Heuristic pairwise sequence alignment

1
Heuristic pairwise sequence alignment
2
Motivation

• Optimal pairwise sequence alignment is slow
  (Complexity?)
• Typical use of sequence alignment is the
  comparison of a single query sequence versus gt
  106 sequences in database
• Heuristic sequence comparison algorithms
  compromise between performance and accuracy

3
Topics

• Heuristic pairwise sequence alignment algorithms
• Scoring schemes and their statistical
  significance
• Case study BLAST and its variations

4
The BLAST algorithm

• Depict query sequence as a set of k-tup words
  (k3 for proteins k11 for DNA).
• Index database sequences in terms of 203 possible
  3-tups.
• Reduce word search space based on match scores
  derived from amino-acid substitution matrices and
  cut-off threshold.
• Find and extend word matches to generate High
  Scoring Segment Pairs (HSPs) (ungapped BLAST)

5
The BLAST algorithm

• Find and extend inter-word regions (gapped BLAST)
• Screen HSPs in terms of Score and E-values.
• Present candidate alignments

6
Flavors of Blast

• Gapped BLAST
• Ungapped BLAST
• PSI-BLAST
• PHI-BLAST
• RPS-BLAST
• BEAUTY (Purely presentation layer)