Jae K. Lee, Ph.D., is a professor of biostatistics and epidemiology in the Department of Health Evaluation Sciences at the University of Virginia School of Medicine, where he designed and teaches a course on Statistical Bioinformatics in Medicine. He earned his doctorate in statistical genetics from the University of Wisconsin, Madison. He was previously a research scientist in the Laboratory of Molecular Pharmacology, National Cancer Institute. Among his current research interests is the integration of statistical and genomic information for the analysis of microarray data.

• Preface xiContributors xiii1 Road to Statistical Bioinformatics 1Challenge 1: Multiple-Comparisons Issue 1Challenge 2: High-Dimensional Biological Data 2Challenge 3: Small-n and Large-p Problem 3Challenge 4: Noisy High-Throughput Biological Data 3Challenge 5: Integration of Multiple, Heterogeneous Biological Data Information 3References 52 Probability Concepts and Distributions for Analyzing Large Biological Data 72.1 Introduction 72.2 Basic Concepts 82.3 Conditional Probability and Independence 102.4 Random Variables 132.5 Expected Value and Variance 152.6 Distributions of Random Variables 192.7 Joint and Marginal Distribution 392.8 Multivariate Distribution 422.9 Sampling Distribution 462.10 Summary 543 Quality Control of High-throughput Biological Data 573.1 Sources of Error in High-Throughput Biological Experiments 573.2 Statistical Techniques for Quality Control 593.3 Issues Specific to Microarray Gene Expression Experiments 663.4 Conclusion 69References 694 Statistical Testing and Significance for Large Biological Data Analysis 714.1 Introduction 714.2 Statistical Testing 724.3 Error Controlling 784.4 Real Data Analysis 814.5 Concluding Remarks 87Acknowledgments 87References 885 Clustering: Unsupervised Learning in Large Biological Data 895.1 Measures of Similarity 905.2 Clustering 995.3 Assessment of Cluster Quality 1155.4 Conclusion 123References 1236 Classification: Supervised Learning with High-dimensional Biological Data 1296.1 Introduction 1296.2 Classification and Prediction Methods 1326.3 Feature Selection and Ranking 1406.4 Cross-Validation 1446.5 Enhancement of Class Prediction by Ensemble Voting Methods 1456.6 Comparison of Classification Methods Using High-Dimensional Data 1476.7 Software Examples for Classification Methods 150References 1547 Multidimensional Analysis and Visualization on Large Biomedical Data 1577.1 Introduction 1577.2 Classical Multidimensional Visualization Techniques 1587.3 Two-Dimensional Projections 1617.4 Issues and Challenges 1657.5 Systematic Exploration of Low-Dimensional Projections 1667.6 One-Dimensional Histogram Ordering 1707.7 Two-Dimensional Scatterplot Ordering 1747.8 Conclusion 181References 1828 Statistical Models, Inference, and Algorithms for Large Biological Data Analysis 1858.1 Introduction 1858.2 Statistical/Probabilistic Models 1878.3 Estimation Methods 1898.4 Numerical Algorithms 1918.5 Examples 1928.6 Conclusion 198References 1999 Experimental Designs on High-throughput Biological Experiments 2019.1 Randomization 2019.2 Replication 2029.3 Pooling 2099.4 Blocking 2109.5 Design for Classifications 2149.6 Design for Time Course Experiments 2159.7 Design for eQTL Studies 215References 21610 Statistical Resampling Techniques for Large Biological Data Analysis 21910.1 Introduction 21910.2 Resampling Methods for Prediction Error Assessment and Model Selection 22110.3 Feature Selection 22510.4 Resampling-Based Classification Algorithms 22610.5 Practical Example: Lymphoma 22610.6 Resampling Methods 22710.7 Bootstrap Methods 23210.8 Sample Size Issues 23310.9 Loss Functions 23510.10 Bootstrap Resampling for Quantifying Uncertainty 23610.11 Markov Chain Monte Carlo Methods 23810.12 Conclusions 240References 24711 Statistical Network Analysis for Biological Systems And Pathways 24911.1 Introduction 24911.2 Boolean Network Modeling 25011.3 Bayesian Belief Network 25911.4 Modeling of Metabolic Networks 273References 27912 Trends and Statistical Challenges in Genomewide Association Studies 28312.1 Introduction 28312.2 Alleles, Linkage Disequilibrium, and Haplotype 28312.3 International HapMap Project 28512.4 Genotyping Platforms 28612.5 Overview of Current GWAS Results 28712.6 Statistical Issues in GWAS 29012.7 Haplotype Analysis 29612.8 Homozygosity and Admixture Mapping 29812.9 Gene Gene and Gene Environment Interactions 29812.10 Gene and Pathway-Based Analysis 29912.11 Disease Risk Estimates 30112.12 Meta-Analysis 30112.13 Rare Variants and Sequence-Based Analysis 30212.14 Conclusions 302Acknowledgments 303References 30313 R and Bioconductor Packages in Bioinformatics: Towards Systems Biology 30913.1 Introduction 30913.2 Brief overview of the Bioconductor Project 31013.3 Experimental Data 31113.4 Annotation 31813.5 Models of Biological Systems 32813.6 Conclusion 33513.7 Acknowledgments 336References 336Index 339

Jae K. Lee, Ph.D., is a professor of biostatistics and epidemiology in the Department of Health Evaluation Sciences at the University of Virginia School of Medicine, where he designed and teaches a course on Statistical Bioinformatics in Medicine. He earned his doctorate in statistical genetics from the University of Wisconsin, Madison. He was previously a research scientist in the Laboratory of Molecular Pharmacology, National Cancer Institute. Among his current research interests is the integration of statistical and genomic information for the analysis of microarray data.

• Preface xiContributors xiii1 Road to Statistical Bioinformatics 1Challenge 1: Multiple-Comparisons Issue 1Challenge 2: High-Dimensional Biological Data 2Challenge 3: Small-n and Large-p Problem 3Challenge 4: Noisy High-Throughput Biological Data 3Challenge 5: Integration of Multiple, Heterogeneous Biological Data Information 3References 52 Probability Concepts and Distributions for Analyzing Large Biological Data 72.1 Introduction 72.2 Basic Concepts 82.3 Conditional Probability and Independence 102.4 Random Variables 132.5 Expected Value and Variance 152.6 Distributions of Random Variables 192.7 Joint and Marginal Distribution 392.8 Multivariate Distribution 422.9 Sampling Distribution 462.10 Summary 543 Quality Control of High-throughput Biological Data 573.1 Sources of Error in High-Throughput Biological Experiments 573.2 Statistical Techniques for Quality Control 593.3 Issues Specific to Microarray Gene Expression Experiments 663.4 Conclusion 69References 694 Statistical Testing and Significance for Large Biological Data Analysis 714.1 Introduction 714.2 Statistical Testing 724.3 Error Controlling 784.4 Real Data Analysis 814.5 Concluding Remarks 87Acknowledgments 87References 885 Clustering: Unsupervised Learning in Large Biological Data 895.1 Measures of Similarity 905.2 Clustering 995.3 Assessment of Cluster Quality 1155.4 Conclusion 123References 1236 Classification: Supervised Learning with High-dimensional Biological Data 1296.1 Introduction 1296.2 Classification and Prediction Methods 1326.3 Feature Selection and Ranking 1406.4 Cross-Validation 1446.5 Enhancement of Class Prediction by Ensemble Voting Methods 1456.6 Comparison of Classification Methods Using High-Dimensional Data 1476.7 Software Examples for Classification Methods 150References 1547 Multidimensional Analysis and Visualization on Large Biomedical Data 1577.1 Introduction 1577.2 Classical Multidimensional Visualization Techniques 1587.3 Two-Dimensional Projections 1617.4 Issues and Challenges 1657.5 Systematic Exploration of Low-Dimensional Projections 1667.6 One-Dimensional Histogram Ordering 1707.7 Two-Dimensional Scatterplot Ordering 1747.8 Conclusion 181References 1828 Statistical Models, Inference, and Algorithms for Large Biological Data Analysis 1858.1 Introduction 1858.2 Statistical/Probabilistic Models 1878.3 Estimation Methods 1898.4 Numerical Algorithms 1918.5 Examples 1928.6 Conclusion 198References 1999 Experimental Designs on High-throughput Biological Experiments 2019.1 Randomization 2019.2 Replication 2029.3 Pooling 2099.4 Blocking 2109.5 Design for Classifications 2149.6 Design for Time Course Experiments 2159.7 Design for eQTL Studies 215References 21610 Statistical Resampling Techniques for Large Biological Data Analysis 21910.1 Introduction 21910.2 Resampling Methods for Prediction Error Assessment and Model Selection 22110.3 Feature Selection 22510.4 Resampling-Based Classification Algorithms 22610.5 Practical Example: Lymphoma 22610.6 Resampling Methods 22710.7 Bootstrap Methods 23210.8 Sample Size Issues 23310.9 Loss Functions 23510.10 Bootstrap Resampling for Quantifying Uncertainty 23610.11 Markov Chain Monte Carlo Methods 23810.12 Conclusions 240References 24711 Statistical Network Analysis for Biological Systems And Pathways 24911.1 Introduction 24911.2 Boolean Network Modeling 25011.3 Bayesian Belief Network 25911.4 Modeling of Metabolic Networks 273References 27912 Trends and Statistical Challenges in Genomewide Association Studies 28312.1 Introduction 28312.2 Alleles, Linkage Disequilibrium, and Haplotype 28312.3 International HapMap Project 28512.4 Genotyping Platforms 28612.5 Overview of Current GWAS Results 28712.6 Statistical Issues in GWAS 29012.7 Haplotype Analysis 29612.8 Homozygosity and Admixture Mapping 29812.9 Gene Gene and Gene Environment Interactions 29812.10 Gene and Pathway-Based Analysis 29912.11 Disease Risk Estimates 30112.12 Meta-Analysis 30112.13 Rare Variants and Sequence-Based Analysis 30212.14 Conclusions 302Acknowledgments 303References 30313 R and Bioconductor Packages in Bioinformatics: Towards Systems Biology 30913.1 Introduction 30913.2 Brief overview of the Bioconductor Project 31013.3 Experimental Data 31113.4 Annotation 31813.5 Models of Biological Systems 32813.6 Conclusion 33513.7 Acknowledgments 336References 336Index 339

Jae K. Lee, Ph.D., is a professor of biostatistics and epidemiology in the Department of Health Evaluation Sciences at the University of Virginia School of Medicine, where he designed and teaches a course on Statistical Bioinformatics in Medicine. He earned his doctorate in statistical genetics from the University of Wisconsin, Madison. He was previously a research scientist in the Laboratory of Molecular Pharmacology, National Cancer Institute. Among his current research interests is the integration of statistical and genomic information for the analysis of microarray data.

• Preface xiContributors xiii1 Road to Statistical Bioinformatics 1Challenge 1: Multiple-Comparisons Issue 1Challenge 2: High-Dimensional Biological Data 2Challenge 3: Small-n and Large-p Problem 3Challenge 4: Noisy High-Throughput Biological Data 3Challenge 5: Integration of Multiple, Heterogeneous Biological Data Information 3References 52 Probability Concepts and Distributions for Analyzing Large Biological Data 72.1 Introduction 72.2 Basic Concepts 82.3 Conditional Probability and Independence 102.4 Random Variables 132.5 Expected Value and Variance 152.6 Distributions of Random Variables 192.7 Joint and Marginal Distribution 392.8 Multivariate Distribution 422.9 Sampling Distribution 462.10 Summary 543 Quality Control of High-throughput Biological Data 573.1 Sources of Error in High-Throughput Biological Experiments 573.2 Statistical Techniques for Quality Control 593.3 Issues Specific to Microarray Gene Expression Experiments 663.4 Conclusion 69References 694 Statistical Testing and Significance for Large Biological Data Analysis 714.1 Introduction 714.2 Statistical Testing 724.3 Error Controlling 784.4 Real Data Analysis 814.5 Concluding Remarks 87Acknowledgments 87References 885 Clustering: Unsupervised Learning in Large Biological Data 895.1 Measures of Similarity 905.2 Clustering 995.3 Assessment of Cluster Quality 1155.4 Conclusion 123References 1236 Classification: Supervised Learning with High-dimensional Biological Data 1296.1 Introduction 1296.2 Classification and Prediction Methods 1326.3 Feature Selection and Ranking 1406.4 Cross-Validation 1446.5 Enhancement of Class Prediction by Ensemble Voting Methods 1456.6 Comparison of Classification Methods Using High-Dimensional Data 1476.7 Software Examples for Classification Methods 150References 1547 Multidimensional Analysis and Visualization on Large Biomedical Data 1577.1 Introduction 1577.2 Classical Multidimensional Visualization Techniques 1587.3 Two-Dimensional Projections 1617.4 Issues and Challenges 1657.5 Systematic Exploration of Low-Dimensional Projections 1667.6 One-Dimensional Histogram Ordering 1707.7 Two-Dimensional Scatterplot Ordering 1747.8 Conclusion 181References 1828 Statistical Models, Inference, and Algorithms for Large Biological Data Analysis 1858.1 Introduction 1858.2 Statistical/Probabilistic Models 1878.3 Estimation Methods 1898.4 Numerical Algorithms 1918.5 Examples 1928.6 Conclusion 198References 1999 Experimental Designs on High-throughput Biological Experiments 2019.1 Randomization 2019.2 Replication 2029.3 Pooling 2099.4 Blocking 2109.5 Design for Classifications 2149.6 Design for Time Course Experiments 2159.7 Design for eQTL Studies 215References 21610 Statistical Resampling Techniques for Large Biological Data Analysis 21910.1 Introduction 21910.2 Resampling Methods for Prediction Error Assessment and Model Selection 22110.3 Feature Selection 22510.4 Resampling-Based Classification Algorithms 22610.5 Practical Example: Lymphoma 22610.6 Resampling Methods 22710.7 Bootstrap Methods 23210.8 Sample Size Issues 23310.9 Loss Functions 23510.10 Bootstrap Resampling for Quantifying Uncertainty 23610.11 Markov Chain Monte Carlo Methods 23810.12 Conclusions 240References 24711 Statistical Network Analysis for Biological Systems And Pathways 24911.1 Introduction 24911.2 Boolean Network Modeling 25011.3 Bayesian Belief Network 25911.4 Modeling of Metabolic Networks 273References 27912 Trends and Statistical Challenges in Genomewide Association Studies 28312.1 Introduction 28312.2 Alleles, Linkage Disequilibrium, and Haplotype 28312.3 International HapMap Project 28512.4 Genotyping Platforms 28612.5 Overview of Current GWAS Results 28712.6 Statistical Issues in GWAS 29012.7 Haplotype Analysis 29612.8 Homozygosity and Admixture Mapping 29812.9 Gene Gene and Gene Environment Interactions 29812.10 Gene and Pathway-Based Analysis 29912.11 Disease Risk Estimates 30112.12 Meta-Analysis 30112.13 Rare Variants and Sequence-Based Analysis 30212.14 Conclusions 302Acknowledgments 303References 30313 R and Bioconductor Packages in Bioinformatics: Towards Systems Biology 30913.1 Introduction 30913.2 Brief overview of the Bioconductor Project 31013.3 Experimental Data 31113.4 Annotation 31813.5 Models of Biological Systems 32813.6 Conclusion 33513.7 Acknowledgments 336References 336Index 339