.

0 (hodnocen0 x )

BK

2nd ed.

Boca Raton : CRC Press/Taylor & Francis, c2012

xlviii, 1042 s. : il. ; 24 cm + 1 CD-ROM

ISBN 978-1-4398-0975-4 (váz.)

Obsahuje bibliografii na s. 981-1025 a rejstřík

000248991

Contents // List of Figures xxv // List of Tables xxxv // Preface xxxix // 1 Introduction 1 // 1.1 Bioinformatics - an emerging discipline ... 1 // 2 The cell and its basic mechanisms 5 // 2.1 The cell... 5 // 2.2 The building blocks of genomic information... 13 // 2.2.1 The deoxyribonucleic acid (DNA) ... 13 // 2.2.2 The DNA as a language... 19 // 2.2.3 Errors in the DNA language... 23 // 2.2.4 Other useful concepts... 24 // 2.3 Expression of genetic information ... 28 // 2.3.1 Transcription... 30 // 2.3.2 Translation... 32 // 2.3.3 Gene regulation... 35 // 2.4 The need for high-throughput methods ... 36 // 2.5 Summary... 37 // 3 Microarrays 39 // 3.1 Microarrays - tools for gene expression analysis ... 39 // 3.2 Fabrication of microarrays ... 41 // 3.2.1 Deposition ... 41 // 3.2.1.1 The Illumina technology ... 42 // 3.2.2 In situ synthesis... 48 // 3.2.3 A brief comparison of cDNA and oligonucleotide technologies ... 55 // 3.3 Applications of microarrays... 57 // 3.4 Challenges in using microarrays in gene expression studies . 58 // ix // x // 3.5 Sources of variability // 3.6 Summary... // Contents // 63 // 67 // 4 Reliability and reproducibility issues in DNA microarray // measurements 69 // 4.1 Introduction ... 69 // 4.2 What is expected from microarrays?... 70 // 4.3 Basic considerations of microarray measurements ... 70 // 4.4 Sensitivity ... 72 // 4.5 Accuracy ... 73 // 4.6 Reproducibility ... 77 // 4.7 Cross-platform consistency ... 78 // 4.8 Sources of inaccuracy and

inconsistencies in microarray measurements ... 82 // 4.9 The MicroArray Quality Control (MAQC) project... 85 // 4.10 Summary... 87 // 5 Image processing 89 // 5.1 Introduction ... 89 // 5.2 Basic elements of digital imaging... 90 // 5.3 Microarray image processing ... 95 // 5.4 Image processing of cDNA microarrays ... 96 // 5.4.1 Spot finding... 99 // 5.4.2 Image segmentation... 100 // 5.4.3 Quantification ... 106 // 5.4.4 Spot quality assessment... Ill // 5.5 Image processing of Affymetrix arrays... 113 // 5.6 Summary... 115 // 6 Introduction to R 119 // 6.1 Introduction to R ... 119 // 6.1.1 About R and Bioconductor... 119 // 6.1.2 Repositories for R and Bioconductor... 120 // 6.1.3 The working setup for R... 121 // 6.1.4 Getting help in R... 122 // 6.2 The basic concepts... 122 // 6.2.1 Elementary computations... 122 // 6.2.2 Variables and assignments... 125 // 6.2.3 Expressions and objects... 126 // 6.3 Data structures and functions ... 128 // 6.3.1 Vectors and vector operations... 128 // 6.3.2 Referencing vector elements... 131 // Contents xi // 6.3.3 Functions... 133 // 6.3.4 Creating vectors... 135 // 6.3.5 Matrices... 137 // 6.3.6 Lists... 141 // 6.3.7 Data frames... 141 // 6.4 Other capabilities ... 144 // 6.4.1 More advanced indexing... 144 // 6.4.2 Missing values ... 145 // 6.4.3 Reading and writing files ... 148 // 6.4.4 Conditional selection and indexing... 150 // 6.4.5 Sorting ... 151 // 6.4.6 Implicit loops... 154 // 6.5 The R environment ... 159 // 6.5.1 The search

path: attach and detach... 159 // 6.5.2 The workspace... 161 // 6.5.3 Packages ... 163 // 6.5.4 Built-in data... 165 // 6.6 Installing Bioconductor ... 165 // 6.7 Graphics ... 167 // 6.8 Control structures in R ... 169 // 6.8.1 Conditional statements ... 170 // 6.8.2 Pre-test loops... 171 // 6.8.3 Counting loops... 172 // 6.8.4 Breaking out of loops ... 173 // 6.8.5 Post-test loops... 173 // 6.9 Programming in R versus C/C++/Java... 174 // 6.9.1 R is “forgiving” - which can be bad... 174 // 6.9.2 Weird syntax errors... 175 // 6.9.3 Programming style... 179 // 6.10 Summary... ... 182 // 6.11 Solved Exercises... 183 // 6.12 Exercises ... 191 // 7 Bioconductor: principles and illustrations 193 // 7.1 Overview... 193 // 7.2 The portal ... 194 // 7.2.1 The main resource categories... 195 // 7.2.2 Working with the software repository ... 195 // 7.3 Some explorations and analyses ... 197 // 7.3.1 The representation of microarray data... 197 // 7.3.2 The annotation of a microarray platform ... 199 // 7.3.3 Predictive modeling using microarray data ... 203 // 7.4 Summary... 205 // Xii Contents // 8 Elements of statistics 207 // 8.1 Introduction ... 207 // 8.2 Some basic concepts... 208 // 8.2.1 Populations versus samples... 208 // 8.2.2 Parameters versus statistics... 209 // 8.3 Elementary statistics ... 211 // 8.3.1 Measures of central tendency: mean, mode, and median 211 // 8.3.1.1 Mean... 211 // 8.3.1.2 Mode... 212 // 8.3.1.3 Median, percentiles, and quantiles... 213 // 8.3.1.4

Characteristics of the mean, mode, and median ... 214 // 8.3.2 Measures of variability... 215 // 8.3.2.1 Range... 215 // 8.3.2.2 Variance... 216 // 8.3.3 Some interesting data manipulations... 218 // 8.3.4 Covariance and correlation... 219 // 8.3.5 Interpreting correlations... 223 // 8.3.6 Measurements, errors, and residuals... 230 // 8.4 Degrees of freedom ... 231 // 8.4.1 Degrees of freedom as independent error estimates . . 232 // 8.4.2 Degrees of freedom as number of additional measurements ... 233 // 8.4.3 Degrees of freedom as observations minus restrictions 233 // 8.4.4 Degrees of freedom as measurements minus model parameters ... 234 // 8.4.5 Degrees of freedom as number of measurements we can // change... 234 // 8.4.6 Data split between estimating variability and model parameters ... 235 // 8.4.7 A geometrical perspective... 235 // 8.4.8 Calculating the number of degrees of freedom... 236 // 8.4.8.1 Estimating ? quantities from n measurements 236 // 8.4.9 Calculating the degrees of freedom for an n x m table . 237 // 8.5 Probabilities ... 241 // 8.5.1 Computing with probabilities... 243 // 8.5.1.1 Addition rule... 243 // 8.5.1.2 Conditional probabilities... 244 // 8.5.1.3 General multiplication rule... 247 // 8.6 Bayes’ theorem... 247 // 8.7 Testing for (or predicting) a disease ... 250 // 8.7.1 Basic criteria: accuracy, sensitivity, specificity, PPV, // NPV... 251 // Contents // xiii // 8.7.2 More about classification criteria: prevalence, incidence, // and various

interdependencies... 253 // 8.8 Summary... 257 // 8.9 Solved problems ... 257 // 8.10 Exercises ... 258 // 9 Probability distributions 261 // 9.1 Probability distributions ... 261 // 9.1.1 Discrete random variables... 262 // 9.1.2 The discrete uniform distribution... 265 // 9.1.3 Binomial distribution ... 266 // 9.1.4 Poisson distribution... 275 // 9.1.5 The hypergeometric distribution... 278 // 9.1.6 Continuous random variables... 281 // 9.1.7 The continuous uniform distribution... 282 // 9.1.8 The normal distribution... 283 // 9.1.9 Using a distribution... 287 // 9.2 Central limit theorem... 291 // 9.3 Are replicates useful? ... 292 // 9.4 Summary... 294 // 9.5 Solved problems ... 295 // 9.6 Exercises ... 296 // 10 Basic statistics in R 299 // 10.1 Introduction ... 299 // 10.2 Descriptive statistics in R... 300 // 10.2.1 Mean, median, range, variance, and standard deviation 300 // 10.2.2 Mode... 304 // 10.2.3 More built-in R functions for descriptive statistics . . 305 // 10.2.4 Covariance and correlation... 307 // 10.3 Probabilities and distributions in R ... 308 // 10.3.1 Sampling... 308 // 10.3.2 Empirical probabilities... 309 // 10.3.3 Standard distributions in R... 315 // 10.3.4 Generating (pseudo-)random numbers... 316 // 10.3.5 Probability density functions... 316 // 10.3.6 Cumulative distribution functions ... 317 // 10.3.7 Quantiles... 319 // 10.3.7.1 The normal distribution... 321 // 10.3.7.2 The binomial distribution... 324 // 10.3.8 Using built-in distributions in

R... 326 // 10.4 Central limit theorem... 330 // 10.5 Summary... 336 // Xiv Contents // 10.6 Exercises ... 337 // 11 Statistical hypothesis testing 339 // 11.1 Introduction ... 339 // 11.2 The framework... 340 // 11.3 Hypothesis testing and significance ... 343 // 11.3.1 One-tailed testing ... 344 // 11.3.2 Two-tailed testing... 348 // 11.4 “I do not believe God does not exist” ... 350 // 11.5 An algorithm for hypothesis testing ... 352 // 11.6 Errors in hypothesis testing... 353 // 11.7 Summary... 357 // 11.8 Solved problems ... 358 // 12 Classical approaches to data analysis 361 // 12.1 Introduction ... 361 // 12.2 Tests involving a single sample... 362 // 12.2.1 Tests involving the mean. The t distribution... 362 // 12.2.2 Choosing the number of replicates... 368 // 12.2.3 Tests involving the variance (a2). The chi-square distribution ... 372 // 12.2.4 Confidence intervals for standard deviation/variance . 376 // 12.3 Tests involving two samples... 377 // 12.3.1 Comparing variances. The F distribution... 377 // 12.3.2 Comparing means... 383 // 12.3.2.1 Equal variances... 386 // 12.3.2.2 Unequal variances... 389 // 12.3.2.3 Paired testing... 389 // 12.3.3 Confidence intervals for the difference of means ii\ — fi2 390 // 12.4 Summary... 39I // 12.5 Exercises ... 394 // 13 Analysis of Variance - ANOVA 395 // 13.1 Introduction ... 395 // 13.1.1 Problem definition and model assumptions... 395 // 13.1.2 The “dot” notation... 399 // 13.2 One-way ANOVA ... 400 // 13.2.1 One-way

Model I ANOVA... 400 // 13.2.1.1 Partitioning the Sum of Squares... 401 // 13.2.1.2 Degrees of freedom... 403 // 13.2.1.3 Testing the hypotheses... 403 // 13.2.2 One-way Model II ANOVA... 407 // Contents // XV // 13.3 Two-way ANOVA ... 410 // 13.3.1 Randomized complete block design ANOVA... 411 // 13.3.2 Comparison between one-way ANOVA and randomized // block design ANOVA ... 414 // 13.3.3 Some examples... 415 // 13.3.4 Factorial design two-way ANOVA ... 419 // 13.3.5 Data analysis plan for factorial design ANOVA ... 424 // 13.3.6 Reference formulae for factorial design ANOVA ... 425 // 13.4 Quality control... 425 // 13.5 Summary... 428 // 13.6 Exercises ... 429 // 14 Linear models in R 433 // 14.1 Introduction and model formulation... // 14.2 Fitting linear models in R ... // 14.3 Extracting information from a fitted model: testing hypotheses // and making predictions ... // 14.4 Some limitations of linear models ... // 14.5 Dealing with multiple predictors and interactions in the linear // models, and interpreting model coefficients ... // 14.5.1 Details on the design matrix creation and coefficients // estimation in linear models... // 14.5.2 ANOVA using linear models... // 14.5.2.1 One-way Model I ANOVA ... // 14.5.2.2 Randomized block design ANOVA... // 14.5.3 Practical linear models for analysis of microarray data // 14.5.4 A two-group comparison gene expression analysis using // a simple t-test... // 14.5.5 Differential expression using the limma library of Bioconductor

... // 14.5.5.1 Two group comparison with single-channel // data... // 14.5.5.2 Multiple contrasts with single-channel data . // 14.6 Summary... // 433 // 435 // 439 // 440 // 443 // 445 // 447 // 448 // 453 // 454 // 455 457 // 457 // 459 // 460 // 15 Experiment design // 463 // 15.1 The concept of experiment design ... 464 // 15.2 Comparing varieties ... 464 // 15.3 Improving the production process ... 466 // 15.4 Principles of experimental design... 468 // 15.4.1 Replication... 468 // 15.4.2 Randomization... 471 // 15.4.3 Blocking... 472 // xvi // Contents // 15.5 Guidelines for experimental design... 472 // 15.6 A short synthesis of statistical experiment designs ... 474 // 15.6.1 The fixed effect design... 475 // 15.6.2 Randomized block design... 47g // 15.6.3 Balanced incomplete block design... 476 // 15.6.4 Latin square design ... 477 // 15.6.5 Factorial design... 477 // 15.6.6 Confounding in the factorial design... 480 // 15.7 Some microarray specific experiment designs ... 481 // 15.7.1 The Jackson Lab approach... 481 // 15.7.2 Ratios and flip-dye experiments... 484 // 15.7.3 Reference design versus loop design... 486 // 15.8 Summary... 4gg // 16 Multiple comparisons 494 // 16.1 Introduction ... 492 // 16.2 The problem of multiple comparisons ... 492 // 16.3 A more precise argument ... 499 // 16.4 Corrections for multiple comparisons ... 501 // 16.4.1 The Sidák correction... 501 // 16.4.2 The Bonferroni correction... 502 // 16.4.3 Holm’s step-wise correction... 503

// 16.4.4 The false discovery rate (FDR)... 504 // 16.4.5 Permutation correction ... 505 // 16.4.6 Significance analysis of microarrays (SAM)... 507 // 16.4.7 On permutation-based methods... 508 // 16.5 Corrections for multiple comparisons in R... 508 // 16.6 Summary... 523 // 17 Analysis and visualization tools 515 // 17.1 Introduction ... 525 // 17.2 Box plots... 52g // 17.3 Gene pies... 52q // 17.4 Scatter plots ... 522 // 17.4.1 Scatter plots in R ... 525 // 17.4.2 Scatter plot limitations ... 526 // 17.4.3 Scatter plot summary... 529 // 17.5 Volcano plots... 529 // 17.5.1 Volcano plots in R... 539 // 17.6 Histograms... 533 // 17.6.1 Histograms summary... 542 // 17.7 Time series... 542 // 17.8 Time series plots in R... 543 // Contents xvii // 17.9 Principal component analysis (PCA) ... 550 // 17.9.1 PCA limitations... 558 // 17.9.2 Principal component analysis in R... 559 // 17.9.3 PCA summary... 560 // 17.10 Independent component analysis (ICA)... 563 // 17.11 Summary ... 564 // 18 Cluster analysis 567 // 18.1 Introduction ... 567 // 18.2 Distance metric ... 568 // 18.2.1 Euclidean distance... 569 // 18.2.2 Manhattan distance... 570 // 18.2.3 Chebychev distance ... 572 // 18.2.4 Angle between vectors... 573 // 18.2.5 Correlation distance... 573 // 18.2.6 Squared Euclidean distance... 574 // 18.2.7 Standardized Euclidean distance... 575 // 18.2.8 Mahalanobis distance... 577 // 18.2.9 Minkowski distance... 577 // 18.2.10 When to use what distance... 578 // 18.2.11 A comparison

of various distances... 580 // 18.3 Clustering algorithms ... 581 // 18.3.1 fc-means clustering... 584 // 18.3.1.1 Characteristics of the A:-means clustering . . 586 // 18.3.1.2 Cluster quality assessment... 588 // 18.3.1.3 Number of clusters in Ł-means... 593 // 18.3.1.4 Algorithm complexity... 593 // 18.3.2 Hierarchical clustering... 594 // 18.3.2.1 Inter-cluster distances and algorithm complexity ... 596 // 18.3.2.2 Top-down versus bottom-up... 597 // 18.3.2.3 Cutting tree diagrams... 599 // 18.3.2.4 An illustrative example... 601 // 18.3.2.5 Hierarchical clustering summary... 603 // 18.3.3 Kohonen maps or self-organizing feature maps (SOFM) 605 // 18.4 Partitioning around medoids (PAM)... 614 // 18.5 Biclustering ... 616 // 18.5.1 Types of biclusters... 617 // 18.5.2 Biclustering algorithms ... 619 // 18.5.3 Differential biclustering... 620 // 18.5.4 Biclustering summary... 621 // 18.6 Clustering in R... 621 // 18.6.1 Partition around medoids (PAM) in R... 629 // 18.6.2 Biclustering in R... 632 // xviii // Contents // 18.7 Summary... 632 // 19 Quality control 635 // 19.1 Introduction ... 635 // 19.2 Quality control for Affymetrix data ... 636 // 19.2.1 Reading raw data (.CEL files)... 636 // 19.2.2 Intensity distributions... 637 // 19.2.3 Box plots... 639 // 19.2.4 Probe intensity images... 639 // 19.2.5 Quality control metrics ... 641 // 19.2.6 RNA degradation curves... 647 // 19.2.7 Quality control plots... 649 // 19.2.8 Probe-level model (PLM) fitting. RLE and NUSE plots

654 // 19.3 Quality control of Illumina data ... 660 // 19.3.1 Reading Illumina data... 660 // 19.3.2 Bead-summary data... 663 // 19.3.2.1 Raw probe data import, visualization, and // quality assessment using “beadarray” ... 663 // 19.3.2.2 Raw probe data import, visualization, and // quality assessment using “lumi”... 665 // 19.3.3 Bead-level data... 669 // 19.3.3.1 Raw bead data import and assessment . . . 669 // 19.3.3.2 Summarizing from bead-level to probe-level // data... 690 // 19.4 Summary... 691 // 20 Data preprocessing and normalization 693 // 20.1 Introduction ... 693 // 20.2 General preprocessing techniques... 694 // 20.2.1 The log transform... 694 // 20.2.2 Combining replicates and eliminating outliers... 696 // 20.2.3 Array normalization... 698 // 20.2.3.1 Dividing by the array mean... 701 // 20.2.3.2 Subtracting the mean... 701 // 20.2.3.3 Using control spots/genes... 703 // 20.2.3.4 Iterative linear regression... 703 // 20.2.3.5 Other aspects of array normalization ... 704 // 20.3 Normalization issues specific to cDNA data... 704 // 20.3.1 Background correction... 704 // 20.3.1.1 Local background correction... 704 // 20.3.1.2 Sub-grid background correction ... 705 // 20.3.1.3 Group background correction... 705 // 20.3.1.4 Background correction using blank spots . . 705 // Contents xix // 20.3.1.5 Background correction using control spots . 705 // 20.3.2 Other spot level preprocessing ... 706 // 20.3.3 Color normalization... 706 // 20.3.3.1 Curve fitting

and correction ... 708 // 20.3.3.2 LOWESS/LOESS normalization... 710 // 20.3.3.3 Piece-wise normalization... 713 // 20.3.3.4 Other approaches to cDNA data normalization 715 // 20.4 Normalization issues specific to Affymetrix data ... 715 // 20.4.1 Background correction... 715 // 20.4.2 Signal calculation... 718 // 20.4.2.1 Ideal mismatch... 718 // 20.4.2.2 Probe values... 719 // 20.4.2.3 Scaled probe values... 720 // 20.4.3 Detection calls... 721 // 20.4.4 Relative expression values... 722 // 20.5 Other approaches to the normalization of Affymetrix data . 722 // 20.5.1 Cyclic Loess... 722 // 20.5.2 The model-based dChip approach ... 723 // 20.5.3 The Robust Multi-Array Analysis (RMA)... 724 // 20.5.4 Quantile normalization... 724 // 20.6 Useful preprocessing and normalization sequences ... 727 // 20.7 Normalization procedures in R... 728 // 20.7.1 Normalization functions and procedures for Affymetrix // data... 728 // 20.7.2 Background adjustment and various types of normalization ... 734 // 20.7.3 Summarization... 735 // 20.8 Batch preprocessing... 738 // 20.9 Normalization functions and procedures for Illumina data . . 739 // 20.10Summary... 743 // 20.11 Appendix: A short primer on logarithms ... 746 // 21 Methods for selecting differentially expressed genes 749 // 21.1 Introduction ... 749 // 21.2 Criteria... 751 // 21.3 Fold change ... 753 // 21.3.1 Description... 753 // 21.3.2 Characteristics... 754 // 21.4 Unusual ratio... 756 // 21.4.1 Description... 756 // 21.4.2 Characteristics...

757 // 21.5 Hypothesis testing, corrections for multiple comparisons, and // resampling ... 758 // XX // Contents // 21.5.1 Description... 75g // 21.5.2 Characteristics... 70q // 21.6 ANOVA ... 760 // 21.6.1 Description... 750 // 21.6.2 Characteristics...  61 // 21.7 Noise sampling... 702 // 21.7.1 Description... 752 // 21.7.2 Characteristics... 753 // 21.8 Model-based maximum likelihood estimation methods ... 764 // 21.8.1 Description... 754 // 21.8.2 Characteristics... 75g // 21.9 Affymetrix comparison calls ... 75g // 21.10 Significance Analysis of Microarrays (SAM) ... 769 // 21.11 A moderated t-statistic... 77O // 21.12 Other methods ... 772 // 21.13 Reproducibility ... 772 // 21.14 Selecting differentially expressed (DE) genes in R... 774 // 21.14.1 Data import and preprocessing... 774 // 21.14.2 Fold change... 775 // 21.14.3 Unusual ratio ... 777 // 21.14.4 Hypothesis testing, corrections for multiple comparisons, and resampling... 779 // 21.15 Summary ... 792 // 21.16 Appendix ... 792 // 21.16.1 A comparison of the noise sampling method with the // full-blown ANOVA approach... 792 // 22 The Gene Ontology (GO) 795 // 22.1 Introduction ... 795 // 22.2 The need for an ontology... 795 // 22.3 What is the Gene Ontology (GO)?... 797 // 22.4 What does GO contain?... 79g // 22.4.1 GO structure and data representation... 798 // 22.4.2 Levels of abstraction, traversing the DAG, the “True // Path Rule”... gOQ // 22.4.3 Evidence codes... gOi // 22.4.4 GO coverage... gQ3 // 22.5 Access

to GO ... gQ3 // 22.6 Other related resources ... g06 // 22.7 Summary... gQ7 // Contents xxi // 23 Functional analysis and biological interpretation of microarray data 809 // 23.1 Over-representation analysis (ORA) ... 809 // 23.1.1 Statistical approaches... 811 // 23.2 Onto-Express... 814 // 23.2.1 Implementation... 814 // 23.2.2 Graphical input interface description... 815 // 23.2.3 Some real data analyses... 820 // 23.2.4 Interpretation of the functional analysis results ... 825 // 23.3 Functional class scoring... 826 // 23.4 The Gene Set Enrichment Analysis (GSEA) ... 826 // 23.5 Summary... 827 // 24 Uses, misuses, and abuses in GO profiling 829 // 24.1 Introduction ... 830 // 24.2 “Known unknowns” ... 830 // 24.3 Which way is up? ... 831 // 24.4 Negative annotations ... 834 // 24.5 Common mistakes in functional profiling ... 834 // 24.5.1 Enrichment versus p-values... 834 // 24.5.2 One-sided versus two-sided testing... 835 // 24.5.3 Reference set... 836 // 24.5.4 Correction for multiple comparisons... 837 // 24.6 Using a custom level of abstraction through the GO hierarchy 837 // 24.7 Correlation between GO terms... 840 // 24.8 GO slims and subsets ... 845 // 24.9 Summary... 846 // 25 A comparison of several tools for ontological analysis 849 // 25.1 Introduction ... 849 // 25.2 Existing tools for ontological analysis ... 850 // 25.3 Comparison of existing functional profiling tools ... 863 // 25.3.1 The statistical model... 864 // 25.3.2 The set of reference genes...

865 // 25.3.3 Correction for multiple experiments ... 865 // 25.3.4 The scope of the analysis... 867 // 25.3.5 Performance issues... 867 // 25.3.6 Visualization capabilities ... 867 // 25.3.7 Custom level of abstraction... 871 // 25.3.8 Prerequisites and installation issues ... 872 // 25.3.9 Data sources... 874 // 25.3.10 Supported input IDs... 874 // xxii // Contents // 25.4 Drawbacks and limitations of the current approach ... // 25.5 Summary... // 26 Focused microarrays - comparison and selection // 26.1 Introduction ... // 26.2 Criteria for array selection ... // 26.3 Onto-Compare... // 26.4 Some comparisons... // 26.5 Summary... // 27 ID Mapping issues // 27.1 Introduction ... // 27.2 Name space issues in annotation databases ... // 27.3 A comparison of some ID mapping tools ... // 27.4 Summary... // 28 Pathway analysis // 28.1 Introduction ... // 28.2 Terms and problem definition ... // 28.3 Over-representation and functional class scoring approaches in // pathway analysis... // 28.3.1 Limitations of the ORA and PCS approaches in pathway // analysis... // 28.4 An approach for the analysis of metabolic pathways... // 28.5 An impact analysis of signaling pathways... // 28.5.1 Method description... // 28.5.2 An intuitive perspective on the impact analysis ... // 28.5.3 A statistical perspective on the impact analysis ... // 28.5.4 Calculating the gene perturbations... // 28.5.5 Impact analysis as a generalization of ORA and PCS . // 28.5.5.1 Gene perturbations for genes with

no upstream activity... // 28.5.5.2 Pathway impact analysis when the expression // changes are ignored... // 28.5.5.3 Impact analysis involving genes with no measured expression change... // 28.5.5.4 Impact analysis in the absence of perturbation // propagation ... // 28.5.5.5 Adding a new dimension to the classical approaches ... // 28.5.6 Some results on real data sets... // 28.6 Variations on the impact analysis theme ... // 876 // 878 // 881 // 881 // 883 // 884 // 885 891 // 893 // 893 // 894 898 901 // 903 // 903 // 904 // 910 // 911 914 914 // 914 // 915 917 // 920 // 921 // 921 // 921 // 923 // 924 // 924 // 926 // 935 // Contents // xxiii // 28.6.1 Using the perturbation accumulation... 935 // 28.6.2 Calculating the perturbation p-value using bootstrapping ... 935 // 28.6.3 Combining the two types of evidence with a normal inversion approach... 936 // 28.6.4 Correcting for multiple comparisons... 938 // 28.6.5 Other extensions... 941 // 28.7 Pathway Guide... 941 // 28.7.1 Data visualization capabilities... 943 // 28.7.2 Portability ... 944 // 28.7.3 Export capabilities... 944 // 28.7.4 Custom pathways support... 944 // 28.7.5 Reliability benchmarks: speed, number of FPs, distribution of p-values under the null distribution ... 946 // 28.8 Kinetic models versus impact analysis... 946 // 28.9 Conclusions... 949 // 28.10 Data sets and software availability ... 950 // 29 Machine learning techniques 951 // 29.1 Introduction ... 951 // 29.2 Main concepts and definitions ...

952 // 29.3 Supervised learning ... 955 // 29.3.1 General concepts... 955 // 29.3.2 Error estimation and validation... 956 // 29.3.3 Some types of classifiers... 959 // 29.3.3.1 Quadratic and linear discriminants... 959 // 29.3.3.2  -Nearest neighbor classifier... 960 // 29.3.3.3 Decision trees ... 961 // 29.3.3.4 Neural Networks... 962 // 29.3.3.5 Support vector machines... 964 // 29.3.4 Feature selection... 968 // 29.4 Practicalities using R ... 970 // 29.4.1 A leukemia dataset... 970 // 29.4.2 Supervised methods... 971 // 29.4.3 Variable importance displays... 973 // 29.4.4 Summary... 973 // 30 The road ahead 977 // 30.1 What next?... 977 // Bibliography 981 // Index // 1027