GOALS:

·      Learn how to get your data into TREE-PUZZLE and the results out.

·      Attain some understanding on what to do with the ml ratio test.

·      Think about models for the substitution process.

·      Learn about the different options in TREE-PUZZLE.

General Remarks:

If no file called infile is present in the program directory, TREE-PUZZLE will prompt you for the infile.  If you work a lot on the same dataset, it might be worth to rename your data set into infile.  The output is always put into the files outfile, outtree and outdist.

You have to rename these files as soon as you are done.  The next time TREE-PUZZLE starts, the files get erased (this is particularly sad if your run took a couple of days).

If you start TREE-PUZZLE from the command line (MS-DOS prompt, or c-shell) using the command "puzzle name_of_infile.phy" the program writes the results into a file called name_of_infile.puzzle, name_of_infile.dist, etc.

The following assumes that you read the notes regarding TREE-PUZZLE from class 10. If you have not done so, do it now (here are the sections on puzzle trees, the ml-ratio test and ml-mapping).

 Open this page in NETSCAPE and click HERE. When prompted save the program TREE-PUZZLE.exe into your work folder.

  Transfer the files testseq5.txt, testseq5.aln and testseq5.phb onto your computer.  These files contains vacuolar/archaeal ATPases from 49 pro- and eukaryotes: 

Daucus carota, Arabidopsis thaliana, Gossypium hirsutum are plants;

Acetabularia acetabulum is a green and Cyanidium caldarium is a red algae

Mus, Homo, Bos (mammals) Gallus (bird), Drosophila, Aedes (insects), Ascidia (tunicate, chordate) are animals

Saccharomyces, Candida, Schizosaccharomyces, Eremothecium and Neurospora are fungi

Dictyostelium discoideum, Entamoeba, Plasmodium falsiparum, Trypanosoma, Nosema, and Giardia are protists or protozoa

Sulfolobus acidocaldarius (70 °C), Sulfolobus solfataricus (70-85 °C), Archaeoglobus fulgidus (83 °C), Methanosarcina barkeri (30-37 °C), Methanosarcina mazeii (37°C), Methanococcus jannaschii (80°C), Haloferax volcanii (37°C), Halobacterium salinarium (ca37°C), Methanobacterium thermoautotrophicum (60-65°C), Desulfurococcus sp. (85-90 °C), Thermococcus sp (75+ °C), Aeropyrum pernix (90°C), Thermoplasma acidophilum (55-60°C), Pyrococcus abysii and Pyrococcus horikoshii 95°C (archaea),

Enterococcus hirae (37°C), Borrelia burgdorferi (33-37°C), Thermus thermophilus (70-80°C), Deinococcus radiodurans (30°C), Chlamydia trachomatis, Chlamydophila pneumoniae (Bacteria) are prokaryotes. (Usually Bacteria have an F- and not an A-ATPase. The bacteria probably obtained the archaeal/vacuolar type ATPase through horizontal gene transfer.) 

Note the German Collection of Microorganisms and Cell Cultures is a good place to find growth temperatures for microorganisms. 

The prokaryotes can be considered as outgroup for the eukaryotes.

   Use maximum likelihood mapping to address the following question: Does the Giardia or the Trichomonas sequence represent the deeper branching lineage among the eukaryotic homologues?  
As we are not sure about the relative branching order of Trypanosoma and Plasmodium, do not assign them to any group. 

For this you need to:

         load testseq5.aln into clustalx,

         save the sequences (they should be already aligned) in phylip format.  Check that truncating the sequence names does not result in duplicate sequence names. 

         Start TREE-PUZZLE by double clicking on the TREE-PUZZLE icon

         Load testseq5.phy (or what ever name you gave the file)

         Select type of analyses: maximum likelihood mapping (“b”)

         Select sequences in 4 clusters (“g”)

         Assign sequences to the 4 groups (i.e. the 4 groups attached to the branch you want to study).  Ask if you cannot figure it out.

         Select number of quartets (“n”) enter “0” to select all (as two of your clusters contain a single sequence only, this number is quite small)

         Select the model of heterogeneity.  (select Gamma distribution with 8 classes, enter the shape parameter as 0.6.  DO NOT have the program find the shape parameter (takes about 30 minutes).

         Interpret the outfiles. 

        What is your conclusion? (Which sequence is deeper branch? Is the support for your finding strong, week, or nonexistent?)

    Select one of the following two problems to address using the maximum likelihood ratio test: 

A)    Do the V-ATPase A subunits from the red and green algae form a monophyletic group with the higher plants?

  For this you

         Delete all prokaryotic sequences from testseq5.aln using clustalx

         Save as phylip file (testseq5b.phy)

         Use the treeview editor to build the appropriate trees to test.  A good starting tree is this one [it is a neighbor joining tree calculated from distances that were estimated using among site rate variation (gamma distribution with alpha=.6) and the JTT substitution model.  I then collapsed all branches that in a maximum likelihood reconstruction were not at least 2.5 standard deviations larger than zero.]  You can load several trees into TREE-PUZZLE simultaneously.  If you do so, TREE-PUZZLE also performs a Kishino-Hasegawa test to determine which of the trees is significantly better than the other ones.  An example for a file with multiple user trees is here. You can use treeview to edit your tree to generate the different starting trees.  You can copy directly from treeview and paste into a text editor.  The only problem is that treeview adds ‘ ‘ to sequences that start with a number e.g. ‘1Acetabularia’.  You need to remove the ‘s before you can read the trees into TREE-PUZZLE! 

         Save the trees (multiple trees should be separated by ; )you want to test in a textfile in the same directory as the TREE-PUZZLE program

         Start TREE-PUZZLE

         Load testseq5b.phy 

         Select tree search procedure and toggle to “User defined trees”

         Select model of heterogeneity (in the interest of time you might leave it at the default setting)

         Calculate branch lengths and likelihood values for the different trees. 

         Use the CHI square distribution (see handout, or Paul's Chi square calculator here, or here) to determine, if the increase in likelihood (2(logL₁-logL₂)) is significant. 

B)    Does the incorporation of among site rate variation lead to a significant increase in likelihood?

For this you

   Delete all prokaryotic sequences from testseq5.aln using clustalx

   Save as phylip file (testseq5b.phy)

   Save this treefile (see above) in your TREE-PUZZLE directory

   Start TREE-PUZZLE

   Load testseq5b.phy 

   Select tree search procedure and toggle to “User defined trees”

   Select the default model of heterogeneity

   start the program

   when prompted enter name of treefile

   When done, save outfile under a different name!

   run TREE-PUZZLE again, but select a different model for substitution (Gamma with 8 categories and alpha=.6)

   Use the CHI square distribution to determine if the increase in likelihood between the two runs (2(logL₁-logL₂)) is significant. 

Does the incorporation of ASRV into the model lead to significant better description of the data?

  If you have time left: