Handbook of Statistical Bioinformatics

Preface
Contents
Part I Single-Cell Analysis
Computational and Statistical Methods for Single-Cell RNA Sequencing Data
1 Introduction
2 Data Preprocessing
2.1 Reads Mapping
2.2 Cell Barcodes Demultiplexing
2.3 UMI Collapsing
2.4 Cell Barcodes Selection
2.5 Summary
3 Data Normalization and Visualization
3.1 Background
3.2 Global Scaling Normalization for UMI Data
3.3 Probabilistic Model-Based Normalization for UMI Data
3.4 Dimension Reduction and Cell Clustering
4 Dropout Imputation
4.1 Background
4.2 Cell-Cell Similarity-Based Imputation
4.3 Gene-Gene Similarity-Based Imputation
4.4 Gene-Gene and Cell-Cell Similarity-Based Imputation
4.5 Deep Neural Network-Based Imputation
4.6 G2S3
4.7 Methods Evaluation and Comparison
5 Differential Expression Analysis
5.1 Background
5.2 DE Methods Ignoring Subject Effects
5.3 DE Methods Considering Subject Effects
5.4 iDESC
5.5 DE Methods Evaluation and Comparison
5.5.1 Type I Error Comparison
5.5.2 Statistical Power Comparison
6 Concluding Remarks
References
Pre-processing, Dimension Reduction, and Clustering for Single-Cell RNA-seq Data
1 Introduction
2 Pre-processing of scRNA-seq Data
2.1 Removal of Batch Effects
2.2 Quality Control and Feature Selection
3 Dimension Reduction and Clustering
3.1 Dimension Reduction
3.2 Clustering
4 Conclusion
References
Integrative Analyses of Single-Cell Multi-Omics Data: A Review from a Statistical Perspective
1 Multi-Omics Data Profiled on Different Cells
2 Multi-Omics Data Profiled on the Same Single Cells
3 Challenges and Future Perspectives
References
Approaches to Marker Gene Identification from Single-Cell RNA-Sequencing Data
1 Introduction
2 Marker Gene Selection Relies on Identifying Differentially Expressed Genes
3 Methods for Marker Gene Selection
3.1 Highest Expressed, Highest Variable
4 Supervised Methods
4.1 COMET
4.2 scGeneFit
5 Unsupervised Methods
5.1 Seurat
5.2 SC3
5.3 SCMarker
5.4 scTIM
5.5 RankCorr
6 Discussion
References
Model-Based Clustering of Single-Cell Omics Data
1 Introduction
2 Single-Cell Transcriptomic Data Clustering
2.1 Single-Cell Transcriptomic Data Structure
2.2 DIMM-SC
2.3 Real Data Example
3 Population-Scale Single-Cell Transcriptomic Data Clustering
3.1 Population-Scale Single-Cell Transcriptomic Data Structure
3.2 BAMM-SC
3.3 Real Data Example
4 Single-Cell Multi-omics Data Clustering
4.1 CITE-seq Data Structure
4.2 BREM-SC
4.3 Real Data Example
5 Concluding Remarks
References
Deep Learning Methods for Single-Cell Omics Data
1 Introduction
2 Factor-Model-Based Deep Learning Approaches
2.1 Regularization and Priors on the Latent Factors
2.1.1 Gaussian Prior and Variational Inference
2.1.2 Adjust for Batch Effects and Confounding Covariates: Identifiability
2.1.3 Adjust for Batch Effects and Confounding Covariates: Implementation
2.1.4 Model Cell Population Structure in the Latent Space
2.2 Distributional Assumptions on Observed Data
2.2.1 Model Observed Data from scRNA-seq
2.2.2 Model Observed Data from scATAC-seq
2.2.3 Model Observed Data from Single-Cell Multiomics Technologies
2.3 Post-training Statistical Analyses
2.3.1 Denoising
2.3.2 Visualization, Clustering, and Trajectory Analysis
2.3.3 Prediction
3 Deep Learning Methods for Dimension Reduction
3.1 Construct the Loss Function
3.2 Extra Penalties and Regularization
4 Discussion
References
Part II Network Analysis
Probabilistic Graphical Models for Gene Regulatory Networks
1 Introduction
2 Probabilistic Graphical Models
2.1 Graphical Model Basics
2.2 Markov Networks
2.3 Bayesian Networks
3 Classic Graphical Models for Reconstructing GRNs
3.1 Frequentist Approach
3.2 Bayesian Approach
3.3 Graphical Models Incorporating Prior Knowledge
4 Testing in Graphical Models
4.1 Parametric Test
4.2 Non-parametric Test for Global Graph Structure
5 Conclusion
References
Additive Conditional Independence for Large and Complex Biological Structures
1 Additive Conditional Independence (ACI)
1.1 Additive Reproducing Kernel Hilbert Spaces and Relevant Linear Operators
2 Variable Selection via ACI
2.1 Nonparametric Variable Selection
2.2 Penalized Least-Square Estimation with RKHS Operators
2.3 Matrix Representation of Operators and Algorithm
2.4 Data Example
3 Graphical Modeling Through ACI
3.1 Nonparametric Graphical Models
3.2 The Additive Conditional Covariance and Partial Correlation Operators
3.3 Operator-Level Estimation and the Algorithm
3.4 Data Examples
References
Integration of Boolean and Bayesian Networks
1 Introduction
2 Methods
2.1 s-p-scores Associated with Networks, SPAN
2.2 Network Learning
3 Results
3.1 An Example
3.2 Real Example
3.3 Complex Example
4 Discussion
References
Computational Methods for Identifying MicroRNA-Gene Regulatory Modules
1 Introduction
2 Identifying MiRNA-Gene Modules by Integrating Heterogeneous Data Sources
2.1 Bipartite Graph-Based Methods
2.2 Nonnegative Matrix Factorization Methods
2.3 Statistical Modeling Approaches
3 Evaluating the Performance of MiRNA-Gene Module Identification Methods
4 Discussion
5 Conclusions
References
Causal Inference in Biostatistics
1 Introduction
1.1 Causation and Association
1.2 Two Conceptual Frameworks: Causal Effect and Causal Discovery
2 Causal Effect
2.1 Approaches to Causal Inference
2.2 Randomized Clinical Trials
2.2.1 Perfect Randomized Trials
2.2.2 Randomized Trials with Missing Data
2.2.3 Randomized Trials with Post-treatment Variables
2.3 Observational Studies
2.3.1 Unconfounded Treatment Assignment Conditional on Measured Covariates
2.3.2 Unmeasured Cofounding
3 Some Current Research Topics
3.1 Heterogenous Treatment Effect and Precision Medicine
3.2 Integrating Data from Randomized Controlled Trials and Observational Studies
3.3 Multiple Treatments
4 Software Appendix
References
Bayesian Balance Mediation Analysis in Microbiome Studies
1 Introduction
2 Bayesian Balance Mediation Model
2.1 Bayesian Balance Mediation Model with a Binary Treatment
2.2 Direct and Mediation Effect and Estimation Based on Predictive Posterior Distribution
3 MCMC Sampling
3.1 MCMC Sampling
3.2 Conditional Distributions
4 Applications to Real Data
4.1 Mediation Analysis at the Phylum Level
4.2 Analysis at the Order Level
5 Simulation Studies
5.1 Data Generation
5.2 Simulation Result
6 Discussion
References
Part III Systems Biology
Identifying Genetic Loci Associated with Complex Trait Variability
1 Introduction
2 The Concept of vQTL
3 Statistical Methods for vQTL Mapping
3.1 Classical Nonparametric Tests
3.2 Regression-Based Methods
3.3 Two-Stage Methods
3.4 Quantile Integral Linear Model (QUAIL)
3.5 Dispersion Effects
4 Applications of vQTL
4.1 Examples of vQTL
4.2 Screening vQTL for Candidate Loci Involved in GxE Interaction
4.3 Variance Polygenic Score
4.4 Other Applications and Future Directions
References
Cell Type-Specific Analysis for High-throughput Data
1 Introduction
2 Cell Type Composition Estimation
3 Cell Type-Specific Differential Analysis
4 Step-by-step Tutorial
References
Recent Development of Computational Methods in the Field of Epitranscriptomics
1 Introduction
2 MeRIP-seq and Other Technologies for RNA Modification Profiling
3 Methods to Analyze MeRIP-seq Data
3.1 Count-Based Methods for Simple Study Designs
3.2 Methods Compatible with Confounding Factors
3.3 A Guide for RNA Differential Methylation Analysis Using RADAR
4 Web Resources on m6A Epitranscriptome
4.1 Web Servers with m6A Site Prediction
4.2 m6A Epitranscriptome Database
5 Discussion
References
Estimation of Tumor Immune Signatures from Transcriptomics Data
1 Introduction
2 Regression-Based Deconvolution Algorithms
2.1 Linear Least Squares Regression
2.2 Support Vector Regression
2.3 Other Deconvolution Methods
3 Gene Set Enrichment-Based Methods and Other Gene-Based Algorithms
3.1 Gene Set Enrichment Analysis (GSEA)
3.2 Single-Sample GSEA (ssGSEA)
4 Benchmark Studies
5 Discussions
References
Cross-Linking Mass Spectrometry Data Analysis
1 Introduction
1.1 Peptide Identification Based on Mass Spectrometry
1.2 Cross-Linked Peptides Identification
1.2.1 Cross-Linker Selection
1.2.2 Chemical Reaction
1.2.3 Enzyme Digestion
1.2.4 Enrichment of Cross-Linked Peptides
1.2.5 LC-MS and MS2 Acquisition
1.2.6 Data Interpretation
1.2.7 Quality Control
1.2.8 Downstream Applications
2 Non-cleavable Cross-Linking Methods
3 Cleavable Cross-Linking Methods
4 Time-Complexity Comparison Between Non-cleavable Methods and Cleavable Methods
5 False Discovery Rate in CL-MS
5.1 Target–Decoy Approach in Linear Peptides Identification
5.2 TDA in Cross-Linked Peptides Identification
6 Downstream Applications
6.1 Protein Structure Analysis
6.2 Protein–Protein Interactions
7 Conclusion and Perspective
7.1 Wet Lab Experiments
7.2 Dry Lab Analysis
References
Cis-regulatory Element Frequency Modules and their Phase Transition across Hominidae
1 Introduction
2 Dual Eigen-Analysis of CREF Matrices
2.1 Cis-regulatory Element Frequency (CREF) Matrices
2.2 Dual Eigen-Analysis
2.2.1 Low-Rank Matrix Approximation
2.2.2 Robust SVD
2.2.3 Dual CREF Eigen-Modules
2.2.4 Cross-Species Correlations and Conservation of Motif Eigen-Modules
2.2.5 Enrichment Analysis of Gene Eigen-Modules
2.3 Stability of CREF Modules
2.3.1 Sensitivity of Eigenvalues and Eigenvectors
2.3.2 Degenerate 2-D Eigen-Space and Its Stability
3 Dual Eigen-Analysis Unravels the Regulatory Evolution from Apes to Human
3.1 Evolutionary Conservation and Divergence of CREF Eigen-Modules
3.2 Conservation of Top Three CREF Modules
3.3 Dual Eigen-Analysis Identifies the Regulators Involved in Major Biological Processes
3.4 Phase Transition of CREF Eigen-Modules
3.5 Human-Specific Phenotypes Corresponding to Its Unique Fourth Gene Eigenvector
3.6 Motifs Present on Alu Elements Underlying the CREF Module Reorganization
3.7 Cis-trans Regulation of the Human-Specific Gene Eigen-Module
4 Discussion
4.1 Definition of cis-regulatory Profiles
4.2 Dual Eigen-Analysis and the Classical PCA
4.3 The Phase Transition in cis-regulatory Evolution
4.4 Conclusion
References
Improved Method for Rooting and Tip-Dating a Viral Phylogeny
1 Introduction
1.1 The Need for Answering the “When” Question Without a Good Outgroup
1.2 Conventional TRAD Methods Assume a Constant and Need to Be Extended to Allow to Change over Time
2 TRAD: Tip Rooting and Ancestor Dating
2.1 Estimating the Root
2.2 Dating the Common Ancestor (TA) and Estimating the Evolutionary Rate
3 Increasing Evolutionary Rate During SARS-CoV-2 Evolution over Time
3.1 Results from the Viral Phylogeny Released by NCBI on April 3, 2021
3.2 Results from the Viral Phylogeny Released by NCBI on September 4, 2021
4 The Difficulty with the “Where” Question
5 Conclusions
References