Multiple Alignment and Phylogenetic Trees

1
Multiple Alignment and Phylogenetic Trees

• Csc 487/687 Computing for Bioinformatics

2
Terms related to graphs

• Graph ? an abstract structure containing nodes
  and edges
• Path ? a consecutive set of edges
• Tree ? a connected graph with exactly one path
  between every two nodes
• Edge length ?a number assigned to each edge
  signifying in some sense the distance between the
  nodes connected by edge
• Path length ? the sum of the lengths of the edges
  that comprise the path

3
Deriving phylogenetic trees

• Phenetic approach
• Measuring a set of distances between species, and
  generate the tree by a hierarchical clustering
  procedure.
• Cladistic approach
• Consider possible pathways of evolution
• Infer the features of the ancestor at each node
• Choose an optimal tree according to some model of
  evolutionary changes.
• Genealogy

4
Phenetic (clustering) approach

• a simple clustering procedure
• Determine a measure of the similarity or
  difference for all pairs
• Create a tree from the set of dissimilarity
  iteratively
• First choose two closely-related species, and
  insert a node to represent their common ancestor.
• Replace the two selected species by a set
  containing both
• Replace the distances from the pair to the others
  by the average distance of the two selected
  species to the others.

5
Phenetic (clustering) approach

• Example
• UPGMA method (Unweighted pair group method with
  arithmetic mean)

6
Cladistic methods

• The goal is to select the correct tree by
  utilizing an explicit model of the evolutionary
  process.
• Maximum parsimony
• Maximum likelihood
• They are specialized to sequence data, starting
  from a multiple sequence alignment

7
Maximum parsimony method

• An optimal tree as the one that introduces the
  fewest mutations.
• Several trees may postulate the same number of
  mutations
• Not unique answer

8
Maximum likelihood method

• Assign quantitative probabilities to mutational
  events
• The optimal tree is the one with the highest
  likelihood of generating the observed data

9
The problem of varying rates of evolution

• Subject to the errors if the rates of
  evolutionary change vary along different branches
  of the tree
• outgroup?a species more distantly related to all
  the species than any pair of them is to each
  other.

10
Computational Consideration

• Cladistic methods?maximum parsimony and maximum
  likelihood?are more accurate than simpler
  clustering method
• But require large amount of computer time
• Approximate answers
• Calculated phylogenies are often approximations