Statistical Bioinformatics: For Biomedical and Life Science Researchers

This book provides an essential understanding of statistical concepts necessary for the analysis of genomic and proteomic data using computational techniques. The author presents both basic and advanced topics, focusing on those that are relevant to the computational analysis of large data sets in biology. Chapters begin with a description of a statistical concept and a current example from biomedical research, followed by more detailed presentation, discussion of limitations, and problems. The book starts with an introduction to probability and statistics for genome-wide data, and moves into topics such as clustering, classification, multi-dimensional visualization, experimental design, statistical resampling, and statistical network analysis. Clearly explains the use of bioinformatics tools in life sciences research without requiring an advanced background in math/statistics Enables biomedical and life sciences researchers to successfully evaluate the validity of their results and make inferences Enables statistical and quantitative researchers to rapidly learn novel statistical concepts and techniques appropriate for large biological data analysis Carefully revisits frequently used statistical approaches and highlights their limitations in large biological data analysis Offers programming examples and datasets Includes chapter problem sets, a glossary, a list of statistical notations, and appendices with references to background mathematical and technical material Features supplementary materials, including datasets, links, and a statistical package available online Statistical Bioinformatics is an ideal textbook for students in medicine, life sciences, and bioengineering, aimed at researchers who utilize computational tools for the analysis of genomic, proteomic, and many other emerging high-throughput molecular data. It may also serve as a rapid introduction to the bioinformatics science for statistical and computational students and audiences who have not experienced such analysis tasks before.

Table of Contents:
Preface xi Contributors xiii 1 Road to Statistical Bioinformatics 1 Challenge 1: Multiple-Comparisons Issue 1 Challenge 2: High-Dimensional Biological Data 2 Challenge 3: Small-n and Large-p Problem 3 Challenge 4: Noisy High-Throughput Biological Data 3 Challenge 5: Integration of Multiple, Heterogeneous Biological Data Information 3 References 5 2 Probability Concepts and Distributions for Analyzing Large Biological Data 7 2.1 Introduction 7 2.2 Basic Concepts 8 2.3 Conditional Probability and Independence 10 2.4 Random Variables 13 2.5 Expected Value and Variance 15 2.6 Distributions of Random Variables 19 2.7 Joint and Marginal Distribution 39 2.8 Multivariate Distribution 42 2.9 Sampling Distribution 46 2.10 Summary 54 3 Quality Control of High-throughput Biological Data 57 3.1 Sources of Error in High-Throughput Biological Experiments 57 3.2 Statistical Techniques for Quality Control 59 3.3 Issues Specific to Microarray Gene Expression Experiments 66 3.4 Conclusion 69 References 69 4 Statistical Testing and Significance for Large Biological Data Analysis 71 4.1 Introduction 71 4.2 Statistical Testing 72 4.3 Error Controlling 78 4.4 Real Data Analysis 81 4.5 Concluding Remarks 87 Acknowledgments 87 References 88 5 Clustering: Unsupervised Learning in Large Biological Data 89 5.1 Measures of Similarity 90 5.2 Clustering 99 5.3 Assessment of Cluster Quality 115 5.4 Conclusion 123 References 123 6 Classification: Supervised Learning with High-dimensional Biological Data 129 6.1 Introduction 129 6.2 Classification and Prediction Methods 132 6.3 Feature Selection and Ranking 140 6.4 Cross-Validation 144 6.5 Enhancement of Class Prediction by Ensemble Voting Methods 145 6.6 Comparison of Classification Methods Using High-Dimensional Data 147 6.7 Software Examples for Classification Methods 150 References 154 7 Multidimensional Analysis and Visualization on Large Biomedical Data 157 7.1 Introduction 157 7.2 Classical Multidimensional Visualization Techniques 158 7.3 Two-Dimensional Projections 161 7.4 Issues and Challenges 165 7.5 Systematic Exploration of Low-Dimensional Projections 166 7.6 One-Dimensional Histogram Ordering 170 7.7 Two-Dimensional Scatterplot Ordering 174 7.8 Conclusion 181 References 182 8 Statistical Models, Inference, and Algorithms for Large Biological Data Analysis 185 8.1 Introduction 185 8.2 Statistical/Probabilistic Models 187 8.3 Estimation Methods 189 8.4 Numerical Algorithms 191 8.5 Examples 192 8.6 Conclusion 198 References 199 9 Experimental Designs on High-throughput Biological Experiments 201 9.1 Randomization 201 9.2 Replication 202 9.3 Pooling 209 9.4 Blocking 210 9.5 Design for Classifications 214 9.6 Design for Time Course Experiments 215 9.7 Design for eQTL Studies 215 References 216 10 Statistical Resampling Techniques for Large Biological Data Analysis 219 10.1 Introduction 219 10.2 Resampling Methods for Prediction Error Assessment and Model Selection 221 10.3 Feature Selection 225 10.4 Resampling-Based Classification Algorithms 226 10.5 Practical Example: Lymphoma 226 10.6 Resampling Methods 227 10.7 Bootstrap Methods 232 10.8 Sample Size Issues 233 10.9 Loss Functions 235 10.10 Bootstrap Resampling for Quantifying Uncertainty 236 10.11 Markov Chain Monte Carlo Methods 238 10.12 Conclusions 240 References 247 11 Statistical Network Analysis for Biological Systems And Pathways 249 11.1 Introduction 249 11.2 Boolean Network Modeling 250 11.3 Bayesian Belief Network 259 11.4 Modeling of Metabolic Networks 273 References 279 12 Trends and Statistical Challenges in Genomewide Association Studies 283 12.1 Introduction 283 12.2 Alleles, Linkage Disequilibrium, and Haplotype 283 12.3 International HapMap Project 285 12.4 Genotyping Platforms 286 12.5 Overview of Current GWAS Results 287 12.6 Statistical Issues in GWAS 290 12.7 Haplotype Analysis 296 12.8 Homozygosity and Admixture Mapping 298 12.9 Gene Gene and Gene Environment Interactions 298 12.10 Gene and Pathway-Based Analysis 299 12.11 Disease Risk Estimates 301 12.12 Meta-Analysis 301 12.13 Rare Variants and Sequence-Based Analysis 302 12.14 Conclusions 302 Acknowledgments 303 References 303 13 R and Bioconductor Packages in Bioinformatics: Towards Systems Biology 309 13.1 Introduction 309 13.2 Brief overview of the Bioconductor Project 310 13.3 Experimental Data 311 13.4 Annotation 318 13.5 Models of Biological Systems 328 13.6 Conclusion 335 13.7 Acknowledgments 336 References 336 Index 339