Notes
Slide Show
Outline
1
Distance Based Phylogeny
  • Leaves are events
  • Labels are evolutionary distances
2
Distance Based
  • Form a distance matrix such that each row and column corresponds to a species and the entries are the distances between
  • Find a tree such that the distance between each pair of leaves is equal to the distance between these species given in the distance matrix (additive tree ®polynomial solution)
  • The distance matrix must be additive
3
 
4
Solutions
  • These conditions usually do not hold- problem becomes  hard.  As a consequence, we invoke a heuristic.  We will use Unweighted Pair Group Method with Arithmetic Mean (UPGMA)
  • UPGMA
    • Bad news- Like any heuristic, it does not always produce the right answer
    • Good news- It always produces an ultrametic tree

5
Unweighted Pair Group  Method with Arithmetic Mean
UPGMA
  • Choose the pair X,Y with the smallest distance
  • Combine X and Y into a singleton, XY
  • The distance between any other point and the compound species XY is the arithmetic mean of the pairwise distances dij where i is a component species of X and j is a component species of Y.
  • Restated, the distance from k to xy is (d(k,x)+d(k,y))/2
6
Unweighted Pair Group  Method with Arithmetic Mean
UPGMA
  • The matrix is now one order smaller
  • Repeat the procedure until only two compound species remain.
7
UPGMA

  • Build an evolutionary tree such that from the point of branching between two (compound) species X and Y, the distance to each of their component species is dij/2.
8
Unweighted Pair Group  Method with Arithmetic Mean
Demonstration of UPGMA
Distance Matrix:
9
UPGMA
10
 
11
Continue Refining the Distance Matrix
12
Next Refinement