Chapter 1
Introduction

Majority of organisms on Earth, though diverse, share a significant biological similarity. There is an abundance of biological sequence data showing that any two mammals can have as many as 99% genes in common. Humans and fruit flies are two very different species that share at least 50% common genes. These striking facts have been discovered largely through biological sequence analysis.

Multiple sequence alignment is a fundamental task in bioinformatics and sequence analysis. In the early 1970s, deoxyribonucleic acid (DNA) sequences were obtained using laborious methods based on 2D chromatography. Thus, the number of sequences is limited and often being studied and annotated individually by scientists. By the late 1970s, Gilbert [1] and Sanger and Coulson [2] proposed DNA sequencing by chemical degradation and enzymatic synthesis, respectively. Their works earned a Nobel Prize in chemistry in 1980. Later, sequences are obtained by many newer methods such as dye-based methods [3], microarrays, mass spectrometry, X-ray, ultracentrifugation, and so on. Since the development of Sanger's method, the volume of sequences being identified and deposited is enormous. The current commercial sequencing such as "454 sequencing" can read up to 20 million bases per run and produce the sequences in hours. With this vast amount of sequences, manually annotating each sequence is infeasible. However, we need to categorize them by family, analyze them, find features that are common between them, and so on. The main step to solve this problem is finding the best way to start with the sequence fundamentals, thus leading the readers to the most modern and practical alignment techniques that have been proven to be effective in biological sequence analysis.

1.1 Motivation

There are two popular trends in sequence analysis. One trend focuses primarily on applying rigorous mathematical methods to bring out the optimal alignment of the sequences, thus leading to revelation of possible hidden biological significance between sequences. The other trend stretches on correctly identifying the actual biological significance between the sequences, where some or all biological features may have already been known. These two trends emerge from specific tasks that bioinformatics scientists are dealing with. The first trend relates to prediction of the sequence structures and homology, evolution of species, or determination of the relationship between sequences in order to categorize and organize sequence databases. The second trend is to perform a daily task inwhich scientists want to arrange similar known features of the sequences into the same columns to see how closely they resemble each other. Thus, the second trend can be seen in evolution analysis, in sequence structure and functional analysis, or in drug design and discovery. In the later case, for each specific virus sequence, drug designers search for possible drug-like compounds from libraries of simple sequence models annotated with functional sites and specific drug-like compounds that can bind [4 5] . Hence, aligning a sequence obtained from a new virus against the library of sequences may lead to a manageable set of sequences and compounds to work with.

Consequently, these two distinctive perspectives lead to different approaches to sequence alignment, and the development of sequence alignment algorithms, in turn, allows scientists to automate these tremendous and time-consuming tasks.

1.2 The Organization of this Book

Multiple sequence alignment study involves many aspects of sequence analysis, and it requires broad and significant background information. Therefore, we present each aspect as a chapter starting with existing methodologies and following by our contributions.

The rest of this chapter provides basic information on biological sequences.

1. Chapter 2 provides fundamentals in pairwise sequence alignment.
2. Chapter 3 describes popular existing quantitative models that have been designed to quantify multiple sequence alignment along with their analysis and evaluations.
3. Chapter 4 describes practical clustering techniques that have been used in multiple sequence alignment.
4. Chapter 5 describes, characterizes, and relates many multiple sequence alignment models.
5. Chapter 6 describes how traditional phylogenetic trees have been constructed and how available sequence knowledge bases can be used to improve the accuracy of reconstructing phylogeny trees.
6. Chapter 7 describes the latest methods developed to improve the run-time efficiency of multiple sequence alignment. A large section of this chapter is devoted to parallel alignment model on reconfigurable networks.
7. Chapter 8 describes several popular existing multiple sequence alignment server and services.
8. Chapter 9 describe several multiple sequence alignment techniques that have been developed to handle short sequences (reads) produced by the next-generation sequencing technique (NSG).
9. Chapter 10 describes a bioinformatics application, genetic variant detection, using multiple sequence alignment of short reads or whole genomes as input.
10. Lastly, Chapter 11 provides a review of ribonucleic acid (RNA) and protein secondary structure prediction using the evolution information inferred from multiple sequence alignments.

1.3 Sequence Fundamentals

DNA is the fundamental unit that characterizes a living organism and its genome, that is, its genetic information set. DNA contains thousands of genes that carry the genetic information of a cell. Each gene holds information of how to build a protein molecule, which serves as building blocks for the cell or performs important tasks for the cell functions. The DNA is positioned in the nucleus, which is organized into chromosomes. Since DNA contains the genetic information of the cell, it must be duplicated before the cell divides. This technique is called duplication. When proteins are required, the corresponding genes are transcribed into RNA (transcription), the noncoding parts of the RNA are removed, and the RNA is transported out of the nucleus. Proteins are built outside of the nucleus based on the code in the RNA (see Figure 1.1). Thus, DNA sequence determines protein sequence and its structure; and the protein structure dictates the protein function. In this section, we present the fundamentals of sequences and their basic components.

Figure 1.1 Transcription and translation processes: DNA RNA protein (the "central dogma" of biology).

1.3.1 Protein

Proteins (also known as polypeptides) are organic compounds consisting of amino acids linked with peptide bonds forming a linear polypeptide chain and folded into a globular form. The linear sequence of amino acids in the protein molecule represents its primary structure. The secondary structure refers to regions of local regularity within a protein fold such as -helices, -turns, or -strands. The size of a protein sequence ranges from a few to several thousand residues. Figure 1.2 shows both the structure and the primary sequence of the yeast 1M2V protein. Tables 1.1 and 1.2 list the amino acids and their common abbreviations. Figure 1.3 shows a substitution matrix to quantify the probable substitution between two amino acids in a protein molecule.

Figure 1.2 The yeast Sec23/24 heterodimer 1M2V: (a) protein structure and (b) primary sequence.

Table 1.1 Common Amino Acids

Table 1.2 Ambiguous Amino Acids

Figure 1.3 BLOSUM62...