Announcement

Collapse
No announcement yet.

(almost) no recombination in flu-B

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    Re: (almost) no recombination in flu-B

    Originally posted by Hanshilinzuo View Post
    It can be explained by random mutation. However, I think it is a good example to detect positive selection in influenza virus.
    No it can't. G743A is silent and appeared on multiple genetic backgrounds.

    Each of these backgrounds picked up 2-6 changes and one was G743A and they all happened at the same time. Random mutation / selection doesn't explain this, which is why I use it as an example. Most of the time the data is not quite this dramatic, but I use G743A to eliminate the error/selection argument (but as noted above, it is still cited).

    The random muttaion / selection mantra also gets very old very fast.

    Comment


    • #17
      Re: (almost) no recombination in flu-B

      Originally posted by niman View Post
      The recombination can be easily missed, because it happens many times and the result is usually a SNP or two (which you then call "random mutations"), because the most common recombination is between closely related sequences.
      However, there are no acceptable model in published peer-review materials. It is impossible to detect recombination between very similar sequences, though there are thousands of handreds of recombination detection methods.

      Comment


      • #18
        Re: (almost) no recombination in flu-B

        Originally posted by Hanshilinzuo View Post
        However, there are no acceptable model in published peer-review materials. It is impossible to detect recombination between very similar sequences, though there are thousands of handreds of recombination detection methods.
        G743A is an example of the "impossible".

        Comment


        • #19
          Re: (almost) no recombination in flu-B

          Could you post some sequences of this example. Then I can have a quick look at the data.

          Comment


          • #20
            Re: (almost) no recombination in flu-B

            maybe some constraint makes G743A more likely, so it appears
            independently on similar backgrounds

            or mixed viruses, some with G743, some with A743 in the same
            droplet


            why SNPs but no double-NPs , triple... via recombination ?
            I'm interested in expert panflu damage estimates
            my current links: http://bit.ly/hFI7H ILI-charts: http://bit.ly/CcRgT

            Comment


            • #21
              Re: (almost) no recombination in flu-B

              Originally posted by Hanshilinzuo View Post
              Could you post some sequences of this example. Then I can have a quick look at the data.
              The nature preceding article(s) can be accessed at no charge and has the sequences (just clik on PDF icon).

              Comment


              • #22
                Re: (almost) no recombination in flu-B

                Originally posted by niman View Post
                The nature preceding article(s) can be accessed at no charge and has the sequences (just clik on PDF icon).
                I will check it later. Due to my experiences in the evolution of viruses, I do not think your idea is right. Anyway you should publish your paper in some peer-viewed journals. It is very important for a research. Try it

                Comment


                • #23
                  Re: (almost) no recombination in flu-B

                  Originally posted by Hanshilinzuo View Post
                  I will check it later. Due to my experiences in the evolution of viruses, I do not think your idea is right. Anyway you should publish your paper in some peer-viewed journals. It is very important for a research. Try it
                  There will be several papers which will address some of the concerns of lab error or vaccine contamination. The recombination, especially for the SNPs, represents a paradigm shift which extends well beyond influenza.

                  I will post a brief history of G743A, so you can see why it is not due to copy errors / selection. The G743A data is quite compelling.

                  Comment


                  • #24
                    Re: (almost) no recombination in flu-B

                    Originally posted by Hanshilinzuo View Post
                    I will check it later. Due to my experiences in the evolution of viruses, I do not think your idea is right. Anyway you should publish your paper in some peer-viewed journals. It is very important for a research. Try it
                    SNPs as acquistions via recombination is a paradigm shift, so you are not alone in your thinking that the idea is not right. However, as more sequences come out, and the examples increase, it will be quite obious that the SNPs that appear at the same time on different backgrounds is a not coincidence or matches due to heavy selection of random errors.

                    Speaking of which, 3 sets of HA and NA sequences are being released from Hong Kong which have H274Y on NA. These sequences were just deposited at Genbank and will be accessible through the front door in a day or two. They are only partials, but the data are quite clear.

                    Most of the oseltamivir resistance recently reported has been on the Brisbane/59 (clade 2B) background. However, this change has also been seen previous on a clade 2C (Hong Kong/2652) background (in China).

                    All three HA sequences are clade 2C, while only 2 of the NA sequneces is clade 2C (the other is an exact match of clade 2B). All three have H274Y and the one with a clade 2B sequence is a reassortant.

                    Of the two isolates that are not reassortants, one is an exact match of previously released sequences from the US and Japan. The other is similar, but has 2 changes. One of the 2 changes encodes N187S, which has also been acquired by 5 of the 8 HA sequences recently released from South Africa (which are clade 2B)



                    Thus, this clade 2B change in South Africa is on a clade 2C background in Hong Kong. These numbers are low, so most will simply wave their hands and say the sharing of N187S between new clade 2B isolates in South Africa and a clade 2C isolate in Hong Kong is a coincidence, but it is not. It is recombination and the fact that both acquired the same change at the same time is due to a common source and the acquisition is via homologous recombination.

                    I will write up a commentary with specifics and links to sequences, which are posted here

                    Comment


                    • #25
                      Re: (almost) no recombination in flu-B

                      Here are the six isolates that have recently (see collection dates) acquired HA N187S (H3 numbering). All have H274Y in NA.

                      Clade 2C
                      A/HongKong/1052/2008 Apr-2008

                      Clade 2B
                      A/Johannesburg/46/2008 11-Jun-2008
                      A/Johannesburg/35/2008 09-Jun-2008
                      A/Johannesburg/34/2008 09-Jun-2008
                      A/Johannesburg/25/2008 03-Jun-2008
                      A/Johannesburg/10/2008 15-May-2008

                      Comment


                      • #26
                        Re: (almost) no recombination in flu-B

                        Originally posted by gsgs View Post
                        maybe some constraint makes G743A more likely, so it appears
                        independently on similar backgrounds

                        or mixed viruses, some with G743, some with A743 in the same
                        droplet


                        why SNPs but no double-NPs , triple... via recombination ?
                        As I said earlier, the first series in Egypt were cloned. There were about 30 clones with the Gharbiya sequence and 10 with the earlier bird sequence. ALL 40 clones had G743A (which was absent in early related Egyptian sequences).

                        G743A was in clade 2.2 in 2006, but all but one were on the same general sequence found in southern Germany, Switzerland, France, and the Czech Republic.

                        None of the 2007 sequences with G743A matched the European sequences yet they appeared on different backgrounds in Russia, Egypt, Kuwait, Ghana, and Nigeria (and subsequently spread throughout Europe on the Uvs Lake background - and probably now are in west Africa, which also now has Uvs Lake).

                        There are doubles and triples, but SNPs are much more common because most recombination is between closely related sequences).
                        Last edited by HenryN; October 1, 2008, 12:05 PM. Reason: typos

                        Comment


                        • #27
                          Re: (almost) no recombination in flu-B

                          Originally posted by Hanshilinzuo View Post
                          I will check it later. Due to my experiences in the evolution of viruses, I do not think your idea is right. Anyway you should publish your paper in some peer-viewed journals. It is very important for a research. Try it
                          Here is a quick run-down on G743A, which should make it pretty clear that the acquistions are not due to selection of random errors.

                          In 2006, G743A on clade 2.2 was concentrated in Europe and on the same clade 2.2.2 background. There were something like 40 isolates from southern Germany, Switzerland, France and the Czech Republic. The only other published clade 2.2 sequence with G743A was in west Africa (one sequence out of many dozen).

                          I was working with NAMRU-3 on H5N1 in Egypt right after a cluster of 3 fatal infections were found in family members in Gharbiya. The NA sequences had N294S, which confered oseltamivir resistance. Since the resistance was in isolates collected prior to oseltamivir treatment, the frequncy of N294S in birds was of interest.

                          In February three isolates were received from chickens in Gharbiya. One of the three was closely related to the human sequences, but didn't have N294S. However, it was clear that the isolate was a mixture, so it was plaque pururified to see if a minor species had N294S. Over 40 clones were isolated and the NA's were sequenced. 30 of the clones were closely related to the human sequences, but they did not have N294S. However, they did have G743A, which was not in the human sequences. The other 10 clones matched the other two chicken sequences, which were related to Egyptian H5N1 sequences from 2006. Once again the 2006 sequences did not have G743A, but the 2007 sequences (from the clones or the two independent isolates) did.

                          Thus, in Gharbiya there were two distinct H5N1 sequences that had acquired the same polymrphism, G743A. The selection for this was not obvious, since the change was silent, and no Egyptian (or Israel, which was similar) sequence had G743A in 2006, but it appeared on two backgrounds at the same time in 2007. Thus, since only two backgrounds were involved, the acquisitions could have been a coincidence, although the number of differences between the 2006 and 2007 sequences was small (2-6 changes), so the chance of a coincidence would have been small.

                          However, shortly thereafter isolates in southern Egypt picked up the change. Once again earlier versions of the southern sequences had been found in patients and the earlier sequences did not have G743A. These sequences were quite different than those in the Nile Delta, yet they picked up the same change. This happened again, so there were four distinct H5N1 in Egypt that had picked up the same change, which would have been VERY hard to explain by random error / selection mechanism.

                          However, these "coincidences" did not stop in Egypt. In Feb, 2007 there was also an H5N1 outbreak in Moscow. The sequences were similar to clade 2.2.3 isolates from Azerbaijan in 2006, but the Moscow sequences had also acquired G743A.

                          NAMRU-3 also had a facility in Ghana, which also reported H5N1 for the first time in the spring of 2007. The sequences were closely related to H5N1 from the Ivory Coast (in 2006), but the Ghana sequences had also acquired G743A.

                          Thus, there were six different backgrounds in three different countries that had acquired the same change at the same time. However, as additional sequences were released, more examples were revealed.

                          Kuwait also had an outbreak in Feb, 2007. The sequences were also clade 2.2.3, but were distinct from the sequences in Moscow. The Kuwait sequences were closley related to clade 2.2.3 from Uvs Lake and Russia collected during a wild bird outbreak in the summer of 2006. Similar sequences were in South Korea and Japan at the end of 2006 and none of the 2006 sequences had G743A, but all of the sequences in Kuwait did. These Uvs Lake sequences in Kuwait subsequently spread throuout Europe in the summer and fall, and all Uvs Lake sequences published to date have G743A.

                          Similarly, there was an H5N1 outbreak in Nigeria in early 2007, which were similar to the first human case there. A small subset of the Nigerian sequences also had G743A.

                          Thus, the G743A in clade 2.2 which was largely limited to a small geographic region in Europe on a clade 2.2.2 background in 2006, suddenly appeared on multiple genetic backgrounds (including multple versions of clade 2.2.1 and clade 2.2.3) in early 2007, and the 2006 precursors of each of these sub-clades did not have G743A.

                          Trying to explain the above acquistions of the G743A SNP via selection of a random mutation is a difficult task, especially since the 2006 precusors for each background had already been isolated and the 2006 sequences did not have G743A.

                          Although the appearance of the same SNP on so many different backgrounds at the same time is unusual, the sharing of the same change by isolates with a smaller number of genetic backgrounds is common, and these acquisition of SNPs via homologous recombination is the MAIN driver of influenza evolution (and accounts for the VAST majority of SNPs).

                          Comment


                          • #28
                            Re: (almost) no recombination in flu-B

                            trying to explain this with recombination is also hard.

                            Why did recombination always pick exactly the same polymorphism ?
                            Why are there no other such examples ?

                            How did the viruses "travel" from one place to another one
                            and were not isolated there, while the "recombinants" were ?

                            Maybe it could help to look at 3d-pictures of N1 to detect some
                            exposure or constraint of position 743 ?
                            I'm interested in expert panflu damage estimates
                            my current links: http://bit.ly/hFI7H ILI-charts: http://bit.ly/CcRgT

                            Comment


                            • #29
                              Re: (almost) no recombination in flu-B

                              Originally posted by gsgs View Post
                              trying to explain this with recombination is also hard.

                              Why did recombination always pick exactly the same polymorphism ?
                              Why are there no other such examples ?

                              How did the viruses "travel" from one place to another one
                              and were not isolated there, while the "recombinants" were ?

                              Maybe it could help to look at 3d-pictures of N1 to detect some
                              exposure or constraint of position 743 ?
                              Please. I have already listed clade 2.2 in 2006 that had G743A. There are MANY other examples, just not quite as dramatic. It traveled in the Uvs Lake strain. G743A is SILENT. 3-D of WHAT, the RNA?

                              Comment

                              Working...
                              X