Announcement

Collapse
No announcement yet.

Only for Tree lovers

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

  • Only for Tree lovers

    Hello,

    For tree lovers, here is some interesting tree I produce with the BioEdit-Tree view interface.

    To be valid, a tree must:

    - be produce with a fully aligned set of sequences;
    - sequences must be cut to the same lenght (if it is not, the tree will only classify the sequence by the lenght of the fragment);
    - tree must be "rooted" with the oldest or the most different strain;

    Also:

    - some trees do not represent genetic distance;
    - some do;
    - remove all () because the device will "bug"
    ________________

    Use the "DNAdist Neigbor phylo tree" with a DNA/RNA alignement to compare genetic distance;
    Use the "ProtDist Neigbor phylo tree" with a protein alignement to compare the virus antigenicity;

    Have fun

    Will follow some examples

  • #2
    Re: Only for Tree lovers

    Notes that theses Tree haved been produce with the HA gene of the hemmaglutinnin;


    Here is a tree that only show sub class;
    Click image for larger version

Name:	tree unrooted.bmp
Views:	1
Size:	501.1 KB
ID:	647715

    Here is a tree that do show genetic distance
    and is rooted with the oldest seq;
    Click image for larger version

Name:	tree rooted.bmp
Views:	1
Size:	502.6 KB
ID:	647714
    Note: Above, the line with 0.1 under represent the 10% distance in the branch lenght.

    Here is also another possible view to show distance between strains;
    Click image for larger version

Name:	tree pie.bmp
Views:	1
Size:	725.7 KB
ID:	647713
    Last edited by Mingus; August 8, 2006, 10:02 AM.

    Comment


    • #3
      Re: Only for Tree lovers

      Mingus - Thank you very much.

      Comment


      • #4
        Re: Only for Tree lovers

        Trees are produce matematically and only represent relationship between strains.

        They represent a ideal situation like if strains would haved evolves linearly.
        They do not take recombination in account. To do that, branch should be added between branch to represent the individuals polymorphism and that kind of tree do not exist.

        So they are usefull tools but are to take with a pinch of doubt always.

        Comment


        • #5
          Re: Only for Tree lovers

          How can you check the bootstrapvalues? I believe with clustalw there is a way?

          Thank you

          Comment


          • #6
            Re: Only for Tree lovers

            Originally posted by Mingus
            Trees are produce matematically and only represent relationship between strains.

            They represent a ideal situation like if strains would haved evolves linearly.
            They do not take recombination in account. To do that, branch should be added between branch to represent the individuals polymorphism and that kind of tree do not exist.

            So they are usefull tools but are to take with a pinch of doubt always.
            You make an extremely valid and important point: Models, such as tree diagrams, have an underlying set of assumptions, and will tend to ignore any data that does not fit those assumptions. This model intrinsically assumes that all mutation is random. Thus it is biased against recombination. You would need a different kind of model to chase the polymorphisms and demonstrate recombination, although you can sometime see it in the trees when there are discordant ages in the nearest match.

            Comment

            Working...
            X