# Chapter 3. Sequence Alignment

BLAST finds statistically significant similarities between sequences by evaluating alignments, but how are sequences aligned? In principle, there are many ways to align two sequences, but in practice, one method is used more often than any other. This chapter explains this technique with the biologist in mind, without using the mathematical notation and jargon that is usually employed to describe such algorithm. Divested of unfamiliar language and notation, these algorithms are quite simple.

Finding the optimal alignment between two sequences can be a computationally complex task. Fortunately, a technique called dynamic programming (DP) makes sequence alignment tractable as long as you follow a few rules. Rather than have you struggle with a confusing definition of DP, this chapter demonstrates how the technique works for sequence alignment and then gets back to the generalities. There are fundamentally two kinds of alignment: global and local. In global alignment, both sequences are aligned along their entire lengths and the best alignment is found. In local alignment, the best subsequence alignment is found. For example, if you want to find the two most similar sentences between two books, you use local alignment. If you want to compare the sentences end to end, use global alignment. This chapter describes global alignment, then local alignment. The example uses English words instead of biological sequences and the algorithms are quite general.

 Forward
 Preface
 Part I: Introduction
 Part III: Practice
 Part IV: Industrial-Strength BLAST
 Part V: BLAST Reference
 Part VI: Appendixes
 Glossary
 Colophon