PREAMBLE:
One of the central question around H5N1 has always been: how likely is it to cause the next pandemic? We may talk about emergence of dominant strains or h2h, but we must also not lose sight of some of the basic assumptions in our thinking.
The most basic one is this: can an avian influenza virus become adapted to humans by gradual evolution (ie without reassortment with a human virus) and establish itself sufficiently to transmit efficiently, not just causing sporadic outbreaks but establishing a stable lineage in a new species?
The biggest support to an affirmative answer comes from the characterization of the 1918 virus, where Taubenberger’s work suggests that this was indeed what happened.
There are however sceptics like Palese who think that the issue is unresolved.
His rationale goes like this: he is not convinced that it is possible to establish that the 1918 virus was avian in origin unless you can compare it to the human influenza samples pre-1918 (which Taubenberger is now actually doing, but more about that later). If you accept his argument, then there is no basis to believe that avian viruses can cause human pandemics. Therefore, he doesn’t think that there are enough grounds to assume that H5N1 can cause a pandemic that way.
Go here 05:11 for translation of Palese interview
Go here for Taubenberger’s rationale for the origin of 1918 virus
A recent paper published online in the Journal of Virology approaches the problem in a very different and novel way, and comes to the same conclusions as Taubenberger, both on the avian origin of the virus, and also that it had entered the human population shortly before 1918.
ABSTRACT:
In the last few years, the genomic sequence data for thousands of influenza A virus strains, including the 1918 pandemic strain, and hundreds of isolates of the avian influenza virus H5N1, which is causing an increasing number of human fatalities, have become publicly available. This large quantity of sequence data allows us to do comparative genomics with the human and avian versions of the virus. We find that the nucleotide compositions of influenza A viruses infecting the two hosts are sufficiently different that we can determine the host at almost 100% accuracy. This assignment works at the segment level, which allows us to construct the reassortment history of individual segments within each strain. We suggest that the different nucleotide compositions can be explained by a host-dependent mutation bias. To support this idea, we estimate the fixation rates for the different polymerase segments and the ratios of synonymous to nonsynonymous changes. Additionally, we provide evidence supporting the hypothesis that the H1N1 influenza virus entered the human population just prior to the 1918 outbreak, with an earliest bound of 1910.
This is quite a technical paper and hard to read, at least for me. Essentially, they have found that avian and human viruses tend to have consistent differences in their nucleotide compositions. The human viruses have a higher percentage of uracil (U) and adenine (A) in their genomes, while the avian viruses have a higher percentage of guanine (G) and cytosine ©.
The only exception to this consistency are H5N1 and H9N2 in humans, and the 1918 H1N1. All these are avian viruses that have infected humans.
This chart shows how it works. (NB. Anything above zero is human, below is avian.) You can see that the 1918 H1N1 is clearly avian, while subsequent human isolates of H1N1 gradually moved up to become the most human of all the samples.
Taking each segment individually, the most interesting one is PB1, shown here. The red crosses are all human samples except H1N1. Notice how the 1918 H1N1, 1957 and 1968 samples are all clearly more avian than their corresponding later samples. This corresponds to what we know from other evidence, that the circulating human virus acquired a new PB1 gene each time in 1957 and 68.
The authors claim that this method of predicting host is 100% consistent.
Using this method, they then computed a parameter that they call the U content, which represents the degree of substitution eg C→U changes over time. Mapping this for the 1918 and 1933 H1N1, they came up with this chart. On extrapolation, they arrived at the conclusion that the 1918 pandemic H1N1 entered the human population no earlier than 1910, probably later.
The following quote from the paper explains it further:
“Extrapolating backwards from the 1918 flu, given the fixation rate matrices, we can put an upper bound on the length of time the H1N1 strain could have evolved within the human population prior to 1918. If the virus had been introduced to human hosts before 1910, the U content would be below the level found in any flu strain in the NCBI database for any host species. So the virus likely did not make the jump to humans more than 8 years before 1918, and probably less. We cannot say for certain that the virus moved straight from an avian to a human host without evolving in another organism for a period of time. However, the nucleotide composition of all the segments from the 1918 H1N1 strain is consistent with its being an avian virus.”
Isn’t an intermediate host suggested by the mismatches in Indonesian human and avian strains ?
interesting. Seems to be mainlay an increase in “T” in human viruses.
For Indonesia I get for the ratios of A,C,G,T in PB1:
Avian H5N1 : 3543,1970,2306,2179 (471) samples
Human H5N1 : 3460,1970,2334,2235 (36 samples)
so there is some increase in T already, but not as much as e.g. in H3N2:
Avian:3600,2064,2284,2050 (97 samples)
Human:3546,1959,2265,2228 (1370 samples)
anon-22: if you have a hard time reading that article then forget about me. But dwasn’t there also suspicion that 1918 crossed thru pigs? And aren’t plenty of pigs in Asia currently carrying H5N1 as well? Seems everyone there is so busy studying chickens, never finding a precise link between birds and humans in all cases, that we need to also be looking elsewhere.
Also, is the lack of data points before 1970 due to that, lack of data, or is there freakish explosion going on demonstrating that not only H5N1 but many other things are on the verge of erupting? We could go more years with not much happening on the H5N1 front, then, whicle no one is paying attnetion, get hammered by a sneak attack by something else.
Swine seem to be between birds and humans in their A+U vs. C+G characteristics. Also 2 or 3 mutations at the RBD in 1918 which are swinish.
OK, let’s test the theory. Shown below is the amino acid sequence of the PB2 segment on an H1N1 Type A virus. Now, please tell us was the host 1. Avian? or 2. Human?
MERIKELRDLMSQSRTREILTKTTVDHMAIIKKYTSGRQEKNPALRMKWMMAMKYPITADKRIMEMIPERNEQGQTLW SKTNDAGSDRVMVSPLAVTWWNRNGPTTSTVHYPKVYKTYFEKVERLKHGTFGPVHFRNQVKIRRRVDINPGHADLSA KEAQDVIMEVVFPNEVGARILTSESQLTITKEKKEELQDCKIAPLMVAYMLERELVRKTRFLPVAGGTSSVYIEVLHL TQGTCWEQMYTPGGEVRNDDVDQSLIIAARNIVRRATVSADPLASLLEMCHSTQIGGIRMVDILRQNPTEEQAVDICK AAMGLRISSSFSFGGFTFKRTSGSSVKREEEVLTGNLQTLKIRVHEGYEEFTMVGRRATAILRKATRRLIQLIVSGRD EQSIAEAIIVAMVFSQEDCMIKAVRGDLNFVNRANQRLNPMHQLLRHFQKNAKVLFQNWGIESIDNVMGMIGILPDMT PSTEMSLRGIRVSRMGVDEYSSTERVVVSIDRFLRVRDQRGNVLLSPEEVSETQGTEKLTITYSSSMMWEINGPESVL VNTYQWIIRNWETVKIQWSQDPTMLYNKMEFEPFQSLVPKAARGQYSGFVRTLFQQMRDVLGTFDTVQIIKLLPFAAA PPEQSRMQFSSLTVNVRGSGMRILIRGNSPVFNYNKATKRLTVLGKDAGALTEDPDEGTAGVESAVLRGFLILGKEDK RYGPALSINELGNLAKGEKANVLIGQGDVVLVMKRKRDSSILTDSQTATKRIRMAIN*CRIV*KRPCFY
Oh no! Side scroll. Help from the Mods?
Sorry to all for the inconvenience.
Influenza A virus (A/duck/Minnesota/1375/1981(H1N1))
I actually tried to use their method and the data from gs above. But it doesn’t work very well. MaMa’s sequence has different proportions than the other dataset, particularly, the number of A was proportionately lower, and the number of G was proportionately higher. I don’t think that averaging the data helps us here. There appears to be a lot of scatter in the data, and if the means and standard deviations were included, there would be a lot of overlap. The whole idea that the strains evolve differently in different organinims is contradicted by the sequence data anyway. The sequence data show that there is a lot of homologous recombination going on, and that even the new Fujian polymorphisms are showing up in human H1N1. There just doesn’t seem to be a bright line difference between mammal and avian subtypes.
Mamabird and anonymous,
Well, I don’t know if you have access to the paper so my remarks may not apply. And as I said their mathematical approach is novel and not something that I will try to test out on my own. :-)
My own focus is to assume their methodology, to the extent that it was reviewed and accepted by a world-renouned journal, is valid, and then follow their results to see what observations or conclusions ensue.
Also, this method is not used to determine swine strains. The bias was only observed and calculated on avian vs human sequences.
a’Akova – at 02:11 Isn’t an intermediate host suggested by the mismatches in Indonesian human and avian strains ?
That would be one explanation that needs to be explored.
However, it could also be due to insufficient samples or other errors. We know so little of the actual data, particularly whether any data that anyone does see is representative of the true extent of the problems both in human and avian outbreaks, that it is very hard to make a judgment. The intermediate host thing is definitely something to look out for, but not something that we are certain enough to build any other theories on, unfortunately.
LauraB – at 06:00 anon-22: if you have a hard time reading that article then forget about me. But dwasn’t there also suspicion that 1918 crossed thru pigs? And aren’t plenty of pigs in Asia currently carrying H5N1 as well? Seems everyone there is so busy studying chickens, never finding a precise link between birds and humans in all cases, that we need to also be looking elsewhere.
First of all, if you read the recent report from scientists writing the WHO report on influenza research, even though they didn’t come outright to say whether there were or were not H5N1 isolates from pigs, the focus appears to have shifted more towards cats than pigs. That gives us a glimpse of what the scientists are thinking.
Even without that, if we examine the theory behind pigs as intermediate hosts, there are actually quite a lot of flaws or incomplete arguments, some of which I wrote up for 1918 here.
In addition, there have been lots of instances of human viruses found in pigs but only very rarely do we have swine viruses found in humans. The only time was 1976 swine flu which never took off. In his recent videocast, Great Teachers - Influenza: Past Pandemics and Future Threats, Taubenberger is now saying very clearly that human virses are more likely to infect pigs than pig viruses to infect humans.
Also, is the lack of data points before 1970 due to that, lack of data, or is there freakish explosion going on demonstrating that not only H5N1 but many other things are on the verge of erupting? We could go more years with not much happening on the H5N1 front, then, whicle no one is paying attnetion, get hammered by a sneak attack by something else.
Yes, to your last question. That is always possible.
I don’t know the answer to the 1970 question, but I would assume that we won’t have as many isolates the further back in time we go.
It’s unlikely to be due to insufficient recent samples. Once you have more than about 30 samples, you can start getting reliable means and standard deviations for normally distributed data. It could certainly be due to sampling bias, particularly, to biases against detecting dual infections. It could also be due to a bias that assumes that the changes in the genome are due to random mutations only.
Think about this: Supposing you were grading term papers and trying to identify plagarism. Would you look for letter frequencies, word frequencies, or would you compare whole paragraphs? Could you use letter frequencies alone to determine when Shakespeare was written, based on comparing 100 2006 documents and 100 1950 documents, and graphing the results? I kind of doubt it, although it might be interesting to try. But I think most people could read a few lines of Shakespeare and know they weren’t reading modern English.
The problem with the random mutation school of analysis is that it predicts slow and stupid evolution. Is that a safe bet to take?
anonymous – at 19:01
Right on. And if you look at the MP segment in the H1N1 1981 MN duck, you can find virtually the identical MP genetic signiture in an H5N1 1981 MN duck, and in human host viruses of that time period.
I believe it is safe to safe that Type A influenza viruses in dually infected hosts are trading genetic information at will, either through reassortment of the entire gene segments, or through recombination of material within segments. Further, it is likely that birds are infecting humans and that humans are infecting birds. So, while you can usually come fairly close to identifying whether a sample is from an avian or human host,I really believe it is a stretch to state:
“We find that the nucleotide compositions of influenza A viruses infecting the two hosts are sufficiently different that we can determine the host at almost 100% accuracy.”
I think I’m getting it…but if you had to sum this up in a few words what significance is this? Is it simply that in 8 years or less from being introduced, H1N1 became pandemic?
Patch – at 20:27
Perhaps the concept that is relevant is that avian viruses that infect humans can be lethal but not very infectious. And that human adapted viruses can be infectious, but not very lethal. And that avian viruses that jump species begin to quickly adapt to their new hosts. And finally, it is that transition period from avian to human that is important to watch.
If the transition occurs through genetic drift as is suggested for the 1918 H1N1, it may take eight years for the virus to become a pandemic problem. However, if the transition occurs through genetic mixing, perhaps we on the verge now of H5N1 becoming an issue, especially the Qinghai variety which is acquiring quite a number of significant human characteristics.
anonymous – at 20:16
I’m not sure I understand your point. I don’t think this paper is trying to elucidate whether the changes observed were due to random mutations or other mechanisms. There was certainly no reference whatsoever to dual infections, (nor do I see how that would affect the results).
All that they are doing, is report a mathematically valid new method which appears to be consistent with what we know. eg the PB1 gene being changed in 1957 and again in 68, or that avian samples all fall into the avian range, human samples into the human range, but samples from humans infected with avian viruses fall into the avian range.
btw, your analogy to Shakespeare is not valid, in that we are able to read Shakespeare and know when it is not modern English, because we already know the different paramenters of Shakespeare vs modern English, but we are not able to read viral sequences in this new way until we have established the parameters that tell us that this new way is valid.
Which is all that this paper is trying to do.
And the results suggest that we can, and it is a valid and useful tool to add to existing tools. That’s all.
It’s important IMO in science to know when to extrapolate and how far. In this instance, not too far at all towards what I think you are talking about.
Patch – at 20:27 I think I’m getting it…but if you had to sum this up in a few words what significance is this? Is it simply that in 8 years or less from being introduced, H1N1 became pandemic?
That would be one of the things that came out of this paper, among others that might be a bit more esoteric for most. :-)
The other one would be that it gives another layer of support to the avian origin of the 1918 H1N1, which as you read from my first post is by no means accepted by all.
This is clipped out of a much longer time sequence of a polymorphisms in the HA gene, tracking the evolution of the new Fujian strain. Note the jumping of the polymorphism between waterfowl, people, and swine. A22, Taubenberger must simply not be looking at sub-gene data. Too bad; he’s missing quite a treat.
If the nucleotides were mutating randomly, then future releases of sequence data will show random divergence and will not dovetail with these kinds of travelogues. Let me make a Prediction based on the existing sequence data: Future released sequences will fit right into such a travelogue. If they don’t, I will sit down and shut up. Actually, I’m cheating, because I already know the answer. This travelogue already shows that the newly released sequences fit right into the existing polymorphism travelogues. And hey, isn’t prediction the very heart and soul of the scientific method?
So let’s have a few more released sequences, say, from Weybridge or the CDC, and see how they fit into the list. This seems like a pretty simple challenge to me.
DQ992760 A/duck/Guiyang/3834/2005 HA (4) 1647 2005 H5N1 DQ992761 A/duck/Guiyang/3996/2005 HA (4) 1689 2005 H5N1 DQ992942 A/duck/Yunnan/4418/2005 HA (4) 1056 2005 H5N1 DQ992900 A/goose/Guiyang/3765/2005 HA (4) 1059 2005 H5N1 DQ992906 A/goose/Guiyang/4006/2005 HA (4) 1017 2005 H5N1 DQ992907 A/goose/Guiyang/4021/2005 HA (4) 1059 2005 H5N1 DQ992908 A/goose/Guiyang/4030/2005 HA (4) 1059 2005 H5N1 CY016459 A/Waikato/11/2005 HA (4) 1737 2005 H1N1 CY013581 A/Waikato/13/2005 HA (4) 1720 2005 H1N1 CY004546 A/pintail duck/Alberta/210/2002 HA (4) 1777 2002 H1N1 CY016260 A/New South Wales/26/2000 HA (4) 1721 2000 H1N1 CY011012 A/Wellington/5/2000 HA (4) 1737 2000 H1N1 AY684891 A/black-headed gull/Sweden/5/99 HA (4) 1760 1999 H16N3 AJ412711 A/swine/Cotes d’Armor/1482/99 HA (4) 1002 1999 H1N1 AJ344002 A/swine/Cotes d’Armor/1515/99 HA (4) 987 1999 H1N1 AB117200 A/Ishikawa/42/1998 (E2) HA (4) 978 1998 H1N1 CY004507 A/mallard/Alberta/201/1998 HA (4) 1777 1998 H1N1 CY016236 A/Nanchang/11/1996 HA (4) 1725 1996 H1N1 CY013813 A/Nanchang/13/1996 HA (4) 1727 1996 H1N1
sorry about the formatting, maybe this is better:
anon_22 – at 20:40
There is really no dispute that the 1918 Brevig Mission H1N1 is very, very avian in its genetic characteristics. However, do we know how it got the way it did?
In other words, did an avian virus infect a human and through genetic adaptation and drift become enough human like to reach pandemic status? Or did an avian host, dually infected with H1N1 and an H3 human virus swap genetic material, become more human like and then jump the species barrier with catastrophic results?
We of course do not know as yet because we neither have a wealth of genetic information from 1917 birds or 1917 humans to say with certainty. However, Dr T is attempting to come with some answers based on Type A positive English patients of a pre-pandemic time frame.
Stay tuned.
Let me summarize the logic of the (for want of a better description) Taubenberger vs Palese debate.
That means one more nail in the coffin for the idea that H5N1 is unlikely to cause the next pandemic because there is no evidence that avian viruses can do that.
Folks understand me, I am not a scientist by any stretch of the imagination, and really do not want to enter into this discussion. However, the below data seems like it would play well here, and if I read the article correctly, Then a change in the NS1 gene increasing the amino acids toward a human form is what you are looking for. (Data from Indonesia human cases would seem the place to start. Or what is the difference in the NS1 gene and the ammo acid built in chickens, cats, pigs and humans.
New paper on H5N1 virulence gene Category: Bird flu • biology Posted on: November 8, 2006 7:44 AM, by revere
<snip>
In 1996 a highly pathogenic H5N1 virus caused an outbreak on a goose farm in southern China, Guangdong province. Two H5N1 viruses were isolated from the geese, who suffered a 40% mortality. One, designated A/Goose/Guandong/1/96 (GS/GD/1/96; see our entry in the Flu Wiki on naming influenza viruses) has been much studied; the other, designated GS/GD/2/96, less so. The reason for the lesser interest in the second virus was it was of low pathogenicity for geese and chickens. Its isolation was more of an incidental finding in the 1996 outbreak. The Chinese scientists have compared the two viruses and obtained interesting information about what makes one virulent to chickens and the other not.
Both viruses have the polybasic amino acid sequence adjacent to their cleavage site. Since one is high path for chickens and geese and the other not, it is clear this is not a sufficient condition (something which we’ve known). When the sequences of all the gene segments in the two viruses were compared, they were identical in the PB2, HA, NA, M2 and NS2 genes at the nucleotide level (the finest level of detail). In four other gene segments there were a total of eight nucleotide differences, of which three were “silent” mutations, that is, they were differences in the nucleotide sequences that didn’t produce a difference in the amino acid sequence of the proteins. The other five changes produced differences in the PA, NP, M1 and NS1 genes. Thus, the differences in virulence to chickens in these two viruses were due to one or more changes in these four genes.
The Harbin team then constructed recombinant variations combining the genes in different ways. It turned out that only the change in position 149 of the NS1 gene determined whether the virus was virulent for chickens. It didn’t seem to affect virulence in geese. Since NS1 _expression in mammals is related to antagonism of a non-specific innate antiviral response, the release of interferons alpha and beta, they went on to show that in birds, too, the difference in NS1 in the two viruses is related to whether it can inhibit interferon production.
< snip >
If I am out in left field, just ignore this stupid post. If not maybe something for the really smart people to chew on.
anon_22 – at 20:53
What would be your view if Dr T found H3 Type A in the 1917-early 1918 patients? Would you take the position that the H1N1 pandemic was a reassorment? If so, why?
Mamabird – at 20:49 anon_22 – at 20:40
There is really no dispute that the 1918 Brevig Mission H1N1 is very, very avian in its genetic characteristics. However, do we know how it got the way it did?
That, is what everyone wants to find out. :-)
In other words, did an avian virus infect a human and through genetic adaptation and drift become enough human like to reach pandemic status? Or did an avian host, dually infected with H1N1 and an H3 human virus swap genetic material, become more human like and then jump the species barrier with catastrophic results?
You are still talking about reassortment, whether it happened in a bird or human. Taubenberger would say that the data from all the gene segments suggest there was no reassortment. And the current test, at least for the largest gene segments, less clearly so for the others, also gives the same result.
We of course do not know as yet because we neither have a wealth of genetic information from 1917 birds or 1917 humans to say with certainty. However, Dr T is attempting to come with some answers based on Type A positive English patients of a pre-pandemic time frame.
Stay tuned.
Exactly. That’s why their work on the pre-1918 samples are so important, and that’s why the fact that they found some flu RNA positive cases for the first wave is some exciting. These may or may not translate into real answers, as we are talking about tiny fragments. But it gives us ground for optimism for further understanding the many complex issues involved.
Anon_451 – at 21:00
One small problem with the NS1 gene segment. The 1918 H1N1 pandemic virus had an NS1 that was virtually all avian in genetic characteristics. In other words it had not changed toward a human like form, so we may not be able to hang our hats on that one segment alone. Probably lots of factors at play here.
anon_22 – at 21:04
Not talking genetic reassortment in my earlier posts. I do not believe it is likely that entire gene segments of two viruses swaped positions to form the 1918 pandemic H1N1. It seems more plausible that a very avian virus turned a bit more human and started things on the downhill slide toward catastrophy.
Having said that, I would also be a bit careful in stating that it was only random point mutations or genetic drift. The swaping of genetic material in other viruses has been well documented at less than a segment level, and Type A influenza may also be subjected to such changes as Dr T, himself has admitted possible.
Mamabird – at 21:02 anon_22 – at 20:53
What would be your view if Dr T found H3 Type A in the 1917-early 1918 patients? Would you take the position that the H1N1 pandemic was a reassorment? If so, why?
I am unclear as to your timeframe, cos that is important. So I’m going to use both.
If we assume that historical records are correct, and that the first wave of the pandemic did not start till at least early 1918, then 1917 samples, especially those of the earlier part of the year, and the years before that, we can assume to be the circulating seasonal strain. They have so far found 31 samples positive for flu RNA of which 15 have tested positive as H3. So it is pretty clear that the circulating seasonal strain was an H3.
Now if they found H3 in early 1918, at the same time as what we think was the first wave was occurring, then one possibility could still be that this particular patient was not infected with the pandemic strain, but the regular seasonal strain. Which doesn’t help us in any way.
What you are trying to suggest, and I’m guessing here, is whether there is any possibility that a human H3 virus in early 1918 aka first wave reassorted with an avian H1 to produce the pandemic strain.
My response to that is whether that kind of reassortment happened or not is not going to be decided by whether we find H3, but by whether the clearly non-pandemic H3, ie those isolated the year before or earlier, had internal genes that were retained in the subsequent pandemic H1N1 virus.
Which would be a crucial next step in the current NIH project.
Stay tuned indeed. :-)
Mamabird – at 21:16
Actually, I am only making a distinction between reassortment and gradual adaptation. We know so little that we are not able to pinpoint it further than that.
Mamabird – at 21:09 Please bear with me as I barely understand this all. You said “The 1918 H1N1 pandemic virus had an NS1 that was virtually all avian in genetic characteristics.” Key word be virtually. I kind of understand the the PB2 receptor binder needs to have the traits to bind to the upper air way of the human to be a human to human spread still, with out the correct NS1 gene, the virulence level in the human form would be very low, if I am reading this correctly. Again I am just guessing here so flame me if you need to :-)
From Niman: Most of the human cases in Indonesia have been from western Java and most have died. H5N1 isolated from these patients has a novel cleavage site that does not match the vast majority of H5N1 birds isolates. Additional birds samples sent to Australia fro sequencing have failed to identify the source of the human infections in Jakarta.
I admit to being way out of my element when it comes to the science, but I didn’t realize that so few (if any? unclear) of the Indo cases had no avian matches. Clearly there must be another factor: mice? cats - who eat infected mice? Being a muslim country I doubt pigs are roaming the streets too often, but mice and cats? oh yes.
I’ll go back to lurking on this thread and leave it to the experts to sort out.
anon_22 – at 20:53: Taubenberger also thinks that the virus had only been in the human population for a short time, before the pandemic.
What is Taubenberger’s definition of “a short time?”
And thanks for the interpretation of this interesting paper.
Anon_451 – at 21:00
None of us (I think) are virologists here, so we are all just offering our opinion from however much we understand, so whatever you can bring to the discussion is always welcome.
The issue of the NS1 gene is important, but may or may not be in the way that you described. Remember that the paper that you quoted was talking about virulence in birds. Flu viruses can cause disease in different species, but the diseases can be very different and so also the pathogenesis. We really do not know how much a study on birds can be extrapolated into virulence in humans.
I think that is partly what Mamabird meant about not hanging our hats on that one.
Plus, it is likely that virulence (as well as transmission) is dependent on multiple gene segments working in co-ordination.
Anon_451 – at 21:22
I’m just terrible with matches. So, let me see if this helps:
Of the 11 proteins that result from the eight genes of the Type A virus, we have yet to identify with any certainty which one or ones are most important in causing a virus to become better adapted to humans. However, there is lots of research that give clues, like the HA receptor binding site and certain amino acid positions in the PB2 segment, just as some examples.
If you look at the 1918 H1N1 proteins you will see a lot of avian characteristics, with some two dozen or so clearly human like point changes creeping into the picture, but not in the NS1 segment. It pretty much looks like every other avian virus NS1 gene segment. Just no human changes there, so it may not be much of a player, at least as to the 1918 H1N1.
LauraB – at 21:28
First of all, don’t lurk. If everybody lurks, we will have no forum. :-)
Secondly, you are asking a very important question. In fact, your question is what we would call ‘right on the money’.
I personally think we need to look at cats a lot more.
Mamabird, and everyone else interested in virology.
If you haven’t done so already, may I strongly recommend that you watch Taubenberger’s latest videocast (link at 20:12 post), especially the part where he explains the receptor binding studies (which, btw, everytime I hear or read an explanation, I learn something new). The most interesting comments are in the Q&A, where he said that the receptor issue is not just whether a virus binds to 2,3 or 2,6 terminals, but which kind of 2,3 or 2,6 terminals. That there appear to be some specificity which needs more exploring, also that we don’t know, for example, whether even distribution of various kinds of 2,6 receptors might play a role in disease or mortality. Of course, I can’t explain as well as he can, that’s why I am making this recommendation. :-)
Pixie – at 21:32
What is Taubenberger’s definition of “a short time?”
Here’s a quote from his paper in Nature, Characterization of the 1918 influenza virus polymerase genes October 2005.
anon_22 – at 21:48
Excellent recommendation.
Perhaps these changes in types and distribution of alpha 2,3 and 2,6 linkages could also lead to some explanation as to why high path H5N1 hits the under 30 crowd so hard. Almost 74% of the deaths within that age group.
nucleotides A+T in %*100
A H S H-A
------------------------
PB2: 5539 5745 5606 206
PB1: 5696 5780 5763 84
PA : 5604 5773 5688 169
HA : 5770 5703 5779 −67
NP : 5255 5410 5402 155
NA : 5671 5730 5744 59
M : 5153 5289 5219 136
(A:avian,H:human,S:swine
NS : 5594 5713 5650 119
5535 5643 5606 108
Perhaps these changes in types and distribution of alpha 2,3 and 2,6 linkages could also lead to some explanation as to why high path H5N1 hits the under 30 crowd so hard.
I think that was exactly where he was heading in his thinking.
sorry. Evil fluwiki-formatting and I mangled one line and I should include deviations. I will do that later. It goes through all segments except HA and swine are between human and avian. For Indonesia there is no difference, so no indication for a mammalean reservoir with this method. Most human sequences are H3N2, most avian H5N1 or H9N2 but these things don’t disturb so much. You get similar differences when you only consider H1N1 or H3N1, but not H5N1 since there is no adaption of H5N1 to mammals yet. When this adaption starts to happen with chains of A2H2H2H..2H or A2H2A , (or with swine first) then we could be in trouble…
they should probably take PB2,PA,NP,M and maybe the non-overlapping part of NS for
their pictures instead of PB2,PB1,PA. And why using logarithms ?
Also the squares are bad, better thin lines, _ and |.
Maybe I can make some pictures… but it’s a bit tedious for me with the graphics.
What’s the explanation ? My first thought would be, that there are just more
T’s and A’s around in humans, is it true ?
here are the improved lists.
what I wrote above about the Indonesian reservoir was not quite true.
There is indeed some evidence for a starting human adaption
in Indonesia H5N1 (except in PB1) although much weaker than with H1N1 or H3N2.
But that adaption took 10–20 years and we only had 2 years in Indonesia of
human H5N1 so far.
well, it could also be, that this adaption didn’t happen in the assumed mammelian reservoir, but just by selection : viruses with more A,T could be more suitable to infect humans.
well again, it’s worse than in average H5N1, though. Especially in NP.
We should examine the other Indonesian sequences from Los Alamos too
and consider whole genomes, not isolated genes …
And maybe weighting the synonymous positions stronger.
Now, every flu-A virus should be assigned a “Humanity”-number,
mesuring it’s potential to infect humans ! And changes in this number will
measure the progress towards pandemic.
In 1918 apparantly there was no such evolution, though.
Once the avian virus went to humans (maybe through swine)
it did not much adaption to increase its humanity-number but already went pandemic
although still having a small humanity-number.
average humanity (suggest a better word !) numbers:
…….H…S…A
------------------
.all:172 87 0
H1N1:194 100 −7
H3N2:179 84 31
H5N1: 38 53 7
.IDN: 45 ? −12
anonymous – at 01:47
Now, every flu-A virus should be assigned a “Humanity”-number, mesuring it’s potential to infect humans ! And changes in this number will measure the progress towards pandemic.
I think what the authors of this paper is saying is not that viruses with more U and A are more likely to infect humans, but that after entering into human hosts, the genome gradually evolved towards more U and A. This is a very important distinction, and it means you can’t use this for prediction, only for discovering the origin of existing samples.
In 1918 apparantly there was no such evolution, though. Once the avian virus went to humans (maybe through swine) it did not much adaption to increase its humanity-number but already went pandemic although still having a small humanity-number.
In 1918, when it entered into humans, it was still avian. Over time, the virus evolved while passing through human hosts and you can see the change in the charts that I posted. (blue asterisks are H1N1 from 1918 and their descendants in human hosts).
Anon-22: I didn’t realize the extent to which it was so transmissible among cats. Given that evidence, my money is on either cats or mice. Have mice been tested? There are everywhere and no one would think twice about finding dead mice around as they have so many predators. Wild birds eat them as well which could contribute to the disease spreading. Either way, it’s rather sobering. While birds are an obvious link to human disease, the real enemy may be elsewhere, undetected.
I promise to contribute if I have something to say. But when you guys start listing H1N1:194 100 −7 H3N2:179 84 31 H5N1: 38 53 7 I am way out of my league!
however, when viruses with more A and T (=U) are more prevalent in humans, that means that selection prefers these viruses. Unless you assume that all these synonymous mutations are entirely neutral. But then, why should mutations at 3rd positions towards A and T _occur_ (rather than survive) more frequent in humans than birds ?
LauraB – at 10:05
I don’t think mice has been tested. The problem is not whether we have tested a whole range of different possible hosts that we don’t have any evidence about. The issue is whether we are even testing those that we do know about sufficiently.
Seroprevalence surveys among humans, poultry, wild birds, cats, pigs, would be my preference, in order of importance.
bump
see here for a list of H5N1 viruses, sorted by host and amount of
A,T in its nucleotides (except HA,NA,NS,M only first half,NP included twice
this gives high correlation of 0.863 with host)
which serves as a measure how human-adapted the virus is.
(and thus, how well it spreads to humans ?)
We can clearly see how the Qinghai and the Java strains are better adapted
to humans than the Vietnam strain
http://www.setbb.com/fluwiki2/viewtopic.php?t=132&mforum=fluwiki2
Top scorers are:
A 5496 H5N1 2006 A/chicken/Nigeria/641/2006(H5N1)
A 5498 H5N1 2006 A/chicken/Cote d’Ivoire/1787–34/2006(H5N1)
A 5499 H5N1 2002 A/chicken/Jiangsu/cz1/2002(H5N1)
H 5499 H5N1 2006 A/Indonesia/CDC699/2006(H5N1)
S 5498 H5N1 2004 A/swine/Guangxi/wz/2004(H5N1)
S 5503 H5N1 2004 A/swine/Anhui/ca/2004(H5N1)
with the suspicion that these last swine viruses went through several
generations of mammals (swine?) before the sample was taken.
__________________
ask experts for their panflu-probability estimates,
and report their responses here !
Let’s merge all the bird flu forums !