Introduction to Bioinformatics: CMSC54610-1, Spring quarter 2006.

Lecture 5

• Chapter 4: Distance methods for phylogeny
  • Review from last week:
    • About distance metrics for biological sequences.
    • Distances in drug design
      • Ariel Fernandez and Sridhar Maddipati, A Priori Inference of Cross Reactivity for Drug-Targeted Kinases, J. Med. Chem. 49, pages 3092-3100, 2006.
    • Consider the scoring matrices in Figure 2.4 on page 39 of the text. If S is the scoring matrix, suppose that the diagonal is k. Then D=k-S.
    • Finding the tree for three nodes: solve 3x3 system of equations
  • Rooted versus unrooted trees: too many but not enough
  • Gene versus species trees: local versus global view
  • UPGMA: clustering technique
    • group closest nodes, define distances for new groups
    • changes the distance matrix (cf. Table 4.2)
  • The four point condition: additive distance matrices revised notes.
  • Finding the right pairs: Sattath and Tversky, 1977 Psychometrika 42, pp. 319-345
• Chapter 5: Character based methods of phylogenetics
  • Parsimony
    • Main difference: internal nodes have a name
    • Informative sites: a heuristic to find the critical determinants
    • In the end, we compute a distance
  • Bootstrapping
• Discussion of the term paper

• Homework problems:
  • 4-4, 4.8 (pages 95-96)
  • Give an example of a distance matrix for which A>B=C and show that there are trees with three different topologies optimally close (distance = A-B) in the L1 norm (see section 3 of the notes on the four point condition)
  • 5-2 (page 116)
• Homework is due by the beginning of the next class.
  • Homework should be turned in physically before the start of the class
  • or sent electronically in standard formats in an emergency.