1/18
3rd Lecture
Name | Mastery | Learn | Test | Matching | Spaced | Call with Kai |
|---|
No analytics yet
Send a link to your students to track their progress
There are three approaches to inferring phylogenies, what are they?
Phenetics, Cladistics, & Model-based approaches.
What are phenetics?
An approach that attempts to infer relationships by classifying organisms based on overall similarity
What are cladistics?
An approach that attempts to infer relationships by classifying organisms based only on shared derived characteristics; better than phenetics
What are model-based approaches?
An approach that attempts to infer relationships, allowing for possibility that not all character state changes are equally informative
Phenetics: What are the three steps?
Find most similar two taxa, link them together and treat them as a group
Find next most similar two taxa, link them, and treat them as a group
Continue until all species have been linked
use character state matrix to draw a neighbor-joining tree of the taxa
Phenetics: True or false. Phenetics reflects the true evolutionary history of the group
False.
Cladistics: What is the principle of parsimony?
The principle of parsimony- All else being equal, the explanation requiring the fewest number of steps is accepted
Cladistics: How do you map traits with parsimony?
Show all hypothesized character state transitions
Minimize number of hypothesized changes
Cladistics: What are the 5 steps to conduct a cladistic analysis?
Consider all possible unrooted trees
Map all character state transitions onto tree
Tree with fewest transitions is accepted as best hypothesis of descent relationships
Equivalent statement: the tree with the fewest apparent homoplasies is accepted.
Root tree using outgroup
Advantages of phenetics?
Simple
Doesn’t need an outgroup
Advantages of cladistics?
Most likely to give the true phylogeny
Problems with parsimony: What are the three general problems with parsimony?
Parsimony minimizes the number of character state transitions
Assumes that all character state transitions are equally likely
Between each possible character
Along each branch of a tree
These assumptions are often not valid
Pyrimidines may be replaced by pyrimidines more often than purines
Some lineages evolve more quickly than others
Problems with parsimony: long-branch attraction
Branches with many autapomorphies will be similar to other long branches by chance alone
Long branches may also attract short branches, but small changes in the data may switch which short branch is attracted to
What is maximum likelihood? What does it not assume?
Maximum likelihood is a model-based approach to tree building. It does not assume equal probabilities of character state changes among all character states.
It allows different probabilities of character state changes among all character states
It accounts for different rates of evolution in different lineages.
Maximum likelihood: Advantages
Takes into account varying rates
May result in more accurate outcomes
Reduces long branch attraction
Maximum likelihood: Disadvantages
Computationally intensive compared to parsimony
Estimates of probabilities of different kinds of change may be inaccurate
True or false? Homoplasies can make a big difference to patterns.
True
True or false? Homoplasy is more likely to affect results
False. Homoplasy is less likely to affect results.
Bootstrapping steps
Make a new data set by randomly choosing loci from data set
Some are sampled repeatedly; some are omitted
Make new tree & see if patterns are the same
Clades that appear repeatedly are considered strongly supported; those that appear only occasionally are not considered to be reliable
Similar to P values, but higher numbers are better
Must include all members and only members of the clade
Relationship among the members in the bootstrapped clade is irrelevant