Neural Networks and Genome Informatics (eBook)

Cover 1
Contents 10
PART I: Overview 14
CHAPTER 1. Neural Networks for Genome Informatics 16
1.1 What Is Genome Informatics? 16
1.2 What Is An Artificial Neural Network? 23
1.3 Genome Informatics Applications 24
1.4 References 25
PART II: Neural Network Foundations 30
CHAPTER 2. Neural Network Basics 32
2.1 Introduction to Neural Network Elements 32
2.2 Transfer Functions 34
2.4 Simple Feed-Forward Network Example 38
2.5 Introductory Texts 39
2.6 References 40
CHAPTER 3. Perceptrons and Multilayer Perceptrons 42
3.1 Perceptrons 42
3.2 Multilayer Perceptrons 46
3.3 References 51
CHAPTER 4. Other Common Architectures 54
4.1 Radial Basis Functions 54
4.2 Kohonen Self-organizing Maps 59
4.4 References 63
CHAPTER 5. Training of Neural Networks 64
5.1 Supervised Learning 64
5.3 Unsupervised Learning 75
5.4 Software for Training Neural Networks 76
5.5 References 76
PART III: Genome Informatics Applications 78
CHAPTER 6. Design Issues - Feature Presentation 80
6.1 Overview of Design Issues 80
6.2 Amino Acid Residues 81
6.3 Amino Acid Physicochemical and Structural Features 82
6.4 Protein Context Features and Domains 84
6.5 Protein Evolutionary Features 86
6.6 Feature Representation 87
6.7 References 89
CHAPTER 7. Design Issues - Data Encoding 92
7.1 Direct Input Sequence Encoding 92
7.2 Indirect Input Sequence Encoding 94
7.3 Construction of Input Layer 96
7.4 Input Trimming 97
7.5 Output Encoding 99
7.6 References 99
CHAPTER 8. Design Issues - Neural Networks 102
8.1 Network Architecture 102
8.2 Network Learning Algorithm 104
8.3 Network Parameters 105
8.4 Training and Test Data 107
8.5 Evaluation Mechanism 110
8.6 References 112
CHAPTER 9. Applications - Nucleic Acid Sequence Analysis 116
9.1 Introduction 116
9.2 Coding Region Recognition and Gene Identification 118
9.3 Recognition of Transcriptional and Translational Signals 120
9.4 Sequence Feature Analysis and Classification 123
9.5 References 124
CHAPTER 10. Applications - Protein Structure Prediction 128
10.1 Introduction 129
10.2 Protein Secondary Structure Prediction 129
10.3 Protein Tertiary Structure Prediction Protein Distance Constraints 134
10.4 Protein Folding Qass Prediction 136
10.5 References 138
CHAPTER 11. Applications - Protein Sequence Analysis 142
11.1 Introduction 142
11.2 Signal Peptide Prediction 143
11.3 Other Motif Region and Site Prediction 146
11.4 Protein Family Classification 149
11.5 References 153
Part IV : Open Problems and Future Directions 156
CHAPTER 12. Integration of Statistical Methods into Neural Network Applications 158
12.1 Problems in Model Development 159
12.2 Training Issues 161
12.3 Interpretation of Results 162
12.4Further Sources of Information 162
12.5 References 162
CHAPTER 13. Future of Genome Informatics Applications 165
13.1 Rule and Feature Extraction from Neural Networks 165
13.2 Neural Network Design Using Prior Knowledge 169
13.3 Conclusions 170
13.4 References 171
Glossary 174
Author Index 206
Subject Index 214