Pacifica (autumnwinds) wrote,
Pacifica
autumnwinds

  • Mood:
I've had a great couple of days and had to share.

A. Pam is planning a bridal shower for me in two weeks. :D

B. WE GOT THE APARTMENT.

C. Got kudos for an idea I had (about estimating Alnus phylogenies with chloroplast DNA), from someone whose opinion I respect.

D. Same person asked me for MY advice on an ecological issue.

E. Got kudos from advisor for doing the research to figure out a sneaky way to order gel stain while avoiding $40 shipping fees.

F. Had a neat conversation with my advisor about orchids and he pledged to give me one of his.

G. Got an A on my test!



1. Describe the KH test of two topologies with RELL approximation (12 pts.). Include the null hypotheses, the test statistic, and null distribution.

2. Let's assume you have an alignment of rRNA sequences that you wish to analyze. Further, you wish to use a separate HKY + I model for stems and loops. Parameterize the substitution model (Hint: this is like parameterizing a GTR + SSR model; 22 pts.)

3. We discussed three methods for selecting a likelihood model for phylogeny estimation on the basis of relative goodness of fit. What are these (3 pts each). Which one would you prefer to use? Justify your choice (10 pts.).

4. Bayesian estimation of any parameter requires the explicit quantification of prior knowledge in the form of prior probabilities. What does this mean in terms of estimating tree topology? Ignore all other parameters (10 pts.).

5. Let's assume you have estimated the following topology: (tree).
The node indicated by the arrow has a nonparametric bootstrap value of 98, a decay index of 42 and a posterior probability of 1.0. Interpret each of these measures of nodal support (12 pts.).

6. Some authors have written that no single method of phylogeny estimation has been shown to outperform all others across all conditions. Perhaps this perception has to do with relationship between the shape of the underlying tree and properties of various methods. Discuss this as it relates to parsimony and likelihood methods and bias, consistency and efficiency (25 pts.).
  • Post a new comment

    Error

    Anonymous comments are disabled in this journal

    default userpic
  • 1 comment