’s not really about consensus, because 100% of scientists can occasionally be wrong. What we are trying to get at here is the truth, but it’s like the blind man and the camel: We don’t always agree on termninology, methods, interpretation, etc. If there is anything I have learned in over 30 years as a professional scientist, it’s to let the data speak for itself, and not pay much attention to what the humans are saying. And not to force it, the answer will open like a flower if you let it.
I think what LMWatBullRun is asking is just what most people think. Opinions and educated guesses are not truths, but often quite useful to know just for background understanding IMO.
As wetDirt has mentioned, we must let the data speak.
Not enough eyes have looked at enough sequences with enough different frameworks in mind to come to useful conclusions at the moment.
Study, Study, Study.
Here is the difference between scientists and engineers-
Scientists seek to further define the nature of reality.
Engineers, like me, seek to find acceptable solutions to human problems.
While I am interested in the science, at this point I am seeking acceptable approximations or best guesses that can be used to provide guidance to reduce the impact of the pending pandemic. If the contributors to this thread cannot be persuaded to provide such, I’ll make up my own. <grin>
Engineers have been doing that for a very long time.
birdman – at 09:31
“Even for a virus whose HA mutates frequently, the father’s changes seem high.”
Again, “seems high” has no statistical grounding.
birdman – at 09:36
“An H3 can reassort with an H5, but the resulting virus would no longer be an H5.”
Incorrect. Reassortment does not mean an exchange of H or N genes only. H5N1 x H3N2 (where ‘x’ stands for reassortment) can yield any of H3N1, H3N2, H5N1, or H5N2. There are 8 genome segments in Influenza, of which H and N are two; trades of the other segments are still qualify as reassortment.
LMWatBullRun – at 17:59
Is that the consensus? IS there a consensus?
Yes and yes. To my knowledge, there is only one biomedical scientist in the world who thinks mutations are rare and recombination very frequent.
There are many virologists listening to the data, and all but one agree on the tune.
a’Akova – at 21:08
Thanks for the reply. Yea, I agree, the “seems high” is a repeat of what all the posters have said about these changes to the HA segment. They seem to think that so many changes is unusual for point mutations, leading one to suggest reassortment.
Now, please clarify for me how the father had so many changes to HA, and lets assume that was due to reassortment. Since the father’s sequence is an H5 hemmagluttin, would that not lead you to the conclusion that the reassortment was from another H5? In other words dual infections. If the reassortment was with an H3, and the father’s HA was still an H5, then the reassortment had to be with one or more of the other 9 genes, not the HA.
Correct? Or am I still missing something here?
correct.
When we count 15HAs, 9NAs, how many different “clades” can we identify for the other genes ? Instead of saying H5N1, wouldn’t it be better to say:
PB1:3,PB2:7,PA:1,HA:5,NP:13,NA:1,M:2,NS:1 or such ?
Birdman – at 22:29
“Now, please clarify for me how the father had so many changes to HA, and lets assume that was due to reassortment. Since the father’s sequence is an H5 hemmagluttin, would that not lead you to the conclusion that the reassortment was from another H5? In other words dual infections. If the reassortment was with an H3, and the father’s HA was still an H5, then the reassortment had to be with one or more of the other 9 genes, not the HA.
Correct? Or am I still missing something here?”
I think what you have not taken into account is that phylogenetic analysis of the father’s HA, even with all these mutations, still places it closest to this cluster, and there are no other HA’s remotely similar to this one except for this cluster.
phylogenetic tree Indonesia Peiris
thanks !
We can’t know whether there are other sequecnes, closer to the father
than the father is to the son.
This is only a 2dim planar version.
But if the data is correct, which I doubt, then the
father would be closer to the other Indonesian sequences
than the son is to the other sequences.
This is strange, since it would mean, that H5N1 mutated away
from the other sequences and then in the father “redid” some of
those mutations.
Looks like a case for Monotreme.
the father should be IDB/560H/06
but I only find IDN/59?0H/06 (cluster)
the son is IDN/546bH/06 (cluster)
anon_22 – at 06:02
Thanks for the reply, and I agree with all that you stated. That has been my position all along. If there was reassortment of the HA segment, it had to be from another H5. The only way for it to work is with a dual infection of H5 viruses. That dispels the allusion from another post that there may have been a seasonal flu virus in the mix (H3N2). There may have been, but not as to the HA changes in the father.
The latest sequence (actually it is the only sequence and is delayed almost 12 months) of H5 from wild birds in Canada has a 25 BP stretch of identity with swine H1. This region includes two polymorphisms, which drive another nail into the “random mutation” explanation of genetic drift. The strech of identity can actually be found in WS/33, the first human flu virus (H1N1) isolated (in 1933).
Of late, this strech has been exclusively found in H1 from pigs, including Ontario 11112, the H1N1 that has the major stretches of identity with one 1977 swine isolate for PA and another 1977 swine isolate for PB2 (and this recent isoalte also has a human PB1 sequence, circa 1996).
The tandem polymorphisms create major problems for random mutations, or even species convergence. The sequening was also done by a lab not affiliated with the Wisconsin lab that generate the “evolution challenged” Candian swine which have 1977 PA and PB2 genes (and one has a 1931 gene).
The latest data again shows that recombination is common, and in fact is so common that one isolate can have polymorphisms contributed by several parents (as is becoming increasingly obvious from the new acquisitions in the Qinghai H5N1 strains in Africa and Europe).
These distributions are NOT random and again demonstrate frequent recombination (especially between closely releated sequences, such as the Qingahi H5N1’s).
A couple of things to keep in mind. First, mistakes do happend. I am starting to come to the conclusion that there may be many mistakes in attribution of sequences to patients in GenBank. See this thread for examples. The Dong Thap cluster reminds me a lot of the Karo cluster. Phylogenetic trees require accurate and complete data. Falsely attributed records and missing data can result in misleading trees. Job one is to clean the databases and make them more complete.
Second, 25bp of identity is not enough to convince me of anything. I routinely BLAST sequences of that size when designing PCR primers. Even when I choose primers from within the coding regions of mammalian genes, it is amazing the amount of near identity hits I get from irrelevant organisms and irrelevant genes, by chance. The more sequences that are deposited, the more often this will occur. As a test, try this: Go to the BLAST search page, type in a random sequence of 25 letters in the search box. Use only A, T, C or G’s. Unclick the low complexity box, then click the BLAST button. A new window will appear. Click the Format! box. Wait a few seconds and you’ll see a new window with alignments. If you do this several times, you’ll be amazed by how much identity there will be between the randomly generated string of 25 letters and real genes from real organisms. There are so many sequences, and more being deposited every day, that it is increasingly difficult to find a string of nucleotides that aren’t a pretty good match for something.
Now, if one starts with a 25 nt string from a viral gene, what are the odds that it will, by chance, match another virus? Very high. The coding regions of genes are highly conserved across hundreds of millions of years of evolution because changes in nucleotides that result in changes of protein often result in genetic death. [this is also true of some non-coding areas as these may have regulatory function]. So, even though mutations occur within these genes, the mutants are not observed because the organisms that contain them die or fail to reproduce. The degree of freedom for variations to occur are actually very limited. The wobble position is one such location - changes there may not affect the function of a gene. In a 25 nt coding region “probe” there will be at a most, 8 degrees of freedom, and likely less. So, the chance that a randomly chosen 25 nt sequence will match another sequence is reasonably high. The odds of this go way up if you systematically work your way down a sequence 25 nts at a time looking for chance alignments.
Long stretches of homology do require explanation. See the first paragraph for one possibility.
I can concur with Monotreme, that I have seen many times short stretches of nucleotides (approximately 15–28 nts), that when blasted will pull up matching sequences from widely distant organisms.
Monotreme, You are comparing apples and oranges. The 25 bp probe of about 50,000 flu sequences produces so very specific results. Comparing this to searching a huge database really is pretty irrelevant as is your “errors” arguments.
There were MANY examples of recombination in the Candia swine and trying to say that one labe made 50–100 “mistakes” really has no credibility (nor does the withdrawl H5N1 swine sequences two weeks before your postings).
Let’s get real and address the data.
Martians MIGHT have created the 50 examples of recombination in the Canadian swine. My guess is that the odds on Martians is slightly better than the lab making that many errors.
also when you compare HA in H5 with HA in H1 and only consider aligned (homologuous) matches ?
Dr. Niman, you really don’t address my arguments relating to the limited degrees of freedom available to viral sequences. My impression is that you search for areas of identity between viral sequences 25 nts at a time, and when you find them, call them evidence of recombination. Ask a statistician what he thinks of the validity of this approach. 50,000 sequences in the database is plenty for random matches if you are starting with viral sequences.
The mistakes I am referring to are not sequencing errors. I am saying the wrong attributions are being made. Examples, here.
When you say “Let’s get real and address the data” and start your next sentence with “Martians..” it’s hard to take you seriously.
anonymous,
The 25 bp in H1 and H5 are in the same position.
Monotreme, You need to look at the flu sequences. They have a great deal or order and the matches are really not as common as you guess, even with probes as short as 10 BP.
I gave you very strong examples of IDENTITY between the swine PA and a 1977 isolate. I also gave you examples of IDENTITY between PB2 and another 1977 isolate. These isolates then hop over to regions of IDENTITY with 1998 and 2002 isolates.
Attributing these matches to “lab error” is well into the Martian category and comparing these to some naming error is pretty irrelant.
Labs to make mistakes and errors do creep into peer reviewed publications, but the recombination in these swine is well beyond the “lab error” possibility.
I once examined this and found that there are typically roughly 2^n different subsequences of length n. (out of 4^n possible ones)
Dr. Niman,
As I’ve said many times, I agree with your suggestion that the swine PA and 1977 sequences are evidence of recombination. I do think this needs to be explained by the laboratories that generated the sequences. Your interpretation is a perfect polymerase, mine is some sort of error. Time will tell who is right. In the meantime, I will point out that there aren’t just a few errors in attibuting sequences to strains, there are many. I found apparent errors of attribution in the first few sequences I examined.
As regards recombination, let me repeat, I have no doubt that it occurs. The question is, how often? You provide a few examples of long stretches of homology which I agree are evidence for recombination. So would most conventional scientists. Where Recombinomics diverges from conventional science is in the claim that recombination occurs frequently. Your only evidence of this is small regions of identity. Apparently 25 nucleotides of identity are sufficient to convince you recombination has occurred. This is unconvincing to me. The only way to resolve this is for a statistician to calculate the odds of chance identity with these small seqments.
Monotreme,
Your hand waving just doesn’t cut it. Lets stick to the data. There are many more examples in the Canadian swine. In fact, there are examples in all 8 genes. I just gave you a few that created some major problems with “random mutations”.
You are using an explanation that really makes no sense, but if you are correct, the “mislabeling” can be easily demonstrated. Do you think the “lab error” is in the 7 recent swine isoaltes, or in the 2 earlier 1977 isolates?
If you think the 1977 isolate are “mislabled”, you can take the 1977 sequence and do a BLAST to see if it matches other 1977 sequences or 2004 sequences. If you think the other 1977 sequence is also mislabeled, you can BLAST that sequence. If you think the 1998 sequence is mislabeled, you can BLAST that sequences. If you think the 2002 sequnece is mislabeled, you can BLAST that sequence.
I think taking the time to prove that the “lab error” explanation is utter nonsense is a waste of time and is in the “Martians did it” category.
If I am wrong, why don’t you do the BLASTS and tell the readers of this board (as well as the labs that submitted the mislabeled sequences), that there DATA can’t be right and tell them and the board why.
Martians are upper nonsense. The 1977 virus could have survived in frozen form in the St.Jude’s lab and then accidently released in 2002 ? There were papers about that virus, but no one mentioned this recombination and conservation. Looks like conspiracy…
I emailed to Olsen, and he insists it didn’t escape from their lab, but offered no explanation. Instead asked about me.. whether I were an employee of recombinomics and such.
Your hand waving just doesn’t cut it.
Have I mentioned that “hand waving” has become one of my least favorite expressions? This thread is on part 5 - it’s hummed along quite nicely for quite a whle and been quite civil in the process. Let’s keep it that way. Surely you can disagree with someone without being insulting. This is where threads go off the rails.
anonymous, You really have to do better than that. The paper on the PA gene is here, so the “escapee” would have to recombine with the isoaltes below and in different places. However, the isoalte for PB2 was a different H1N1 from 1977, so that one would have had to escape also. However, most of a 1998 isolate from North Carolina was also a exact match with one of the Candian sequences, as was a 2002 isolate from Korea with another Canadian sequence, so the data would require four “escapees” just for the PA and PB2. In addition, one of the Canadian isolates has extensive homology with the 5′ end of the 1931 isolate listed below, so that one had to “escape” but change somewhat and tehn recombine. For the other genes, more “esapees” would be required and they all had to get into many different swine on many different farms in southern Canada.
One escapee is a stretch. Two are getting close to nonsense. Three are in the utter nonsense, and four or five are in the Martian range. The “lab error / escapee” argument is well beyond the Martian range.
A/swine/Alberta/56626/03(H1N1)
A/swine/Ontario/11112/04(H1N1) identity between 721–1319
A/swine/Ontario/23866/04(H1N1) identity between 992–1745
A/swine/Ontario/48235/04(H1N2) identity between 150–2016
A/swine/Ontario/53518/03(H1N1) identity between 25–1469
A/swine/Ontario/55383/04(H1N2) identity between 589–1787
A/swine/Ontario/57561/03(H1N1) identity between 541–1817
A/Swine/Tennessee/24/77 (H1N1)
A/Swine/1976/31 (H1N1)
pogge, Most of the readers of this board have little scientific background. The few that do put up comments that are very charitably call “hand waving”. To try to say that labs made dozens or errors is not really addressing the data.
The “humming” was really a major disinformation campaign, with serious gaps in credibility. It’s somewhat like members of the flat earth society running out and grabbing an atlas of maps and saying “See, the earth is flat because all of these maps show that it is flat”.
Statements made without a shred of evidence, are commonly called “hand waving” and the easiest way to disprove the hand waving call is to simply come up with a shred of evidence.
At this time, no one has indicated that multiple entries at GenBank are mislabled, or that multiple strains of H1N1 escaped from labs and converged on southern Canada (and in reality, they would have to to show that all samples were mislabled or all samples escaped.
comments that are very charitably call “hand waving”.
Find a more charitable way to put it anyway. I wasn’t asking for a debate about the moderating.
anonymous, No conspiracy required. As I said earlier, many of the sequencers really don’t know how to read their sequences. Even if all those H1N1’s escaped from freezers and floated into pigs on farms in southern Canada, that would only address the exact match aspect. It would not adddress the recombination, which Monotreme has acknowledged is real, if the Olsen lab didn’t make dozens of errors.
Thus, your comments suggest the authors didn’t mention the recombination to you, they just asked if you were a recombinomics employee!
The bottom line is their paper has OBVIOUS examples of recombination, which they maintain was not an artifact, but still failed to mention the recombination in their peer reviewed paper (and probably still maintain that there is no recombination - because it was in all samples and virtually all 8 gene segments and failing to mention it would be a gross oversight).
Many of the upholders of the influenza dogma maintain they can’t find recombination. I am not doubting their comments, just their ability to recognize recombination, even when it is quite obvious. They have used “differential evolution” to expalin recombination, which I would also put in the nonsense category.
As I said earlier, the influenza community is largely in the second stage of denial (athough for the swine sequence identities, Monoteme is clearly at stage 1):
Stage 1. It can’t be true.
Stage 2. Even if it is true, it can’t be important.
Stage 3. We knew it all along.
Niman,
Let’s work on this communication issue together.
Monotreme and his cooperation is your best possible opportunity to have the readers here on FluWiki understand and follow your insights into recombination. His questions are pointed, no doubt, because he wishes to fully understand your viewpoints and the evidence that you suggest is pertinent. Please be certain you understand that he is your best potential ally if you are able to present your information in an organized and complete fashion.
If we go through the same revolutions that we have in the past with our weak forms of argument based on personalities and avoiding the clear-cut communication of the facts, then we’ve again wasted it all.
Last time, someone baited you inappropriately. This time, you are baiting a potential ally inappropriately.
Take a few minutes and think about how great it will be to have another fine sounding board like Monotreme to help you hone your explanation of elegant recombination and rapid, rules-based recombination.
I won’t be available until much later today, so I hope to return and find unrelenting evidence posted here that everyone can understand to communicate exactly what you’ve discovered and its impact on our futures.
NS1 – at 12:58
Agreed
Monotreme,
Have you considered that even the small probes being used are returning homologies with similar positioning within the segment? So we’re not randomly matching a substring within a larger, potentially unrelated, string. We’re matching the substring in a place/position where we expect it to be.
Changes the probability, right?
Niman,
Your stages are entertaining, but not relevant among friends.
Be happy that some here may still be at Stage 1 and have not reverted to
Stage 0 - Don’t care if it’s true, because the explanations are so erratic?
Encourage those people at Stage 1 to remain there until they can get enough clarity in explanation to understand.
You baiting of Monotreme is very poorly considered.
I am one of very few here who follow your varied explanations, have investigated your claims and feel that you are truly onto something very important.
Let’s bring a few more people back into the investigation mode by working with a high level of civility. You will find less confusion when people are less inflamed. Then perhaps your explanations will begin to build into something sound in their minds.
You are here, as we are, to
Gather and Solve
Niman,
Please understand that many here, including the chief investigators, may not have a full or clear understanding of what you are describing and what your aren’t describing.
Several of your key and foundational concepts are still poorly understood because so much of the intercourse here is in invective, rather than well-defined academic verbiage.
I don’t think that most here recall that you’ve regularly reported your observations of rapid genetic acquisition, for example.
thanks niman. No, I didn’t look at all the genes. I wished, there were a table with all the data, so we needn’t redo it. But when we can’t explain something we won’t first search the reason in recombination. I’m not so comfortable with the explanation that there just were no opportunities of recombination in those swine and thus they didn’t mutate. There should be some other reason…We should test this. Olsen didn’t mention recombination, I mentioned it to him earlier. But that’s not what I meant. I vaguely remember a paper where they examined this virus for another reason and didn’t mention the recombination, which I found strange. I don’t remember the paper now.
NS1, Actually, I wasn’t baiting, but looking for same reasonable response. Instead I get the sequences are mislabled or the H1N1’s are all labe escapees (these sequences have been discribed fairly extenisvely in the past. Now the data from H5 birds indocate they are serving as a reservoir for birds (and humans, but I won’t even attempt to describe the human side. I actually will be discussing thsi at the Vaccine meeting next month, and thought I might get some insight, but I was clearly wrong.
I began posting with major reservations, and those reservations appear to be well placed. I don’t think this site is appropriate for a serious scientific discussion. There are a few with scientific backgrounds, but they really haven’t done much with flu sequences. Others have no background, so they ask many basic questions, which I have probably answered too many times in the past.
The risk / reward ratio on these boards is close to zero and the discussions are very hard to justify.
niman, when you want insight, discuss with the experts ! Write papers ! Don’t expect too much from a forum like this. Sometimes you might get some interesting discussion, but this is unpredictable. BTW. this isn’t your thread, it’s monotreme’s.
anonymous, Olsen has never mentioned recombination in any papers.
For the Canadian swine, I gave many genes because the recombination is quite common and easy to see in the swine because they evolve in “slow motion” so big chunks remain after several years (in many of these examples they remain for 27 years). Even the human sequences evolve slowly in swine (which is why the PB1 sequence matches human sequences of the mid 90′s even though the swine isolates are from 2003 and 2004 - this slow motion is the rule not the exception).
This slow evolution is explained by most in the field by a lack of selection pressure. I know of no one in the field who thinks it is because all of the sequences are mislabeled, or all of the isolates have recently escaped from labs.
You are correct on insight. It really isn’t here at this time (and probably never will be - absent insight, I was looking for reasonable objections, but instead got the Martian Chronicles.
I mis-stated the risk / reward ratio. It is not near zero, it is near infinity.
Oh dear…
this is of negative infinite unlikelyhood. You were the only one here with the Martian argument… The slow mutation rate could be somehow encrypted in the sequences. Or some host contribution. I can’t imagine it’s just lack of recombination with H3N2 or other H1N1. Olsen didn’t comment. Apparantly there is no paper about it. Not even some speculation… Neither yours nor Monotrme’s explanations look good.
and I couldn’t verify that flu mutates slower in swine in general. (examinating all swine sequences from genbank, or only all H1-, I don’t remember)
Monotreme,
I admire your ability to remain engaged in the dialogue without resorting to insults. Thank you. Also thank you to pogge and NS1 for (so far at least) keeping this thread on the rails.
Niman says: “I know of no one in the field who thinks it is because all of the sequences are mislabeled, or all of the isolates have recently escaped from labs.”
Is this a misrepresentation of what you have said, or have you indeed indicated you believe all of the sequences are mislabeled, or all the isolates escaped?
I appreciate your willingness to consider every avenue (even those about which you have serious doubts) with an open mind.
Scaredy Cat, I am talking about the Canadian swine sequences. Just one throws the “random mutations” into the circular file. They ALL have to be mislabeled (in most genes, but clearly in PB2 or PA) or all are escapees because they all have major regions of IDENTITY (which is why the “lab error” argument is clearly in the Martian Chronicles).
niman,
in other words, you believe the “lab error” argument is highly implausible?
Yes, because the “lab error” would require 50–100 errors by the same lab, who according to anonymous, has denied any errors.
niman – at 13:38 I began posting with major reservations, and those reservations appear to be well placed. I don’t think this site is appropriate for a serious scientific discussion.
It is rarely useful to denigrate the abilities of those with whom you are holding a discussion. Doing so places you in the position of committing the logical fallacies of ad hominem and/or “appeal to authority.” Each of those fallacies weakens your case, and makes it appear that your arguments cannot stand on their own merit.
If you truly do not think this site is appropriate for your discussion, there is an easy solution to that.
Scaredy Cat, I didn’t give much background on the Canadian swine, because the sequence originally came out a few months ago. There were some legnthy discussions, and I e-mail Monotreme some of the sequences. He agreed that the data showed clear-cut recombination. However, now he is backtracking and trying to say that the clearcut recombiantion is due to lab error. However, for it to be lab error, a rediculously high number of errors would be required.
I wrote up commentaries on PA and PB2 because they were the first two genes I looked at in depth because the recombination was so obvious. However, not only was the recombination clear cut, but the matches with parental sequences were IDENTICAL. The most common sequence among the swine isolates were two different 1977 H1N1 isolates from Tennessee. One was in six of the 2003/2004 PB2 genes and the other was in six of the 2003/2004 PA genes. The regions of identity were quite large, representing more than half of the gene in most instances. However, not only did the swine have the 1977 genes (where the match region was an EXACT match), but other portions of the genes were EXACT matches with a 1998 (from North Carolina) or 2002 isolate (from Korea). Thus, for the 1998 isoalte, not only did it show reconbination, but it showed that the 1998 protion had not changed in 6 years, because it too was exact. The same thing was true for the 2002 gene. Thus, these regions had been replicated for 2–27 years without error and involved several parental sequences. The same type of recombination was true for the other 6 genes, but the parental sequences were not as old (they were from the 90′s, not 70′s), but still showed that the sequence could be maintained for many years without chnage. In fact one of the Canadian PA genes had high homoolgy with a 1931 swine H1N1 (this was not exact, but was clear that most of the 5′ region of the 2004 gene came from a 1931 gene).
Thus, for the data to be lab error, the lab would have to keep sticking the 1977 sequence in at some places, and then stick in the 1998 sequence or the 2002 sequence. These insertaions were at different position and of varying length. Then they would have hasd to use another 1977 gene for another 2003/2004 set of genes, and again stick those sequences in at various places. This process would have to be repeated for the other 6 gene segments, so the number of errors would have to be 50–100.
Labs don’t make that many mistakes. It is unlikely that they even had the matching sequences, but even if the sequences existed, it would be hard to imagine getting three sources into one sequence. The paper had 56 sequences from isolates that were not wholly human so the mulitple origins would have to be repeated again and again.
Even if there were errors or mislabing in most of the sequences, the ones without errors would still create a problem because they were EXACT matches of regions of genes from 1977 or the 1990′s, even though the swine were replicating the H1N1 or H1N2 in 2003 or 2004.
Lab error is well beyond implausible.
I would like to interrupt this train(wreck) of thought by going back to a point raised by Monotreme on 11 July in this thread, page4. In that thread, he cites a 1988 paper, “Evolution to predominance of swine influenza virus hemagglutinin mutants of predictable phenotype during single infections of the natural host. “, and uses it to argue that mutations are the predominant mechanism in the shift from L to H in a certain H1N1 strain. I read the article, and have some serious concerns about whether it can be used as fodder in this discussion.
There are several methodological and analytical problems with using this paper. First, the H and L sera are defined by phenotye, not genotype. While the paper cites a single marker, L or G at position 153 of the hemagglutinin gene, it does not say that they sequenced the entire thing and that is the sole way to tell the two strains apart. They do not establish that the viruses are monoclonal otherwise. Thus, the original sera might have any number of differences between them in addition to position 153.
They also mention that the H sample virus was contaminated at a low level by L virus, but blew it off by saying that there were only about 40 individual virions per dose, so it’s not important.
They also only infected the pigs intranasally, they did not try the oral route. It’s possible that the initial infection in the oropharynx seleceted against gut receptors, and by the time the piglets were swallowing snot, there weren’t enough of the L virons to try for a gut approach.
Again, they didn’t sequence their results, they only identified them with monoclonal antibodies. So if recombination was at work, and there was some heterogeneity in the parent virus mixture, there’s no way to know.
Because they didn’t sequence either the parent viruses or the daughter viruses, but only looked at position 153, their whole argument is kind of weak. We still don’t know if this is binding site selection or random mutation, or something else.
Scaredy Cat,
hee is more background. Swine flu viruses in both the United States and Canada have been undergoing some significant chnages via reassortment, highlighting the dual infections. Prior to the mid 1990′s only H1N1 was found in North America (called classical swine). This was like the H1N1 from the first flu virus from pigs in Iowa in 1931. However, in the mid-90′s H3N2 was found in pigs. These viruses had three human genes (HA3, NA2, and PB1). The still have three classical swine genes (M, NP, and NS). That was followed by H1N2 which swapped out human H3 for swine H1. At teh same time there was swapping in humans and H1N2 was identified (H1 from H1N1 and N2 from H3N2). Then human H1N2 was in swine (all 8 genes were human). Some of teh recent Candian swine had 7 swine genes with human PB1, which is the case for Ontario/11112, which has a 25 BP match wity the mallard H5 that was just released. This region contained two tandom polymorphisms not seen in other H5s.
Thus, the above chnages show that pigs are picking up human and avina genes and they are reassorting and recombining, which is why they have examples of recombiantion in slow motion. The PB1 in these recent isolates are human, but like human PB1 in the 1990′s although aportion recently reappear3ed in recent human PB1′s.
The swine are acting as a reservoir and creating new human and avina genes via recombination and this recombination is NOT rare (or cherry picked).
Scaredy Cat, The subject of the Candain swine was also re-opened because of the H5 sequence from a mallard in British Columbia. Last August, birds were swabbed as part of a banding experiment across southern Canada. In Bristish Columbia, a remarkable 24% of the birds tested were H5 positive (H5N2 or H5N9). The HA released this week and although it was low path and was most closely related to North American isolates, it had two unique polymorphsims, which were also in one of the Canadian with the 1977 PA and PB2 genes, identifying donor swine sequences ro the wild bird H5 in British Columbia.
Journal of General Virology (1999), 80, 2535–2543. Printed in Great Britain
Evolutionary aspects of recombination in RNA viruses Michael Worobey and Edward C. Holmes Department of Zoology, University of Oxford, South Parks Road, Oxford OX1 3PS, UK
This provides interesting data and commentary on recombination.
The Holmes paper predicted that as more sequences became available, there would be more example. The paper you cited was written in 1999. The flu database expanded most dramatically in 2004, after H5N1 exploded out of China and caused hman fatalities in Vietnam and Thailnad.
The “natural” recombinants the emerged after selection are detailed many times over here, even with an incomplete database that excludes regions of genes were recombinantion has the highest probability.
Science Teacher – at 18:08
Thank you, The information was very helpful to me.
Science Taecher, The paper you linked commented several times on “natural” isolates because it is well known that not all experiments mimic nature. A good example of that is the in vitro errors on copying. As noted in the Holmes piece, what is seen in the test tube may not appear in nature because most of what is made in vitro isn’t selected in nature. Most scientists modify their hypothesis or procedures when the data they generated fails to properly represent nature.
The sequneces at GenBank are about as close to nature as you are going to get. Some PCR data may be more reflective, because of selection of isolates (like the loss of S227N because the virus was isolated in chicken eggs). However, in genral, the sequences at GenBank are going to be a closer approximation than some in vitro data.
I gave you a link to many clearcut examples of recombination. The Canadian swine offer many more. These are not “cherry picked”. The swine do present a clearer picture because the virus evolves more slowly in swine, so it can be caught in the act of recombining.
There are additional clear cut examples in sequences by others. This data is NOT lab error. The swine sequences are useful because there are so many examples in just one paper.
Earlier I had posted another example of an exact match of M in a 2003 chicken in Korea and a 1998 swine in Hong Kong (this data came out of St Jude and Hong Kong).
The examples can go on and on, but there has NOT been a counter agrument to the DATA, other than “lab error”.
Dr. Niman, personally i agree with your comments and am very interested in reading your opinions as they do add breadth to the virology discussion. Thank you for the links. My link offered historical perspective on the topic. There seem to be many folks on the web who are under the impression that recombination was invented by you and that it is a new concept. Your rules of application are “new” and to me, worth exploring. There are many threads in the tapestry of science and I hope we can all work together as we explore different ‘threads’.
Science Teacher, Actually, the paper you linked refers back to a Mike Lai paper in 1992 for demonstrating recombination in coronaviruses. I actually took his course (on RNA viruses) in graduate school at USC. He was working on the 5th floor in Microbiology, and I was on the second floor in Pathology. We both regularly went to the Cold Spring Hatbor Symposium on RNA Tumor Virus each May (where I actually spent my honeymoon!).
I know Mike well, and recombination per se, is not new, Mike was working on coronaviruses, and he was frequently cited when SARS broke out. Most coronavirologist accept recombination. It too has a segmented genome and can reasort, but there is no doubt that recombination playes a role. Most don’t realize however, how big that role is. Even the “point mutations” are actually recombination, and in flu the upholders or flu dogma are furmly tethered to reassortment, which obviously happens, but recombination is the real name of the game.
Maybe a little off topic…:) But I’m thinking of the feline sequences. Isn’t it quite common to eat cats and dogs in east Asia?
Alot to comment on here. I won’t be able to get to most of it until later tonight, but I wanted to get in some thank-you’s in while I have a chance.
pogge, I appreciate your moderation. I think you do a great job as a moderator.
beehiver, glad to see that someone agrees with me :-)
NS1 and Scaredy Cat, your kind words mean alot to me. Please understand that even if we don’t agree on the issue of Recombinomics, I have the highest degree of respect for you. I think we do agree on almost everthing else.
No-one should worry that I will respond to Dr.Niman with insults. I have been harshly critical of the WHO and the CDC and have pulled no punches there, but I will not attack Dr. Niman, for a variety of reasons, regardless of provocation.
Now the Theory of Recombinomics, that’s another thing ;-)
More later…
This is getting pretty long so I’ve opened up the next thread in the series.