Essential bioinformatics

P1: JZP
0521840988pre

CB1022/Xiong

0 521 84098 8

January 10, 2006

This page intentionally left blank

ii

15:7

P1: JZP
0521840988pre

CB1022/Xiong

0 521 84098 8

January 10, 2006

ESSENTIAL BIOINFORMATICS
Essential Bioinformatics is a concise yet comprehensive textbook of bioinformatics that
provides a broad introduction to the entire field. Written specifically for a life science
audience, the basics of bioinformatics are explained, followed by discussions of the stateof-the-art computational tools available to solve biological research problems. All key areas
of bioinformatics are covered including biological databases, sequence alignment, gene
and promoter prediction, molecular phylogenetics, structural bioinformatics, genomics,
and proteomics. The book emphasizes how computational methods work and compares
the strengths and weaknesses of different methods. This balanced yet easily accessible text
will be invaluable to students who do not have sophisticated computational backgrounds.
Technical details of computational algorithms are explained with a minimum use of mathematical formulas; graphical illustrations are used in their place to aid understanding. The
effective synthesis of existing literature as well as in-depth and up-to-date coverage of all
key topics in bioinformatics make this an ideal textbook for all bioinformatics courses
taken by life science students and for researchers wishing to develop their knowledge of
bioinformatics to facilitate their own research.
Jin Xiong is an assistant professor of biology at Texas A&M University, where he has taught
bioinformatics to graduate and undergraduate students for several years. His main research
interest is in the experimental and bioinformatics analysis of photosystems.

i

15:7

P1: JZP
0521840988pre

CB1022/Xiong

0 521 84098 8

January 10, 2006

ii

15:7

P1: JZP
0521840988pre

CB1022/Xiong

0 521 84098 8

January 10, 2006

Essential
Bioinformatics
JIN XIONG
Texas A&M University

iii

15:7

cambridge university press
Cambridge, New York,;  Melbourne, Madrid, Cape Town, Singapore, São Paulo
Cambridge University Press
The Edinburgh Building, Cambridge cb2 2ru, UK
Published in the United States of America by Cambridge University Press, New York
 
Information on this title:  
© Jin Xiong 2006
This publication is in copyright. Subject to statutory exception and to the provision of
relevant collective licensing agreements, no reproduction of any part may take place
without the written permission of Cambridge University Press.
First published in print format 2006
isbn-13
isbn-10

978-0-511-16815-4 eBook (EBL)
0-511-16815-2 eBook (EBL)

isbn-13
isbn-10

978-0-521-84098-9 hardback
0-521-84098-8 hardback

isbn-13
isbn-10

978-0-521-60082-8
0-521-60082-0

Cambridge University Press has no responsibility for the persistence or accuracy of urls
for external or third-party internet websites referred to in this publication, and does not
guarantee that any content on such websites is, or will remain, accurate or appropriate.

P1: JZP
0521840988pre

CB1022/Xiong

0 521 84098 8

January 10, 2006

Contents

Preface ■ ix
SECTION I
1

INTRODUCTION AND BIOLOGICAL DATABASES

Introduction ■ 3
What Is Bioinformatics? ■ 4
Goal ■ 5
Scope ■ 5
Applications ■ 6
Limitations ■ 7
New Themes ■ 8
Further Reading ■ 8

2

Introduction to Biological Databases ■ 10
What Is a Database? ■ 10
Types of Databases ■ 10
Biological Databases ■ 13
Pitfalls of Biological Databases ■ 17
Information Retrieval from Biological Databases ■ 18
Summary ■ 27
Further Reading ■ 27

SECTION II
3

SEQUENCE ALIGNMENT

Pairwise Sequence Alignment ■ 31
Evolutionary Basis ■ 31
Sequence Homology versus Sequence Similarity ■ 32
Sequence Similarity versus Sequence Identity ■ 33
Methods ■ 34
Scoring Matrices ■ 41
Statistical Significance of Sequence Alignment ■ 47
Summary ■ 48
Further Reading ■ 49

4

Database Similarity Searching ■ 51
Unique Requirements of Database Searching ■ 51
Heuristic Database Searching ■ 52
Basic Local Alignment Search Tool (BLAST) ■ 52
FASTA ■ 57
Comparison of FASTA and BLAST ■ 60
Database Searching with the Smith–Waterman Method ■ 61

v

15:7

P1: JZP
0521840988pre

vi

CB1022/Xiong

0 521 84098 8

CONTENTS

Summary ■ 61
Further Reading ■ 62

5

Multiple Sequence Alignment ■ 63
Scoring Function ■ 63
Exhaustive Algorithms ■ 64
Heuristic Algorithms ■ 65
Practical Issues ■ 71
Summary ■ 73
Further Reading ■ 74

6

Profiles and Hidden Markov Models ■ 75
Position-Specific Scoring Matrices ■ 75
Profiles ■ 77
Markov Model and Hidden Markov Model ■ 79
Summary ■ 84
Further Reading ■ 84

7

Protein Motifs and Domain Prediction ■ 85
Identification of Motifs and Domains in Multiple Sequence
Alignment ■ 86
Motif and Domain Databases Using Regular Expressions ■ 86
Motif and Domain Databases Using Statistical Models ■ 87
Protein Family Databases ■ 90
Motif Discovery in Unaligned Sequences ■ 91
Sequence Logos ■ 92
Summary ■ 93
Further Reading ■ 94

SECTION III
8

GENE AND PROMOTER PREDICTION

Gene Prediction ■ 97
Categories of Gene Prediction Programs ■ 97
Gene Prediction in Prokaryotes ■ 98
Gene Prediction in Eukaryotes ■ 103
Summary ■ 111
Further Reading ■ 111

9

Promoter and Regulatory Element Prediction ■ 113
Promoter and Regulatory Elements in Prokaryotes ■ 113
Promoter and Regulatory Elements in Eukaryotes ■ 114
Prediction Algorithms ■ 115
Summary ■ 123
Further Reading ■ 124

SECTION IV
10

MOLECULAR PHYLOGENETICS

Phylogenetics Basics ■ 127
Molecular Evolution and Molecular Phylogenetics ■ 127
Terminology ■ 128
Gene Phylogeny versus Species Phylogeny ■ 130

January 10, 2006

15:7

P1: JZP
0521840988pre

CB1022/Xiong

0 521 84098 8

vii

CONTENTS

Forms of Tree Representation ■ 131
Why Finding a True Tree Is Difficult ■ 132
Procedure ■ 133
Summary ■ 140
Further Reading ■ 141

11

Phylogenetic Tree Construction Methods and Programs ■ 142
Distance-Based Methods ■ 142
Character-Based Methods ■ 150
Phylogenetic Tree Evaluation ■ 163
Phylogenetic Programs ■ 167
Summary ■ 168
Further Reading ■ 169

SECTION V
12

STRUCTURAL BIOINFORMATICS

Protein Structure Basics ■ 173
Amino Acids ■ 173
Peptide Formation ■ 174
Dihedral Angles ■ 175
Hierarchy ■ 176
Secondary Structures ■ 178
Tertiary Structures ■ 180
Determination of Protein Three-Dimensional Structure ■ 181
Protein Structure Database ■ 182
Summary ■ 185
Further Reading ■ 186

13

Protein Structure Visualization, Comparison,
and Classification ■ 187
Protein Structural Visualization ■ 187
Protein Structure Comparison ■ 190
Protein Structure Classification ■ 195
Summary ■ 199
Further Reading ■ 199

14

Protein Secondary Structure Prediction ■ 200
Secondary Structure Prediction for Globular Proteins ■ 201
Secondary Structure Prediction for Transmembrane Proteins ■ 208
Coiled Coil Prediction ■ 211
Summary ■ 212
Further Reading ■ 213

15

January 10, 2006

Protein Tertiary Structure Prediction ■ 214
Methods ■ 215
Homology Modeling ■ 215
Threading and Fold Recognition ■ 223
Ab Initio Protein Structural Prediction ■ 227
CASP ■ 228
Summary ■ 229
Further Reading ■ 230

15:7

P1: JZP
0521840988pre

viii

CB1022/Xiong

0 521 84098 8

CONTENTS

16

RNA Structure Prediction ■ 231
Introduction ■ 231
Types of RNA Structures ■ 233
RNA Secondary Structure Prediction Methods ■ 234
Ab Initio Approach ■ 234
Comparative Approach ■ 237
Performance Evaluation ■ 239
Summary ■ 239
Further Reading ■ 240

SECTION VI
17

GENOMICS AND PROTEOMICS

Genome Mapping, Assembly, and Comparison ■ 243
Genome Mapping ■ 243
Genome Sequencing ■ 245
Genome Sequence Assembly ■ 246
Genome Annotation ■ 250
Comparative Genomics ■ 255
Summary ■ 259
Further Reading ■ 259

18

Functional Genomics ■ 261
Sequence-Based Approaches ■ 261
Microarray-Based Approaches ■ 267
Comparison of SAGE and DNA Microarrays ■ 278
Summary ■ 279
Further Reading ■ 280

19

Proteomics ■ 281
Technology of Protein Expression Analysis ■ 281
Posttranslational Modification ■ 287
Protein Sorting ■ 289
Protein–Protein Interactions ■ 291
Summary ■ 296
Further Reading ■ 296

APPENDIX

Appendix 1. Practical Exercises ■ 301
Appendix 2. Glossary ■ 318
Index ■ 331

January 10, 2006

15:7

P1: JZP
0521840988pre

CB1022/Xiong

0 521 84098 8

January 10, 2006

Preface

With a large number of prokaryotic and eukaryotic genomes completely sequenced
and more forthcoming, access to the genomic information and synthesizing it for
the discovery of new knowledge have become central themes of modern biological
research. Mining the genomic information requires the use of sophisticated computational tools. It therefore becomes imperative for the new generation of biologists to be familiar with many bioinformatics programs and databases to tackle
the new challenges in the genomic era. To meet this goal, institutions in the United
States and around the world are now offering graduate and undergraduate students
bioinformatics-related courses to introduce them to relevant computational tools
necessary for the genomic research. To support this important task, this text was written to provide comprehensive coverage on the state-of-the-art of bioinformatics in a
clear and concise manner.
The idea of writing a bioinformatics textbook originated from my experience of
teaching bioinformatics at Texas A&M University. I needed a text that was comprehensive enough to cover all major aspects in the field, technical enough for a collegelevel course, and sufficiently up to date to include most current algorithms while at
the same time being logical and easy to understand. The lack of such a comprehensive text at that time motivated me to write extensive lecture notes that attempted to
alleviate the problem. The notes turned out to be very popular among the students
and were in great demand from those who did not even take the class. To benefit a
larger audience, I decided to assemble my lecture notes, as well as my experience and
interpretation of bioinformatics, into a book.
This book is aimed at graduate and undergraduate students in biology, or any practicing molecular biologist, who has no background in computer algorithms but wishes
to understand the fundamental principles of bioinformatics and use this knowledge
to tackle his or her own research problems. It covers major databases and software
programs for genomic data analysis, with an emphasis on the theoretical basis and
practical applications of these computational tools. By reading this book, the reader
will become familiar with various computational possibilities for modern molecular
biological research and also become aware of the strengths and weaknesses of each
of the software tools.
The reader is assumed to have a basic understanding of molecular biology and biochemistry. Therefore, many biological terms, such as nucleic acids, amino acids, genes,
transcription, and translation, are used without further explanation. One exception is
protein structure, for which a chapter about fundamental concepts is included so that
ix

15:7

P1: JZP
0521840988pre

x

CB1022/Xiong

0 521 84098 8

January 10, 2006

PREFACE

algorithms and rationales for protein structural bioinformatics can be better understood. Prior knowledge of advanced statistics, probability theories, and calculus is of
course preferable but not essential.
This book is organized into six sections: biological databases, sequence alignment,
genes and promoter prediction, molecular phylogenetics, structural bioinformatics,
and genomics and proteomics. There are nineteen chapters in total, each of which
is relatively independent. When information from one chapter is needed for understanding another, cross-references are provided. Each chapter includes definitions
and key concepts as well as solutions to related computational problems. Occasionally there are boxes that show worked examples for certain types of calculations. Since
this book is primarily for molecular biologists, very few mathematical formulas are
used. A small number of carefully chosen formulas are used where they are absolutely necessary to understand a particular concept. The background discussion of
a computational problem is often followed by an introduction to related computer
programs that are available online. A summary is also provided at the end of each
chapter.
Most of the programs described in this book are online tools that are freely available
and do not require special expertise to use them. Most of them are rather straightforward to use in that the user only needs to supply sequences or structures as input,
and the results are returned automatically. In many cases, knowing which programs
are available for which purposes is sufficient, though occasionally skills of interpreting the results are needed. However, in a number of instances, knowing the names
of the programs and their applications is only half the journey. The user also has to
make special efforts to learn the intricacies of using the programs. These programs
are considered to be on the other extreme of user-friendliness. However, it would be
impractical for this book to try to be a computer manual for every available software
program. That is not my goal in writing the book. Nonetheless, having realized the
difficulties of beginners who are often unaware of or, more precisely, intimidated by
the numerous software programs available, I have designed a number of practical Web
exercises with detailed step-by-step procedures that aim to serve as examples of the
correct use of a combined set of bioinformatics tools for solving a particular problem.
The exercises were originally written for use on a UNIX workstation. However, they
can be used, with slight modifications, on any operating systems with Internet access.
In the course of preparing this book, I consulted numerous original articles and
books related to certain topics of bioinformatics. I apologize for not being able to
acknowledge all of these sources because of space limitations in such an introductory
text. However, a small number of articles (mainly recent review articles) and books
related to the topics of each chapter are listed as “Further Reading” for those who
wish to seek more specialized information on the topics. Regarding the inclusion of
computational programs, there are often a large number of programs available for
a particular task. I apologize for any personal bias in the selection of the software
programs in the book.

15:7

P1: JZP
0521840988pre

CB1022/Xiong

0 521 84098 8

January 10, 2006

xi

PREFACE

One of the challenges in writing this text was to cover sufficient technical background of computational methods without extensive display of mathematical formulas. I strived to maintain a balance between explaining algorithms and not getting
into too much mathematical detail, which may be intimidating for beginning students and nonexperts in computational biology. This sometimes proved to be a tough
balance for me because I risk either sacrificing some of the original content or losing
the reader. To alleviate this problem, I chose in many instances to use graphics instead
of formulas to illustrate a concept and to aid understanding.
I would like to thank the Department of Biology at Texas A&M University for the
opportunity of letting me teach a bioinformatics class, which is what made this book
possible. I thank all my friends and colleagues in the Department of Biology and
the Department of Biochemistry for their friendship. Some of my colleagues were
kind enough to let me participate in their research projects, which provided me with
diverse research problems with which I could hone my bioinformatics analysis skills.
I am especially grateful to Lisa Peres of the Molecular Simulation Laboratory at Texas
A&M, who was instrumental in helping me set up and run the laboratory section
of my bioinformatics course. I am also indebted to my former postdoctoral mentor,
Carl Bauer of Indiana University, who gave me the wonderful opportunity to learn
evolution and phylogenetics in great depth, which essentially launched my career in
bioinformatics. Also importantly, I would like to thank Katrina Halliday, my editor
at Cambridge University Press, for accepting the manuscript and providing numerous suggestions for polishing the early draft. It was a great pleasure working with
her. Thanks also go to Cindy Fullerton and Marielle Poss for their diligent efforts in
overseeing the copyediting of the book to ensure a quality final product.
Jin Xiong

15:7

P1: JZP
0521840988pre

CB1022/Xiong

0 521 84098 8

January 10, 2006

xii

15:7

P1: JZP
0521840988c01

CB1022/Xiong

0 521 84098 8

January 10, 2006

SECTION ONE

Introduction and Biological Databases

1

9:48

P1: JZP
0521840988c01

CB1022/Xiong

0 521 84098 8

January 10, 2006

2

9:48

P1: JZP
0521840988c01

CB1022/Xiong

0 521 84098 8

January 10, 2006

CHAPTER ONE

Introduction

Quantitation and quantitative tools are indispensable in modern biology. Most biological research involves application of some type of mathematical, statistical, or
computational tools to help synthesize recorded data and integrate various types
of information in the process of answering a particular biological question. For example, enumeration and statistics are required for assessing everyday laboratory experiments, such as making serial dilutions of a solution or counting bacterial colonies,
phage plaques, or trees and animals in the natural environment. A classic example in
the history of genetics is by Gregor Mendel and Thomas Morgan, who, by simply counting genetic variations of plants and fruit flies, were able to discover the principles of
genetic inheritance. More dedicated use of quantitative tools may involve using calculus to predict the growth rate of a human population or to establish a kinetic model for
enzyme catalysis. For very sophisticated uses of quantitative tools, one may find application of the “game theory” to model animal behavior and evolution, or the use of millions of nonlinear partial differential equations to model cardiac blood flow. Whether
the application is simple or complex, subtle or explicit, it is clear that mathematical and computational tools have become an integral part of modern-day biological
research. However, none of these examples of quantitative tool use in biology could be
considered to be part of bioinformatics, which is also quantitative in nature. To help the
reader understand the difference between bioinformatics and other elements of quantitative biology, we provide a detailed explanation of what is bioinformatics in the
following sections.
Bioinformatics, which will be more clearly defined below, is the discipline of quantitative analysis of information relating to biological macromolecules with the aid of
computers. The development of bioinformatics as a field is the result of advances in
both molecular biology and computer science over the past 30–40 years. Although
these developments are not described in detail here, understanding the history of this
discipline is helpful in obtaining a broader insight into current bioinformatics research. A succinct chronological summary of the landmark events that have had major
impacts on the development of bioinformatics is presented here to provide context.
The earliest bioinformatics efforts can be traced back to the 1960s, although the
word bioinformatics did not exist then. Probably, the first major bioinformatics project
was undertaken by Margaret Dayhoff in 1965, who developed a first protein sequence
database called Atlas of Protein Sequence and Structure. Subsequently, in the early
1970s, the Brookhaven National Laboratory established the Protein Data Bank for
archiving three-dimensional protein structures. At its onset, the database stored less
3

9:48

P1: JZP
0521840988c01

4

CB1022/Xiong

0 521 84098 8

January 10, 2006

INTRODUCTION

than a dozen protein structures, compared to more than 30,000 structures today.
The first sequence alignment algorithm was developed by Needleman and Wunsch
in 1970. This was a fundamental step in the development of the field of bioinformatics, which paved the way for the routine sequence comparisons and database
searching practiced by modern biologists. The first protein structure prediction algorithm was developed by Chou and Fasman in 1974. Though it is rather rudimentary by
today’s standard, it pioneered a series of developments in protein structure prediction.
The 1980s saw the establishment of GenBank and the development of fast database
searching algorithms such as FASTA by William Pearson and BLAST by Stephen
Altschul and coworkers. The start of the human genome project in the late 1980s
provided a major boost for the development of bioinformatics. The development and
the increasingly widespread use of the Internet in the 1990s made instant access to,
and exchange and dissemination of, biological data possible.
These are only the major milestones in the establishment of this new field. The
fundamental reason that bioinformatics gained prominence as a discipline was the
advancement of genome studies that produced unprecedented amounts of biological
data. The explosion of genomic sequence information generated a sudden demand
for efficient computational tools to manage and analyze the data. The development
of these computational tools depended on knowledge generated from a wide range of
disciplines including mathematics, statistics, computer science, information technology, and molecular biology. The merger of these disciplines created an informationoriented field in biology, which is now known as bioinformatics.

WHAT IS BIOINFORMATICS?
Bioinformatics is an interdisciplinary research area at the interface between computer science and biological science. A variety of definitions exist in the literature
and on the world wide web; some are more inclusive than others. Here, we adopt the
definition proposed by Luscombe et al. in defining bioinformatics as a union of biology and informatics: bioinformatics involves the technology that uses computers for
storage, retrieval, manipulation, and distribution of information related to biological
macromolecules such as DNA, RNA, and proteins. The emphasis here is on the use of
computers because most of the tasks in genomic data analysis are highly repetitive or
mathematically complex. The use of computers is absolutely indispensable in mining
genomes for information gathering and knowledge building.
Bioinformatics differs from a related field known as computational biology. Bioinformatics is limited to sequence, structural, and functional analysis of genes and
genomes and their corresponding products and is often considered computational
molecular biology. However, computational biology encompasses all biological areas
that involve computation. For example, mathematical modeling of ecosystems, population dynamics, application of the game theory in behavioral studies, and phylogenetic construction using fossil records all employ computational tools, but do not
necessarily involve biological macromolecules.

9:48

P1: JZP
0521840988c01

CB1022/Xiong

0 521 84098 8

SCOPE

Beside this distinction, it is worth noting that there are other views of how the two
terms relate. For example, one version defines bioinformatics as the development and
application of computational tools in managing all kinds of biological data, whereas
computational biology is more confined to the theoretical development of algorithms
used for bioinformatics. The confusion at present over definition may partly reflect
the nature of this vibrant and quickly evolving new field.

GOALS
The ultimate goal of bioinformatics is to better understand a living cell and how it
functions at the molecular level. By analyzing raw molecular sequence and structural
data, bioinformatics research can generate new insights and provide a “global” perspective of the cell. The reason that the functions of a cell can be better understood
by analyzing sequence data is ultimately because the flow of genetic information is
dictated by the “central dogma” of biology in which DNA is transcribed to RNA, which
is translated to proteins. Cellular functions are mainly performed by proteins whose
capabilities are ultimately determined by their sequences. Therefore, solving functional problems using sequence and sometimes structural approaches has proved to
be a fruitful endeavor.

SCOPE
Bioinformatics consists of two subfields: the development of computational tools and
databases and the application of these tools and databases in generating biological
knowledge to better understand living systems. These two subfields are complementary to each other. The tool development includes writing software for sequence,
structural, and functional analysis, as well as the construction and curating of biological databases. These tools are used in three areas of genomic and molecular biological
research: molecular sequence analysis, molecular structural analysis, and molecular
functional analysis. The analyses of biological data often generate new problems and
challenges that in turn spur the development of new and better computational tools.
The areas of sequence analysis include sequence alignment, sequence database
searching, motif and pattern discovery, gene and promoter finding, reconstruction of
evolutionary relationships, and genome assembly and comparison. Structural analyses include protein and nucleic acid structure analysis, comparison, classification,
and prediction. The functional analyses include gene expression profiling, protein–
protein interaction prediction, protein subcellular localization prediction, metabolic
pathway reconstruction, and simulation (Fig. 1.1).
The three aspects of bioinformatics analysis are not isolated but often interact
to produce integrated results (see Fig. 1.1). For example, protein structure prediction depends on sequence alignment data; clustering of gene expression profiles
requires the use of phylogenetic tree construction methods derived in sequence
analysis. Sequence-based promoter prediction is related to functional analysis of

January 10, 2006

5

9:48

P1: JZP
0521840988c01

6

CB1022/Xiong

0 521 84098 8

January 10, 2006

INTRODUCTION

Figure 1.1: Overview of various subfields of bioinformatics. Biocomputing tool development is at the
foundation of all bioinformatics analysis. The applications of the tools fall into three areas: sequence
analysis, structure analysis, and function analysis. There are intrinsic connections between different
areas of analyses represented by bars between the boxes.

coexpressed genes. Gene annotation involves a number of activities, which include
distinction between coding and noncoding sequences, identification of translated
protein sequences, and determination of the gene’s evolutionary relationship with
other known genes; prediction of its cellular functions employs tools from all three
groups of the analyses.

APPLICATIONS
Bioinformatics has not only become essential for basic genomic and molecular
biology research, but is having a major impact on many areas of biotechnology
and biomedical sciences. It has applications, for example, in knowledge-based drug
design, forensic DNA analysis, and agricultural biotechnology. Computational studies
of protein–ligand interactions provide a rational basis for the rapid identification of
novel leads for synthetic drugs. Knowledge of the three-dimensional structures of proteins allows molecules to be designed that are capable of binding to the receptor site
of a target protein with great affinity and specificity. This informatics-based approach

9:48

P1: JZP
0521840988c01

CB1022/Xiong

0 521 84098 8

LIMITATIONS

significantly reduces the time and cost necessary to develop drugs with higher potency,
fewer side effects, and less toxicity than using the traditional trial-and-error approach.
In forensics, results from molecular phylogenetic analysis have been accepted as evidence in criminal courts. Some sophisticated Bayesian statistics and likelihood-based
methods for analysis of DNA have been applied in the analysis of forensic identity. It
is worth mentioning that genomics and bioinformtics are now poised to revolutionize our healthcare system by developing personalized and customized medicine. The
high speed genomic sequencing coupled with sophisticated informatics technology
will allow a doctor in a clinic to quickly sequence a patient’s genome and easily detect
potential harmful mutations and to engage in early diagnosis and effective treatment
of diseases. Bioinformatics tools are being used in agriculture as well. Plant genome
databases and gene expression profile analyses have played an important role in the
development of new crop varieties that have higher productivity and more resistance
to disease.

LIMITATIONS
Having recognized the power of bioinformatics, it is also important to realize its limitations and avoid over-reliance on and over-expectation of bioinformatics output.
In fact, bioinformatics has a number of inherent limitations. In many ways, the role
of bioinformatics in genomics and molecular biology research can be likened to the
role of intelligence gathering in battlefields. Intelligence is clearly very important in
leading to victory in a battlefield. Fighting a battle without intelligence is inefficient
and dangerous. Having superior information and correct intelligence helps to identify
the enemy’s weaknesses and reveal the enemy’s strategy and intentions. The gathered
information can then be used in directing the forces to engage the enemy and win
the battle. However, completely relying on intelligence can also be dangerous if the
intelligence is of limited accuracy. Overreliance on poor-quality intelligence can yield
costly mistakes if not complete failures.
It is no stretch in analogy that fighting diseases or other biological problems using
bioinformatics is like fighting battles with intelligence. Bioinformatics and experimental biology are independent, but complementary, activities. Bioinformatics depends
on experimental science to produce raw data for analysis. It, in turn, provides useful
interpretation of experimental data and important leads for further experimental
research. Bioinformatics predictions are not formal proofs of any concepts. They
do not replace the traditional experimental research methods of actually testing
hypotheses. In addition, the quality of bioinformatics predictions depends on the
quality of data and the sophistication of the algorithms being used. Sequence data
from high throughput analysis often contain errors. If the sequences are wrong or
annotations incorrect, the results from the downstream analysis are misleading as
well. That is why it is so important to maintain a realistic perspective of the role of
bioinformatics.

January 10, 2006

7

9:48

P1: JZP
0521840988c01

8

CB1022/Xiong

0 521 84098 8

January 10, 2006

INTRODUCTION

Bioinformatics is by no means a mature field. Most algorithms lack the capability and sophistication to truly reflect reality. They often make incorrect predictions
that make no sense when placed in a biological context. Errors in sequence alignment, for example, can affect the outcome of structural or phylogenetic analysis. The
outcome of computation also depends on the computing power available. Many
accurate but exhaustive algorithms cannot be used because of the slow rate of computation. Instead, less accurate but faster algorithms have to be used. This is a necessary
trade-off between accuracy and computational feasibility. Therefore, it is important
to keep in mind the potential for errors produced by bioinformatics programs. Caution
should always be exercised when interpreting prediction results. It is a good practice
to use multiple programs, if they are available, and perform multiple evaluations. A
more accurate prediction can often be obtained if one draws a consensus by comparing results from different algorithms.

NEW THEMES
Despite the pitfalls, there is no doubt that bioinformatics is a field that holds great
potential for revolutionizing biological research in the coming decades. Currently, the
field is undergoing major expansion. In addition to providing more reliable and more
rigorous computational tools for sequence, structural, and functional analysis, the
major challenge for future bioinformatics development is to develop tools for elucidation of the functions and interactions of all gene products in a cell. This presents
a tremendous challenge because it requires integration of disparate fields of biological knowledge and a variety of complex mathematical and statistical tools. To gain
a deeper understanding of cellular functions, mathematical models are needed to
simulate a wide variety of intracellular reactions and interactions at the whole cell
level. This molecular simulation of all the cellular processes is termed systems biology.
Achieving this goal will represent a major leap toward fully understanding a living system. That is why the system-level simulation and integration are considered the future
of bioinformatics. Modeling such complex networks and making predictions about
their behavior present tremendous challenges and opportunities for bioinformaticians. The ultimate goal of this endeavor is to transform biology from a qualitative
science to a quantitative and predictive science. This is truly an exciting time for
bioinformatics.

FURTHER READING
Attwood, T. K., and Miller, C. J. 2002. Progress in bioinformatics and the importance of being
earnest. Biotechnol. Annu. Rev. 8:1–54.
Golding, G. B. 2003. DNA and the revolution of molecular evolution, computational biology,
and bioinformatics. Genome 46:930–5.
Goodman, N. 2002. Biological data becomes computer literature: New advances in bioinformatics. Curr. Opin. Biotechnol. 13:68–71.

9:48

P1: JZP
0521840988c01

CB1022/Xiong

0 521 84098 8

FURTHER READING

Hagen. J. B. 2000. The origin of bioinformatics. Nat. Rev. Genetics 1:231–6.
Kanehisa, M., and Bork, P. 2003. Bioinformatics in the post-sequence era. Nat. Genet. 33
Suppl:305–10.
Kim, J. H. 2002. Bioinformatics and genomic medicine. Genet. Med. 4 Suppl:62S–5S.
Luscombe, N. M., Greenbaum, D., and Gerstein, M. 2001. What is bioinformatics? A proposed
definition and overview of the field. Methods Inf. Med. 40:346–58.
Ouzounis, C. A., and Valencia, A. 2003. Early bioinformatics: The birth of a discipline – A personal
view. Bioinformatics 19:2176–90.

January 10, 2006

9

9:48

P1: JZP
0521840988c02

CB1022/Xiong

0 521 84098 8

January 10, 2006

CHAPTER TWO

Introduction to Biological Databases

One of the hallmarks of modern genomic research is the generation of enormous
amounts of raw sequence data. As the volume of genomic data grows, sophisticated
computational methodologies are required to manage the data deluge. Thus, the very
first challenge in the genomics era is to store and handle the staggering volume of information through the establishment and use of computer databases. The development
of databases to handle the vast amount of molecular biological data is thus a fundamental task of bioinformatics. This chapter introduces some basic concepts related to
databases, in particular, the types, designs, and architectures of biological databases.
Emphasis is on retrieving data from the main biological databases such as GenBank.

WHAT IS A DATABASE?
A database is a computerized archive used to store and organize data in such a way
that information can be retrieved easily via a variety of search criteria. Databases
are composed of computer hardware and software for data management. The chief
objective of the development of a database is to organize data in a set of structured
records to enable easy retrieval of information. Each record, also called an entry,
should contain a number of fields that hold the actual data items, for example, fields
for names, phone numbers, addresses, dates. To retrieve a particular record from the
database, a user can specify a particular piece of information, called value, to be found
in a particular field and expect the computer to retrieve the whole data record. This
process is called making a query.
Although data retrieval is the main purpose of all databases, biological databases
often have a higher level of requirement, known as knowledge discovery, which refers
to the identification of connections between pieces of information that were not
known when the information was first entered. For example, databases containing
raw sequence information can perform extra computational tasks to identify sequence
homology or conserved motifs. These features facilitate the discovery of new biological
insights from raw data.

TYPES OF DATABASES
Originally, databases all used a flat file format, which is a long text file that contains
many entries separated by a delimiter, a special character such as a vertical bar (|).
Within each entry are a number of fields separated by tabs or commas. Except for the
10

14:42

P1: JZP
0521840988c02

CB1022/Xiong

0 521 84098 8

January 10, 2006

TYPES OF DATABASES

raw values in each field, the entire text file does not contain any hidden instructions
for computers to search for specific information or to create reports based on certain
fields from each record. The text file can be considered a single table. Thus, to search
a flat file for a particular piece of information, a computer has to read through the
entire file, an obviously inefficient process. This is manageable for a small database,
but as database size increases or data types become more complex, this database style
can become very difficult for information retrieval. Indeed, searches through such files
often cause crashes of the entire computer system because of the memory-intensive
nature of the operation.
To facilitate the access and retrieval of data, sophisticated computer software
programs for organizing, searching, and accessing data have been developed. They
are called database management systems. These systems contain not only raw data
records but also operational instructions to help identify hidden connections among
data records. The purpose of establishing a data structure is for easy execution of the
searches and to combine different records to form final search reports. Depending
on the types of data structures,these database management systems can be classified
into two types: relational database management systems and object-oriented database
management systems. Consequently, databases employing these management systems are known as relational databases or object-oriented databases, respectively.

Relational Databases
Instead of using a single table as in a flat file database, relational databases use a set
of tables to organize data. Each table, also called a relation, is made up of columns
and rows. Columns represent individual fields. Rows represent values in the fields of
records. The columns in a table are indexed according to a common feature called
an attribute, so they can be cross-referenced in other tables. To execute a query in
a relational database, the system selects linked data items from different tables and
combines the information into one report. Therefore, specific information can be
found more quickly from a relational database than from a flat file database.
Relational databases can be created using a special programming language called
structured query language (SQL). The creation of this type of databases can take a great
deal of planning during the design phase. After creation of the original database, a
new data category can be easily added without requiring all existing tables to be modified. The subsequent database searching and data gathering for reports are relatively
straightforward.
Here is a simple example of student course information expressed in a flat file
which contains records of five students from four different states, each taking a different course (Fig. 2.1). Each data record, separated by a vertical bar, contains four
fields describing the name, state, course number and title. A relational database
is also created to store the same information, in which the data are structured as
a number of tables. Figure 2.1 shows how the relational database works. In each
table, data that fit a particular criterion are grouped together. Different tables can
be linked by common data categories, which facilitate finding of specific information.

11

14:42

P1: JZP
0521840988c02

12

CB1022/Xiong

0 521 84098 8

January 10, 2006

INTRODUCTION TO BIOLOGICAL DATABASES

Figure 2.1: Example of constructing a relational database for five students’ course information originally
expressed in a flat file. By creating three different tables linked by common fields, data can be easily
accessed and reassembled.

For example, if one is to ask the question, which courses are students from Texas
taking? The database will first find the field for “State” in Table A and look up for
Texas. This returns students 1 and 5. The student numbers are colisted in Table B,
in which students 1 and 5 correspond to Biol 689 and Math 172, respectively. The
course names listed by course numbers are found in Table C. By going to Table C, exact
course names corresponding to the course numbers can be retrieved. A final report is
then given showing that the Texans are taking the courses Bioinformatics and Calculus. However, executing the same query through the flat file requires the computer to
read through the entire text file word by word and to store the information in a temporay memory space and later mark up the data records containing the word Texas. This
is easily accomplishable for a small database. To perform queries in a large database
using flat files obviously becomes an onerous task for the computer system.

Object-Oriented Databases
One of the problems with relational databases is that the tables used do not describe
complex hierarchical relationships between data items. To overcome the problem,
object-oriented databases have been developed that store data as objects. In an
object-oriented programming language, an object can be considered as a unit that
combines data and mathematical routines that act on the data. The database is structured such that the objects are linked by a set of pointers defining predetermined relationships between the objects. Searching the database involves navigating through the
objects with the aid of the pointers linking different objects. Programming languages
like C++ are used to create object-oriented databases.
The object-oriented database system is more flexible; data can be structured based
on hierarchical relationships. By doing so, programming tasks can be simplified for
data that are known to have complex relationships, such as multimedia data. However,

14:42

P1: JZP
0521840988c02

CB1022/Xiong

0 521 84098 8

January 10, 2006

BIOLOGICAL DATABASES

Figure 2.2: Example of construction and query of an object-oriented database using the same student
information as shown in Figure 2.1. Three objects are constructed and are linked by pointers shown
as arrows. Finding specific information relies on navigating through the objects by way of pointers. For
simplicity, some of the pointers are omitted.

this type of database system lacks the rigorous mathematical foundation of the
relational databases. There is also a risk that some of the relationships between objects
may be misrepresented. Some current databases have therefore incorporated features
of both types of database programming, creating the object–relational database management system.
The above students’ course information (Fig. 2.1) can be used to construct an
object-oriented database. Three different objects can be designed: student object,
course object, and state object. Their interrelations are indicated by lines with arrows
(Fig. 2.2). To answer the same question – which courses are students from Texas
taking – one simply needs to start from Texas in the state object, which has pointers
that lead to students 1 and 5 in the student object. Further pointers in the student
object point to the course each of the two students is taking. Therefore, a simple
navigation through the linked objects provides a final report.

BIOLOGICAL DATABASES
Current biological databases use all three types of database structures: flat files,
relational, and object oriented. Despite the obvious drawbacks of using flat files in
database management, many biological databases still use this format. The justification for this is that this system involves minimum amount of database design and the
search output can be easily understood by working biologists.

13

14:42

P1: JZP
0521840988c02

14

CB1022/Xiong

0 521 84098 8

January 10, 2006

INTRODUCTION TO BIOLOGICAL DATABASES

Based on their contents, biological databases can be roughly divided into three
categories: primary databases, secondary databases, and specialized databases.
Primary databases contain original biological data. They are archives of raw sequence
or structural data submitted by the scientific community. GenBank and Protein Data
Bank (PDB) are examples of primary databases. Secondary databases contain computationally processed or manually curated information, based on original information from primary databases. Translated protein sequence databases containing
functional annotation belong to this category. Examples are SWISS-Prot and Protein Information Resources (PIR) (successor of Margaret Dayhoff’s Atlas of Protein
Sequence and Structure [see Chapter 1]). Specialized databases are those that cater
to a particular research interest. For example, Flybase, HIV sequence database, and
Ribosomal Database Project are databases that specialize in a particular organism
or a particular type of data. A list of some frequently used databases is provided in
Table 2.1.

Primary Databases
There are three major public sequence databases that store raw nucleic acid sequence
data produced and submitted by researchers worldwide: GenBank, the European
Molecular Biology Laboratory (EMBL) database and the DNA Data Bank of Japan
(DDBJ), which are all freely available on the Internet. Most of the data in the databases
are contributed directly by authors with a minimal level of annotation. A small number
of sequences, especially those published in the 1980s, were entered manually from
published literature by database management staff.
Presently, sequence submission to either GenBank, EMBL, or DDBJ is a precondition for publication in most scientific journals to ensure the fundamental molecular
data to be made freely available. These three public databases closely collaborate
and exchange new data daily. They together constitute the International Nucleotide
Sequence Database Collaboration. This means that by connecting to any one of
the three databases, one should have access to the same nucleotide sequence data.
Although the three databases all contain the same sets of raw data, each of the individual databases has a slightly different kind of format to represent the data.
Fortunately, for the three-dimensional structures of biological macromolecules,
there is only one centralized database, the PDB. This database archives atomic coordinates of macromolecules (both proteins and nucleic acids) determined by x-ray
crystallography and NMR. It uses a flat file format to represent protein name, authors,
experimental details, secondary structure, cofactors, and atomic coordinates. The
web interface of PDB also provides viewing tools for simple image manipulation.
More details of this database and its format are provided in Chapter 12.

Secondary Databases
Sequence annotation information in the primary database is often minimal. To
turn the raw sequence information into more sophisticated biological knowledge,
much postprocessing of the sequence information is needed. This begs the need for

14:42

P1: JZP
0521840988c02

CB1022/Xiong

0 521 84098 8

January 10, 2006

15

BIOLOGICAL DATABASES

TABLE 2.1. Major Biological Databases Available Via the World Wide Web

Databases and
Retrieval
Systems
AceDB
DDBJ
EMBL
Entrez
ExPASY
FlyBase
FSSP
GenBank
HIV databases
Microarray
gene
expression
database
OMIM
PIR
PubMed
Ribosomal
database
project
SRS
SWISS-Prot
TAIR

Brief Summary of Content

URL

Genome database for
Caenorhabditis elegans
Primary nucleotide sequence
database in Japan
Primary nucleotide sequence
database in Europe
NCBI portal for a variety
of biological databases
Proteomics database
A database of the Drosophila
genome
Protein secondary structures
Primary nucleotide sequence
database in NCBI
HIV sequence data and related
immunologic information
DNA microarray data and
analysis tools

 

Genetic information of human
diseases
Annotated protein sequences
Biomedical literature
information
Ribosomal RNA sequences and
phylogenetic trees derived
from the sequences
General sequence retrieval
system
Curated protein sequence
database
Arabidopsis information
database

 
 
 
 
 
 
 
 
 

 
 
 
 

 
 
 

secondary databases, which contain computationally processed sequence information derived from the primary databases. The amount of computational processing work varies greatly among the secondary databases; some are simple archives of
translated sequence data from identified open reading frames in DNA, whereas others
provide additional annotation and information related to higher levels of information
regarding structure and functions.
A prominent example of secondary databases is SWISS-PROT, which provides
detailed sequence annotation that includes structure, function, and protein family assignment. The sequence data are mainly derived from TrEMBL, a database of

14:42

P1: JZP
0521840988c02

16

CB1022/Xiong

0 521 84098 8

January 10, 2006

INTRODUCTION TO BIOLOGICAL DATABASES

translated nucleic acid sequences stored in the EMBL database. The annotation of
each entry is carefully curated by human experts and thus is of good quality. The protein annotation includes function, domain structure, catalytic sites, cofactor binding,
posttranslational modification, metabolic pathway information, disease association,
and similarity with other sequences. Much of this information is o