3020-Exercise #3

C3020 — Molecular Evolution

Exercises #3: Phylogenetics

Consider the following sequences for five taxa 1-5 and the known outgroup O, which has the ancestral states

1 ACAAACAGTT CGATCGATTT GCAGTCTGGG

2 ACAAACAGTT TCTAGCGATT GCAGTCAGGG

3 ACAGACAGTT CGATCGATTT GCAGTCTCGG

4 ACTGACAGTT CGATCGATTT GCAGTCAGAG

5 ATTGACAGTT CGATCGATTT GCAGTCAGGA

O TTTGACAGTT CGATCGATTT GCAGTCAGGG

1. Make a distance matrix using raw distances (number of differences) for the five ingroup sequences.

2. Infer the UPGMA tree for these sequences from your matrix. Label the branches with their lengths.

3. Using the parsimony criterion,

b. How many nucleotide changes (steps) does the tree rooted tree (((((1,2),3),4),5),0) require? How many homoplasious changes does it require? Mark the number of changes that occur on each branch.

c. How many nucleotide changes does the tree ((((1,4),(2,3)),5),0) require? How many of these are homoplasious?

d. Which one of these two trees is "better"? Why?

e. For what clade of taxa does a C in the third nucleotide position represent a synapomorphy?

f. Name two nucleotide positions that contain a symplesiomorphic state for the clade (1,2,3).

4. Using neighbor-joining,

b. On this tree, what are the branch lengths for the branches leading from the last common ancestor (LCA) of 1 and 2 to taxon 1 and to taxon 2?

c. Estimate S on the unrooted tree in which only taxa 1 and 3 are joined as neighbors. What are the lengths for the branches leading to 1 and to 3 from their LCA?

d. Which of these two trees is better? What do you do with next with the better one?

e. Estimate S for the tree ((1,2),3),4,5) -- a possible "next step" tree from the tree in question 4a.

f. Estimate S for the tree ((1,2),4),5,3) -- another possibility for the next step.

g. Which of these last two trees is better?

h. Calculate all the individual branch lengths on this tree.

i. Draw the rooted tree for the tree you chose as better, assuming that the outgroup attaches to the branch leading to taxon 5 and only taxon 5..

5. You now have 3 trees -- a UPGMA tree (problem #2), a parsimony tree (#3d), and a NJ tree (#4i). (Although you have not examined all possible phylogenies for either parsimony or NJ, assume that the best tree from question 3 is the most parsimonious tree and the best tree from question 4 is the best NJ tree. These are, in fact, the optimal tree for each method.)

b. If there are differences, explain what might have caused the incongruence you observe. Which tree(s) do you find most convincing?

6. Morphological data do not resolve the relationships among the mammalian orders primates, artiodactyls, and rodents. You would like to use molecular data to establish whcih lineage diverged first from the others. In your laboratory, your research assistant obtains sequences from a cow, a human, and a mouse for three genes: psi-n-globin (a pseudogene of globin, the oxygen-transporting protein in blood), histone A1 (one of the proteins that packs DNA in chromatin), and 18S ribosomal RNA. You use parsimony for a phylogenetic analysis.

b. What other information do you absolutely need to resolve the relationships among these taxa?

c. Suppose the gene you chose gives the phylogeny ((primate,rodent),artiodactyl), and the node supporting a primate-mouse clade has a bootstrap value of 55 and a Bremer support of 2.

What does a bootstrap value of 55 mean? -- explain in a sentence or two. What does a Bremer support of 2 mean?
How confident are you that you have the true phylogeny? Justify your answer.
If this is not the true tree, what are some reasons you might have gotten this tree anyway? What should you do to address these problems?

d. Suppose now that the gene you chose gives the same phylogeny as above, but the node supporting a primate-mouse clade has a bootstrap value of 95 and a Bremer support of 9. Do you have more or less confidence in your phylogeny? Why? Is it possible still that you do not have the correct taxonomic phylogeny? If so, why might you have gotten this tree?