BLAST

This is the Title of the Book, eMatter Edition

Copyright © 2012 O’Reilly & Associates, Inc. All rights reserved.

55

Chapter 4

CHAPTER 4

Sequence Similarity

The fact that the human genome is often referred to as the Book of Life is an apt

description because nucleic acids and proteins are often represented (and manipu-

lated) as text files. Chapter 3 described common algorithms for aligning sequences of

letters, and score is the metric used to determine the best alignment. This chapter

shows what scores really are. Some of the introduced terms come from information

theory, so the chapter begins with a brief introduction to this branch of mathemat-

ics. It then explores the typical ways to measure sequence similarity. You’ll see that

this approach fits well with the sequence-alignment algorithms described in

Chapter 3. The last part of the chapter focuses on the statistical significance of

sequence similarity in a database search. The theories discussed in this chapter apply

only to local alignment. There is currently no theory for global alignment.

Introduction to Information Theory

In common usage, the word information conveys many things. Forget everything you

know about this word because you’re going to learn the most precise definition.

Information is a decrease in uncertainty. You can also think of information as a

degree of surprise.

Suppose you’re taking care of a child and the response to