Trees - maximum parsimony

Question 1

What is Occam’s razor

Answer 1

an operational principle in logic and problem solving where the hypothesis with the fewest assumptions is preferred

–> With this parsimony is the ideal where the best tree is the shortest

Question 2

Describe homoplasy in terms of parsimony analysis

Answer 2

with homoplasy there is the existence of more than the minimum number of steps for one or more characters - characters can undergo parallelism, reversal or convergence

Question 3

What does the parsimony principle state

Answer 3

with other things being equal, the preferred tree is the one with the least homoplasy (the fewest possible changes of character state)

Question 4

What is the optimality criterion for Occam’s razor

Answer 4

maximum parsimony (MP)

Question 5

What is the method for finding the best parsimonious tree

Answer 5

• tree space is searched for the shortest trees
• each tree is evaluated for length or total changes
• the most commonly used method for optimizing character changes onto a tree to determine its length is that of Fitch

Question 6

How does the Fitch method work for finding the best parsimonious tree

Answer 6

• assume all changes between all states of character are possible and all count equally
• examine the interior nodes of the tree in sequence from tips to root (downward pass) and then from root to tip (upward pass)
• the states at each node are the intersection of the states at adjacent nodes if the sets of possible states overlep
• if they do not overlap the states at the node are the union of the states at adjacent nodes

Question 7

Describe ACCTRAN optimization

Answer 7

this places the initial character change as close to the root as possible, often leading to the hypothesis of reversal later

Question 8

Describe DELTRAN optimization

Answer 8

places the initial character change as close to the terminal tips as possible, often leading the hypothesis of convergent or parallel evolution

Question 9

What is a problem with the maximum parsimony method

Answer 9

1. long branch attraction
2. constraints on tree searching

Question 10

How can long branch attraction be minimized under parsimony criterion

Answer 10

1. Adding more data does not usually help, unless their rates of evolution are extremely different
2. A better approach is to add taxa, especially those with intermediate phylogenetic positions or those that are basal within their clades

Question 11

can molecular studies use fossils

Answer 11

no

Question 12

Describe long branch attraction

Answer 12

(here distantly related lineages are incorrectly inferred to be closely related) this is a cause of systematic error particularly troublesome in molecular systematics where there are a large number of characters, limited numbers of states, rates of evolution may be variable and high, and homoplasy is common

Question 13

Describe constraints on tree searching when it comes to finding prisonous trees

Answer 13

• for studies with modestly large numbers of taxa, there is an astronomical number of trees to be evaluated

Question 14

Describe an exhaustive search

Answer 14

• Every possible tree is evaluated,
• this gives certainty that the best tree has been found
• the shape of the tree landscape is known
• used for n < ~14 taxa

Question 15

Describe implicit enumeration search

Answer 15

• omits evaluation of set of trees that can be ruled out as the possible shortest trees
• gives certainty that the best tree has been found
• shape of the tree landscape is not known
• possible for n < ~18 taxa

Question 16

Describe traditional search / heuristic search

Answer 16

• includes different methods of sampling the tree landscape
• a starting tree is obtained, then branches are swapped to find a better tree
• used for a larger number of taxa
• no guarantee of finding the best tree
• tree landscape sampled by not completely known

Question 17

What are the 3 branch swapping algorithms (tree search algorithms)

Answer 17

1. nearest neighbor interchange (NNI) (least costly in computational time but least likely to find shorter tree)
2. Subtree pruning and regrafting (SPR) - this tries all possible subtrees regrafted on all possible branches
3. Tree bisection and reconnection (TBR) - breaks all internal branches and reconnects all possible pairs of branches from the two subtrees obtained at each round (this is most costly in terms of computational time but most likely to find shorter solution) - widely used

Question 18

Describe subtree pruning and regrafting (SPR)

Answer 18

this is a tree searching algorithm used for maximum parsimony where it tries all possible subtrees regrafted on all possible branches

Question 19

Describe tree bisection and reconnection (TBR)

Answer 19

this is a tree searching algorithm used for maximum parsimony which breaks all internal branches and reconnects all possible pairs of branches obtained from the two subtrees at each round