DEDICATION -- PREFACE -- CONTENTS -- MOLECULAR BIOLOGY FOR MOLECULAR BIOLOGY FOR -- 1. Introduction -- 2. Our Molecular Selves -- 2.1. Cells, DNAs, and Chromosomes -- 2.2. Genes and Genetic Variations -- 2.2.1. Mutations and Genetic Diseases -- 3. Our Biological Machineries -- 3.1. From Gene to Protein: The Central Dogma -- 3.2. The Genetic Code -- 3.3. Gene Regulation and Expression -- 4. Tools of the Trade -- 4.1. Basic Operations -- 4.1.1. Cutting DNA -- 4.1.2. Copying DNA -- 4.1.3. Separating DNA -- 4.1.4. Matching DNA -- 4.2. Putting It Together -- 4.2.1. Genome Sequencing -- 4.2.2. Gene Expression Profiling -- 5. Conclusion -- STRATEGY AND PLANNING OF BIOINFORMATICS EXPERIMENTS -- 1. Multi-Dimensional Bioinformatics -- 2. Reasoning and Strategy -- 3. Planning -- DATA MINING TECHNIQUES FOR THE PRACTICAL BIOINFORMATICIAN -- 1. Feature Selection Methods -- 1.1. Curse of Dimensionality -- 1.2. Signal-to-Noise Measure -- 1.3. T-Test Statistical Measure -- 1.4. Fisher Criterion Score -- 1.5. Entropy Measure -- 1.6. x2 Measure -- 1.7. Information Gain -- 1.8. Information Gain Ratio -- 1.9. Wilcoxon Rank Sum Test -- 1.10. Correlation-Based Feature Selection -- 1.11. Principal Component Analysis -- 1.12. Use of Feature Selection in Bioinformatics -- 2. Classification Methods -- 2.1. Decision Trees -- 2.2. Bayesian Methods -- 2.3. Hidden Markov Models -- 2.4. Artificial Neural Networks -- 2.5. Support Vector Machines -- 2.6. Prediction by Collective Likelihood of Emerging Patterns -- 3. Association Rules Mining Algorithms -- 3.1. Association Rules -- 3.2. The Apriori Algorithm -- 3.3. The Max-Miner Algorithm -- 3.4. Use of Association Rules in Bioinformatics -- 4. Clustering Methods -- 4.1. Factors Affecting Clustering -- 4.2. Partitioning Methods -- 4.3. Hierarchical Methods -- 4.4. Use of Clustering in Bioinformatics -- 5. Remarks.

TECHNIQUES FOR RECOGNITION OF TRANSLATION INITIATION SITES -- 1. Translation Initiation Sites -- 2. Data Set and Evaluation -- 3. Recognition by Perceptrons -- 4. Recognition by Artificial Neural Networks -- 5. Recognition by Engineering of Support Vector Machine Kernels -- 6. Recognition by Feature Generation, Feature Selection, and Feature Integration -- 6.1. Feature Generation -- 6.2. Feature Selection -- 6.3. Feature Integration for Decision Making -- 7. Improved Recognition by Feature Generation, Feature Selection, and Feature Integration -- 8. Recognition by Linear Discriminant Function -- 9. Recognition by Ribosome Scanning -- 10. Remarks -- HOW NEURAL NETWORKS FIND PROMOTERS USING RECOGNITION OF MICRO-STRUCTURAL PROMOTER COMPONENTS -- 1. Motivation and Background -- 1.1. Problem Framework -- 1.2. Promoter Recognition -- 1.3. ANN-Based Promoter Recognition -- 2. CharacteristicMotifs of Eukaryotic Promoters -- 3. Motif-Based Search for Promoters -- 3.1. Evaluation Study by Fickett and Hatzigeorgiou -- 3.2. Enhancers May Contribute to False Recognition -- 4. ANNs and Promoter Components -- 4.1. Description of Promoter Recognition Problem -- 4.2. Representation of Nucleotides for Network Processing -- 5. Structural Decomposition -- 5.1. Parallel Composition of Feature Detectors -- 5.2. First- and Second-Level ANNs -- 5.3. Cascade Composition of Feature Detectors -- 5.4. Structures Based on Multilayer Perceptrons -- 6. Time-Delay Neural Networks -- 6.1. Multistate TDNN -- 6.2. Pruning ANN Connections -- 7. Comments on Performance of ANN-Based Programs for Eukaryotic Promoter Prediction -- NEURAL-STATISTICAL MODEL OF TATA-BOX MOTIFS IN EUKARYOTES -- 1. Promoter Recognition via Recognition of TATA-Box -- 2. PositionWeight Matrix and Statistical Analysis of the TATA-Box and Its Neighborhood.

2.1. TATA Motifs as One of the Targets in the Search for Eukaryotic Promoters -- 2.2. Data Sources and Data Sets -- 2.3. Recognition Quality -- 2.4. Statistical Analysis of TATA Motifs -- 2.4.1. PWM -- 2.4.2. Numerical Characterization of Segments Around TATA-Box -- 2.4.3. Position Analysis of TATA Motifs -- 2.4.4. Characteristics of Segments S1, S2, and S3 -- 2.5. Concluding Remarks -- 3. LVQ ANN for TATA-Box Recognition -- 3.1. Data Preprocessing: Phase 1 -- 3.2. Data Preprocessing: Phase 2- Principal Component Analysis -- 3.3. Learning Vector Quantization ANN -- 3.4. The Structure of an LVQ Classifier -- 3.5. LVQ ANN Training -- 3.6. Initial Values for Network Parameters -- 3.6.1. Genetic Algorithm -- 3.6.2. Searching for Good Initial Weights -- 4. Final Model of TATA-BoxMotif -- 4.1. Structure of Final System for TATA-Box Recognition -- 4.2. Training of the ANN Part of the Model -- 4.3. Performance of Complete System -- 4.3.1. Comparison of MLVQ and System with Single LVQ ANN -- 4.4. The Final Test -- 5. Summary -- TUNING THE DRAGON PROMOTER FINDER SYSTEM FOR HUMAN PROMOTER RECOGNITION -- 1. Promoter Recognition -- 2. Dragon Promoter Finder -- 3. Model -- 4. Tuning Data -- 4.1. Promoter Data -- 4.2. Non-Promoter Data -- 5. Tuning Process -- 6. Discussions and Conclusions -- RNA SECONDARY STRUCTURE PREDICTION -- 1. Introduction to RNA Secondary Structures -- 2. RNA Secondary Structure Determination Experiments -- 3. RNA Structure Prediction Based on Sequence -- 4. Structure Prediction in the Absence of Pseudoknot -- 4.1. Loop Energy -- 4.2. First RNA Secondary Structure Prediction Algorithm -- 4.3. Speeding up Multi-Loops -- 4.4. Speeding Up Internal Loops -- 5. Structure Prediction in the Presence of Pseudoknots -- 6. Akutsu's Algorithm -- 6.1. Definition of Simple Pseudoknot -- 6.2. RNA Secondary Structure Prediction with Simple Pseudoknots.

7. Approximation Algorithm for Predicting Secondary Structure with General Pseudoknots -- PROTEIN SUBCELLULAR LOCALIZATION PREDICTION -- 1. Motivation -- 2. Biology of Localization -- 3. Experimental Techniques for Determining Localization Sites -- 3.1. Traditional Methods -- 3.1.1. Immunofluorescence Microscopy -- 3.1.2. Gradient Centrifugation -- 3.2. Large-Scale Experiments -- 3.2.1. ImmunofluorescentMicroscopy -- 3.2.2. Green Fluorescent Protein -- 3.2.3. Comments on Large-Scale Experiments -- 4. Issues and Complications -- 4.1. How Many Sites are There and What are They? -- 4.2. Is One Site Per Protein an Adequate Model? -- 4.3. How Good are the Predictions? -- 4.3.1. Which Method Should I Use? -- 4.4. A Caveat Regarding Estimated Prediction Accuracies -- 4.5. Correlation and Causality -- 5. Localization and Machine Learning -- 5.1. Standard Classifiers Using (Generalized) Amino Acid Content -- 5.2. Localization Process Modeling Approach -- 5.3. Sequence Based Machine Learning Approaches with Architectures Designed to Reflect Localization Signals -- 5.4. Nearest Neighbor Algorithms -- 5.5. Feature Discovery -- 5.6. Extraction of Localization Information from the Literature and Experimental Data -- 6. Conclusion -- HOMOLOGY SEARCH METHODS -- 1. Overview -- 2. Edit Distance and Alignments -- 2.1. Edit Distance -- 2.2. Optimal Alignments -- 2.3. More Complicated Objectives -- 2.3.1. Score Matrices -- 2.3.2. Gap Penalties -- 3. Sequence Alignment: Dynamic Programming -- 3.1. Dynamic Programming Algorithm for Sequence Alignment -- 3.1.1. Reducing Memory Needs -- 3.2. Local Alignment -- 3.3. Sequence Alignment with Gap Open Penalty -- 4. Probabilistic Approaches to Sequence Alignment -- 4.1. Scoring Matrices -- 4.1.1. PAM Matrices -- 4.1.2. BLOSUM Matrices -- 4.1.3. Weaknesses of this Approach -- 4.2. Probabilistic Alignment Significance.

5. Second Generation Homology Search: Heuristics -- 5.1. FASTA and BLAST -- 5.2. Large-Scale Global Alignment -- 6. Next-Generation Homology Search Software -- 6.1. Improved Alignment Seeds -- 6.2. Optimized Spaced Seeds and Why They Are Better -- 6.3. Computing Optimal Spaced Seeds -- 6.4. Computing More Realistic Spaced Seeds -- 6.4.1. Optimal Seeds for Coding Regions -- 6.4.2. Optimal Seeds for Variable-Length Regions -- 6.5. Approaching Smith-Waterman Sensitivity Using Multiple Seed Models -- 6.6. Complexity of Computing Spaced Seeds -- 7. Experiments -- Acknowledgments -- ANALYSIS OF PHYLOGENY: A CASE STUDY ON SAURURACEAE -- 1. The What,Why, and How of Phylogeny -- 1.1. What is Phylogeny? -- 1.2. Why Study Phylogeny? -- 1.3. How to Study Phylogeny -- 2. Case Study on Phylogeny of Saururaceae -- 3. Materials and Methods -- 3.1. Plant Materials -- 3.2. DNA Extraction, PCR, and Sequencing -- 3.3. Alignment of Sequences -- 3.4. Parsimony Analysis of Separate DNA Sequences -- 3.5. Parsimony Analysis of Combined DNA Sequences -- 3.6. Parsimony Analysis of Morphological Data -- 3.7. Analysis of Each Morphological Characters -- 4. Results -- 4.1. Phylogeny of Saururaceae from 18S Nuclear Genes -- 4.2. Phylogeny of Saururaceae from trnL-F Chloroplast DNA Sequences -- 4.3. Phylogeny of Saururaceae from matR Mitochondrial Genes -- 4.4. Phylogeny of Saururaceae from Combined DNA Sequences -- 4.5. Phylogeny of Saururaceae from Morphological Data -- 5. Discussion -- 5.1. Phylogeny of Saururaceae -- 5.2. The Differences Among Topologies from 18S, trnL-F, and matR -- 5.3. Analysis of Important Morphological Characters -- 6. Suggestions -- 6.1. Sampling -- 6.2. Selecting Out-Group -- 6.3. Gaining Sequences -- 6.4. Aligning -- 6.5. Analyzing -- 6.6. Dealing with Morphological Data -- 6.7. Comparing Phylogenies Separately from Molecular Data and Morphological Data.

6.8. Doing Experiments.