Phylogenetics

1
Phylogenetics
2
Topics

• Motivation
• General Issues
• Algorithms
• Implementations

3
Motivation

• Thou art my brother but a few sequences removed
• Generate topologies of evolutionary relationships
  of sequences
• When used on a large scale, suggestive of
  evolutionary relationships between species
• Heuristic forays into NP-complete territory

4
Species phylogeny ! Sequence phylogeny

• Sequences evolve at different rates
• Within a single species
• Between different species
• Especially in bacteria, horizontal transfer
  (Napsters been around for ages) quite common

5
General Principles

• Start with pairwise distances
• Path length Distance (Evolutionary time)
• Work from leaves to node to generate tree
• (opposite of binary tree generation)
• Its easier to be rootless than to be rooted
• Binary tree approximation of higher order trees
• Edges do not imply direct links (Missing
  links/incomplete data), only a representation of
  sequence evolution

6
Algorithms

• UPGMA (Unweighted pair group method with
  arithmetic averages)
• If my cousin looks more like me than my brother,
  he must be my lost brother, and perhaps my
  brother my cousin?
• Neighbor joining
• Some sequences evolve faster than others
• Parsimony
• Its just a bruise, not Kaposis sarcoma!
• Maximum Likelihood
• Given the facts, Watson, the answer is
  elementary!

7
UPGMA

• No. of clusters no. of sequences no. of leaf
  nodes
• Inter cluster distance Average pairwise
  distance
• While (no. of clusters gt 1)
• Connect pair of closest clusters (at distance d)
  with intermediate node at distance d/2 from each
  of them
• Caveat Satisfies minimal distance requirement,
  but may result in spurious topologies because
  of constant rate evolution assumption