reading

Cover
Title Page
List of Contributors
Preface
Acknowledgments
Part I: Platforms for Molecular Data Acquisition and Analysis
1. 1 Clinical Data Collection and Patient Phenotyping
  1. 1.1 Clinical Data Collection
  2. 1.2 Patient Phenotyping
  3. 1.3 Concluding Remarks
  4. References
2. 2 Biobanking, Ethics, and Relevant Legal Issues
  1. 2.1 Introduction
  2. 2.2 Brief Historical Derivation to the Ethical Guidelines in Medical Research
  3. References
  4. 2.3 Biobanking: Definition, Role, and Guidelines of National and International Biobanks
  5. References
  6. 2.4 Tasks of Ethics Committees in Research with Biobank Materials
  7. Further Reading
3. 3 Nephrogenetics and Nephrodiagnostics:
  1. 3.1 Introduction
  2. 3.2 Applications of Molecular Diagnostics
  3. 3.3 Aims of Present‐Day Molecular Genetic Investigations
  4. 3.4 Material Used for Genetic Testing
  5. 3.5 Clinical, Genetic, and Allelic Heterogeneity
  6. 3.6 Oligogenic Inheritance
  7. 3.7 ADPKD, Phenotypic Heterogeneity, and Genetic Modifiers
  8. 3.8 Collagen IV Nephropathies, Genetic and Phenotypic Heterogeneity, and Genetic Modifiers
  9. 3.9 CFHR5 Nephropathy, Phenotypic Heterogeneity, and Genetic Modifiers
  10. 3.10 Unilocus Mutational and Phenotypic Diversity (UMPD)
  11. 3.11 Next‐Generation Sequencing (NGS)
  12. 3.12 Conclusions
  13. Acknowledgments
  14. References
4. 4 The Use of Transcriptomics in Clinical Applications
  1. 4.1 Introduction
  2. 4.2 Clinical Applications of Transcriptomics: Cases and Potential Examples
  3. References
  4. Further Reading
5. 5 miRNA Analysis
  1. 5.1 miRNA Biogenesis, Function, and Annotation
  2. 5.2 Annotation of miRNAs
  3. 5.3 miRNAs: Location, Stability, and Research Methods
  4. 5.4 Use of miRNA In Vivo
  5. 5.5 miRNAs as Potential Therapeutic Agents and Biomarkers: Lessons Learned So Far
  6. 5.6 Conclusion
  7. References
6. 6 Proteomics of Body Fluids
  1. 6.1 Introduction
  2. 6.2 General Workflow for Obtaining High‐Quality Proteomics Results
  3. 6.3 Body Fluids
  4. 6.4 Sample Collection and Storage
  5. 6.5 Sample Preparation for MS/MS Analysis
  6. 6.6 Analytical Instruments
  7. 6.7 Data Processing and Bioinformatics Analysis
  8. 6.8 Validation of Findings
  9. 6.9 Clinical Applications of Body Fluid Proteomics
  10. 6.10 Conclusions
  11. References
7. 7 Peptidomics of Body Fluids
  1. 7.1 Introduction
  2. 7.2 Clinical Application of Peptidomics
  3. 7.3 Different Types of Body Fluids Used in Biomarker Research
  4. 7.4 Sample Preparation and Separation Methods for Mass Spectrometric Analysis
  5. 7.5 Identification of Peptides and Their Posttranslational Modifications
  6. 7.6 Urinary Peptidomics for Clinical Application
  7. 7.7 Concluding Remarks
  8. References
8. 8 Tissue Proteomics
  1. 8.1 Introduction
  2. 8.2 Tissue Proteomics Workflow
  3. 8.3 Tissue Sample Collection and Storage
  4. 8.4 Sample Preparation
  5. 8.5 Overcoming Tissue Complexity and Protein Dynamic Range: Separation Techniques
  6. 8.6 Instrumentation
  7. 8.7 Quantitative Proteomics
  8. 8.8 Functional Annotation of Proteomics Data
  9. 8.9 Application of MS‐Based Tissue Proteomics in Bladder Cancer Research
  10. 8.10 Conclusions
  11. References
9. 9 Tissue MALDI Imaging
  1. 9.1 Introduction
  2. 9.2 Experimental Procedures
  3. 9.3 Applications in Clinical Research
  4. References
10. 10 Metabolomics of Body Fluids
  1. 10.1 Introduction to Metabolomics
  2. 10.2 Analytical Techniques
  3. 10.3 Statistical Tools and Systems Integration
  4. 10.4 Metabolomics in CKD
  5. 10.5 Conclusions
  6. References
11. 11 Statistical Inference in High‐Dimensional Omics Data
  1. 11.1 Introduction
  2. 11.2 From Raw Data to Expression Matrices
  3. 11.3 Brief Introduction R and Bioconductor
  4. 11.4 Feature Selection
  5. 11.5 Sample Classification
  6. 11.6 Real Data Example
  7. 11.7 Multi‐Platform Data Integration
  8. 11.8 Discussion and Further Challenges
  9. References
12. 12 Epidemiological Applications in ‐Omics Approaches
  1. 12.1 Overview: Importance of Study Design and Methodology
  2. 12.2 Principles of Hypothesis Testing
  3. 12.3 Selection of Appropriate Epidemiological Study Design for Hypothesis Testing
  4. 12.4 Types of Epidemiological Study Designs
  5. 12.5 Selection of Appropriate Statistical Analyses for Hypothesis Testing
  6. 12.6 Summary
  7. References
Part II: Progressing Towards Systems Medicine
1. 13 Introduction into the Concept of Systems Medicine
  1. 13.1 Medicine of the Twenty‐First Century: From Empirical Medicine and Personalized Medicine to Systems Medicine
  2. 13.2 The Emerging Concept of Systems Medicine
  3. 13.3 Networking Among All Key Stakeholders
  4. 13.4 Coordinated European Efforts for Dissemination and Implementation
  5. 13.5 The Contributions of Academia in Systems Medicine
  6. 13.6 Data Generation: Omics Technologies
  7. 13.7 Data Integration: Identifying Disease Modules and Multilayer Disease Modules
  8. 13.8 Modeling: Computational and Animal Disease Models for Understanding the Systemic Context of a Disease
  9. 13.9 Examples and Success Stories of Systems Medicine‐Based Approaches
  10. 13.10 Limitations, Considerations, and Future Challenges
  11. References
2. 14 Knowledge Discovery and Data Mining
  1. 14.1 Introduction
  2. 14.2 Knowledge Discovery Process
  3. 14.3 Data Mining in Scientific Applications
  4. 14.4 Bioinformatics Data Mining Tools
  5. 14.5 Conclusions
  6. References
3. 15 ‐Omics and Clinical Data Integration
  1. 15.1 Introduction
  2. 15.2 Data Sources
  3. 15.3 Integration of Different Data Sources
  4. 15.4 Integration of Different ‐Omics Data
  5. 15.5 Visualization of Integrated ‐Omics Data
  6. 15.6 Integration of ‐Omics Data into Models
  7. 15.7 Data Integration and Human Health
  8. 15.8 Conclusions
  9. References
4. 16 Generation of Molecular Models and Pathways
  1. 16.1 Introduction
  2. 16.2 PKN Construction Through Expert Biocuration
  3. 16.3 Modeling and Simulating the Dynamical Behavior of Networks
  4. 16.4 Conclusions
  5. References
5. 17 Database Creation and Utility
  1. 17.1 Introduction
  2. 17.2 Database Systems
  3. 17.3 Biological Databases
  4. 17.4 Conclusions
  5. References
Part III: Test Cases CKD and Bladder Carcinoma
1. 18 Kidney Function, CKD Causes, and Histological Classification
  1. 18.1 Introduction
  2. 18.2 The Evaluation of Glomerular Filtration Rate
  3. 18.3 Causes of CKD
  4. 18.4 Assessment of Disease Progression and Response to Therapy for the Individual: Interval Renal Biopsy
  5. 18.5 Recent Advances: Pathology at the Molecular Level
  6. 18.6 Digital Pathology
  7. 18.7 Conclusions
  8. References
2. 19 CKD: Diagnostic and Other Clinical Needs
  1. 19.1 The Evolving Concept of Chronic Kidney Disease
  2. 19.2 A Growing Epidemic
  3. 19.3 Increasing Mortality from Chronic Kidney Disease
  4. 19.4 The Issue of Cause and Etiologic Therapy
  5. 19.5 Unmet Medical Needs: Biomarkers and Therapy
  6. 19.6 Conclusions
  7. Acknowledgments
  8. References
3. 20 Molecular Model for CKD
  1. 20.1 Introduction
  2. 20.2 Data‐Driven Approaches and Multiomics Data Integration
  3. 20.3 Chronic Kidney Disease (CKD) Case Study
  4. 20.4 Final Remarks
  5. Acknowledgments
  6. Conflict of Interest Statement
  7. References
4. 21 Application of Omics and Systems Medicine in Bladder Cancer
  1. 21.1 Introduction
  2. 21.2 Bladder Cancer Pathology and Clinical Needs
  3. 21.3 Systems Medicine in Bladder Cancer
  4. 21.4 Outlook
  5. Acknowledgments
  6. References
Index
Wiley Series on Mass Spectrometry
End User License Agreement

List of Tables

Chapter 01
1. Table 1.1 Coding on the case report form module.
2. Table 1.2 Well‐designed and poorly designed data fields.
3. Table 1.3 Distribution of codes across the 14 eMERGE phenotyping algorithms.
Chapter 03
1. Table 3.1 Genes with evidence that they modify the disease outcome of monogenic renal conditions.
Chapter 04
1. Table 4.1 Examples of transcriptomics clinical applications for PCR, microarray, and RNA‐seq disease.
2. Table 4.2 Technical comparison between PCR, microarray, and RNA‐Seq methodologies.
Chapter 05
1. Table 5.1 miRNA detection methods.
2. Table 5.2 Available web‐tools for miRNA target prediction.
3. Table 5.3 Available web‐tools for miRNA target validation.
4. Table 5.4 miRNAs and target validation methods.
Chapter 06
1. Table 6.1 Comparison of tissue and body fluids proteomics in human.
2. Table 6.2 Characteristics of four body fluids.
3. Table 6.3 Sample collection guidelines references.
4. Table 6.4 Advantages and disadvantages of various separation techniques.
Chapter 07
1. Table 7.1 Proteomics and peptidomics urinary databases.
2. Table 7.2 Protease regulation in human diseases.
Chapter 08
1. Table 8.1 An overview of the studied tissue proteomes.
2. Table 8.2 An overview on the workflow for collection of fresh‐frozen tissue specimens and associated critical factors.
3. Table 8.3 Overview on the application of LCM on human tissues in combination with proteomic analysis.
4. Table 8.4 Comparison of the three main enzymatic digestion strategies applied in bottom‐up proteomics.
5. Table 8.5 Specification of commonly used mass spectrometers to analyze tissue proteome.
6. Table 8.6 Examples of tools/resources applied for functional annotations of proteomics datasets.
7. Table 8.7 Overview on tissue proteomics studies performed in the context of bladder cancer.
Chapter 09
1. Table 9.1 An overview of the most commonly used ionization sources for MSI experiments.
2. Table 9.2 Advantages and disadvantages of MALDI‐MSI.
Chapter 10
1. Table 10.1 A general comparison of analytical techniques in metabolomics.
Chapter 11
1. Table 11.1 R packages dedicated to different omics tasks such as feature selection, sample classification, and data integration.
2. Table 11.2 Top 10 ranking of the 1828 peptides in the proteomic dataset under different feature selection methods.
3. Table 11.3 Top 10 ranking of the 227 metabolites in the metabolites dataset under different feature selection methods.
4. Table 11.4 Classifier performance as a function of training set size.
5. Table 11.5 Demographic and clinical data.
Chapter 14
1. Table 14.1 Data mining terminologies.
Chapter 15
1. Table 15.1 Main data sources available for data integration.
2. Table 15.2 Visualization tools focused on interaction networks.
Chapter 16
1. Table 16.1 Example of the use of membership functions.
2. Table 16.2 Tools cited in the text by task.
Chapter 17
1. Table 17.1 List of available data analysis software for scientific applications.
2. Table 17.2 Types of biological databases.
Chapter 18
1. Table 18.1 International Society of Nephrology/Renal Pathology Society (ISN/RPS) 2003 classification of lupus nephritis.
2. Table 18.2 Classification scheme for ANCA‐associated glomerulonephritis.
3. Table 18.3 Glomerular classification of DN.
Chapter 20
1. Table 20.1 Model development in several disease/conditions and omics platforms.
2. Table 20.2 Databases within the scope of CKD or of general scope that can be filtered by conditions and ‘omics platforms based on expression profiles.
3. Table 20.3 Overall view of standalone and web‐based tools to assist in pathway(s) visualization and in the development of disease models.
4. Table 20.4 Overlap of the associated genes in four related kidney diseases (CRI, Chronic Kidney Insufficiency; DN, Diabetic Nephropathy; AKI, Acute Kidney Failure; GN, Glomerulonephritis) using data sourced from DisGeNET.
5. Table 20.5 Patients with CKD stages II–IV were analyzed.
6. Table 20.6 Instance counting of the associated number of features per high‐throughput acquisition omic technologies (PRO, proteomics; MET, metabolomics; MIR, microRNA transcriptomics) after applying each of the statistical criteria thresholds.
7. Table 20.7 Gene ontology (GO, BP, biological process; MF, molecular function, performed in the Cytoscape ClueGO [16] app.
Chapter 21
1. Table 21.1 Current treatment modalities across the bladder cancer disease development and progression.
2. Table 21.2 Overview of the literature review including the search criteria and the results obtained.

List of Illustrations

Chapter 03
1. Figure 3.1 (a) The thin basement membrane nephropathy (TBMN) genetic map of Cyprus. Shown are villages and cities where 26 families of TBMN have been detected so far. All patients carry founder mutations in the COL4A3/COL4A4 genes. Families in villages shown with blue dots carry the founder mutation in the COL4A3 gene, p.G1334E. (b) The CFHR5 genetic map of Cyprus. Shown are 12 villages where 23 families of CFHR5 nephropathy have been detected. All patients carry the same CFHR5 exon 2–3 duplication, as a result of a founder effect.
2. Figure 3.2 (a) Patient with CFHR5 nephropathy. Light micrograph from a representative kidney biopsy in family CY5308 stained with PAS showing mesangial hypercellularity (arrows) (×400). (b) Immunofluorescence from a representative kidney biopsy in family CY5308 showing diffuse granular C3 staining on the mesangium and capillary wall. (c) Electron micrograph from a representative kidney biopsy in family CY5308 showing the presence of subendothelial deposits (SD) as well as mesangial deposits (MD). A red blood cell (RBC) is seen in Bowman’s capsule (×8000).
3. Figure 3.3 Pedigrees of the two original Greek–Cypriot families segregating CFHR5 nephropathy that were described in London.
Chapter 04
1. Figure 4.1 Gene coverage for RNA‐Seq data. First gene track (blue line) shows the gene structure of MECOM by RefSeq gene annotation. Thick lines show exon locations and arrowed lines intronic locations. The second and third tracks show the coverage of MECOM for samples 1 and 2, respectively. In gray, the depth of coverage of sample 1 indicates that there is expression where the peaks are located. The extent of the expression is proportional to the height of the peaks, and the amplitude represents how much of the gene sequence is expressed, usually coinciding with exon boundaries.
2. Figure 4.2 Visualization of a SNV detected by RNA‐Seq. The 4 bp insertion (TCTG) is represented as a purple line on the alignment track for the NPM1 gene. The gene track (blue) shows the codons present in this particular location (blue and dim blue) as well as the amino acid it encodes for (letters). The sequence of this particular location is shown, and the black arrow gives information about the strand of the shown nucleotides (forward). The third track shows the coverage for each of the represented bases. Finally, the alignment is shown where horizontal gray lines represent a match with the reference sequence and a colored horizontal line represents a mismatch, which is colored with the same colors as the sequence nucleotides, being T red, A green, C blue, and G orange; black represents a deletion.
3. Figure 4.3 Visual representation of splicing detected for two sequenced patients with RNA‐Seq. Bottom blue track represents gene structure of IRF1 gene. Thick lines show exons bases and arrowed line intronic bases. Green track shows the coverage of sample 3 (green blocks) as well as the number of reads that span the distance to the next splice acceptor region (green line). Red tracks show coverage of sample 5 (red blocks) and number of reads that span the distance to the next splice acceptor region (red line) connecting two exons. The last red track shows an alternative splicing event with 221 reads that are connecting exons 3 and 5 but skipping exon 4. This shows that exon 4 is partially skipped by some transcripts in sample 5.
Chapter 05
1. Figure 5.1 mRNA biogenesis and function. (A) In the first step of miRNA biogenesis, miRNA is transcribed similar to mRNA transcription with the help of polymerase II or III and various transcription factors followed by precursor miRNA (pre‐miRNA) generation via a canonical and noncanonical pathway. (A1) Canonical pathway: miRNA is transcribed in a primary form (pri‐miRNA with a hairpin loop) and then spliced by the Drosha/DGCR8 complex to form the pre‐miRNA. (A2) Noncanonical pathway: A “mirtron” is generated from mRNA after splicing in a complex called spliceosome. Next, the mirtron reforms to the shape of a pre‐miRNA. (B) The pre‐miRNA is transferred from the nucleus to the cytoplasm with the help of exportin‐5 in a Ran‐GTP‐dependent manner. (C) In the cytoplasm the protein DICER together with TRBP cleaves the pre‐miRNA hairpin loop and releases the two strands. One strand will serve as the mature miRNA (18–22 nt length), while the other, in most cases, will be degraded. (D) The mature miRNA then binds to a protein complex including AGO2, forming the so‐called RISC. AGO2 then leads the miRNA to the mRNA target. (E) If the miRNA sequence is completely complementary to the 3′ UTR of the mRNA, it binds and AGO2 cleaves the mRNA, completely stopping the translation of the protein. (F) If the miRNA sequence is partially complementary with the mRNA’s 3′ UTR, GW182 is recruited in the RISC to aid with the connection and inhibit translation. AGO2, Argonaute 2; DGCR8, DiGeorge syndrome critical region 8; FXR1, fragile X mental retardation‐related protein 1; PACT, protein kinase R‐activating protein; RISC, RNA‐induced silencing complex; TRBP, HIV‐1 transactivating response (TAR) RNA‐binding protein. The picture was designed using Servier Medical ART (http://smart.servier.fr/servier‐medical‐art).
2. Figure 5.2 miRNA‐transporting molecules. During biogenesis or even after fulfilling their biological function, miRNAs can be incorporated into exosomes or bound to lipoproteins (e.g., HDL) or RNA‐binding proteins (e.g., AGO2). Release of the encapsulated or bound miRNAs can occur via exocytosis during cell lysis and apoptosis. Encapsulation or binding to these proteins most likely protects the miRNAs from degradation in the hostile biofluid environment. The figures were designed using Servier Medical ART (http://smart.servier.fr/servier‐medical‐art).
3. Figure 5.3 (a) Antagomir function. The normal pathway of the miRNA biogenesis and action is shown in summary in A, A1, B1, A2, and A3 with the final result to be the suppression of gene expression. With the addition of the antagomirs (B2), the action of the miRNAs is inhibited. Antagomirs bind in a complementary manner to the miRNAs, leading to translation of the targeted mRNA (B4). (b) miRNA mimics function. The normal pathway of the miRNA biogenesis and action is shown in summary in A, A1, and C1. AGO2 molecules are available in excess to any mature miRNA and complete its function (A2 and A3). With the addition of the miRNA mimics (C2), the free AGO2 proteins are capable to bind to more miRNAs than normal, leading to additional inhibition of translation of the mRNA target molecules.
4. Figure 5.4 A possible workflow of how research on miRNAs could lead to clinically useful miRNAs.
Chapter 06
1. Figure 6.1 Graphical depiction of a general workflow of proteomics analysis.
Chapter 07
1. Figure 7.1 Illustration of the workflow for peptidomic analysis.
Chapter 08
1. Figure 8.1 Application and significance of tissue proteomics in clinical research. Through a global proteomic profiling of tissue specimens, disease mechanisms can be highlighted at the molecular level, leading to the discovery of potential biomarkers and drug targets.
2. Figure 8.2 Schematic representation of tissue proteomics workflow. The main steps during the tissue proteomics studies are marked in bold.
Chapter 09
1. Figure 9.1 The general principles of MALDI‐MSI. Laser ablation leads to the desorption and ionization of matrix and analyte ions. The detected ions yield the generation of mass spectra at discrete spatial coordinates and the spatial distribution of any of the ions present in these spectra can be visualized following the generation of a molecular image.
2. Figure 9.2 The different approaches in matrix deposition for MALDI‐MS profiling and MALDI‐MSI.
3. Figure 9.3 Sample preparation workflow for the MALDI‐MSI proteomic analysis of fresh‐frozen (a) and FFPE (b) tissues. The steps common to both workflows are highlighted in (c).
4. Figure 9.4 The MALDI‐MSI data cube represented by m/z values (x‐axis), signal intensity (y‐axis), and spatial coordinates (z‐axis).
5. Figure 9.5 The MALDI‐MS spectra preprocessing steps. Left panel, in the 3 000–18 000 m/z range: Raw (a), following baseline correction (b) and following smoothing (c). Right panel, in the 10 000–10 500 m/z range: Raw spectrum (d), following baseline correction (e) and following smoothing (f).
6. Figure 9.6 The generation of overview MALDI spectra. (top) mass spectra obtained from two separate x–y coordinates and (bottom) the resulting average and skyline spectra.
Chapter 10
1. Figure 10.1 The ‐omics progression toward phenotype characterization. Four major ‐omics fields build from the previous to describe the phenotype. As can been seen, metabolomics is the most distal of the ‐omics fields and therefore can most accurately describe the true phenotype.
2. Figure 10.2 Metabolomics workflow.
3. Figure 10.3 Example of a partial ¹H NMR of urine with selected metabolite annotations.
4. Figure 10.4 RSPA alignment scheme.
5. Figure 10.5 (a) PCA plot of urine samples from CKD patients. (b) Orthogonal partial least squares discriminant analysis (O‐PLS‐DA) scores plot of the same sample set with CKD stage 1 and 5 being the discriminating factor. (c) O‐PLS‐DA loading plot. The color bar represents the correlation coefficient of the corresponding variable in discriminating between the groups. On the y‐axis, the O‐PLS covariance measures the degree and direction of a given peak with that of variable. The layout of the plot resembles the original metric of the NMR spectra, meaning stereochemistry of the plot is preserved and is therefore straightforward to interpret for an NMR spectroscopist. This supervised analysis allows a more targeted search for biomarkers of a certain predetermined variable.
Chapter 11
1. Figure 11.1 Different methods for data integration.
Chapter 12
1. Figure 12.1 Types of epidemiological studies utilized in ‐omics research. The types of analytical epidemiological studies most frequently applied in ‐omics research are presented. Observational studies are most commonly used for the discovery and validation of diagnostic and prognostic biomarkers, while experimental studies are most frequently employed for evaluating the utility of predictive biomarkers.
2. Figure 12.2 Hierarchy of research design. The hierarchy of observational and interventional epidemiological study designs, in relation to the type of biomarker (BM) under evaluation, is displayed. Optimal study designs for the evaluation of diagnostic, prognostic, and predictive biomarkers (BM) are annotated.
3. Figure 12.3 Temporal association between collecting sample and clinical information in relation to assessing the presence of disease and/or patient outcomes of interest, according to type of epidemiological study design applied.
Chapter 13
1. Figure 13.1 Overview of the empirical medicine, personalized medicine, and the emerging systems medicine.
2. Figure 13.2 The network‐within‐a‐network concept as a cornerstone of the systems medicine approach.
3. Figure 13.3 Systems medicine toward implementation of P4 medicine and a paradigm shift in healthcare.
4. Figure 13.4 The main contributions of academia in systems medicine involve generation, integration, and modeling of disease‐related data in an interassociated and interlocking manner.
Chapter 14
1. Figure 14.1 The end‐to‐end knowledge discovery (KDD) process. This figure shows the flow of information from the acquisition stage through to the final output via numerous steps.
2. Figure 14.2 Possible research applications of KDD techniques.
3. Figure 14.3 Forms of data preprocessing. Examples of data handling routines generally used in the transformation of acquired raw data prior to modeling are shown.
4. Figure 14.4 Data mining tasks. The decision tree of various data mining methods is shown.
5. Figure 14.5 Data mining applications. Commonly employed tasks in data mining such as classification, regression analysis, or clustering are depicted.
6. Figure 14.6 The rise of text mining approaches in science and discovery from 1995 to 2015. Information retrieved from PubMed based on “text mining OR literature mining” keywords. *search performed July 2016.
7. Figure 14.7 Growing interest in data mining techniques in scientific research. Information retrieved from PubMed based on “data mining OR machine learning” keywords. *search performed July 2016.
8. Figure 14.8 General example of machine learning application. A training set of labeled gene expression data is used as an input for the learning algorithm to build a model able to predict the label (diabetic nephropathy (DN) or healthy) for future gene expression data (unlabeled, test set). Based on the model, labels are assigned to the test set, and performance of the model is evaluated.
Chapter 15
1. Figure 15.1 Overview of the systems biology approach toward the realization of personalized medicine. The many different ‐omics data currently available (reddish box) must be deposited in one or more database in order to systematically organize the information and to facilitate the data integration process. Integrated ‐omics data are analyzed (light‐blue box) by means of statistical methods, computer‐aided modeling, and visually represented in order to understand emergent properties and design principles of the biological system (orange cartoon). Personalized ‐omics data (green box), the knowledge of emergent properties/design principles, and different scientific competencies (purple box) will ultimately merge, allowing to develop personalized medical treatment of multifactorial diseases.
2. Figure 15.2 From uninformative to meaningful ‐omics data visualization with Cytoscape. Adding data on the network in a rational way improves the information content of the representation. On the top part, an example of hairball layout obtained using data from high‐confidence protein–protein interactions [120] (a). On the lower level, examples of ‐omics data mapped on a network modifying its elements (nodes and edges color/dimension, layout) exploiting Cytoscape apps (Node Chart Plug‐in for transcriptomics (b), CyFluxViz for fluxomics, (c) and chemViz2 for metabolomics (d)).
3. Figure 15.3 Network showing a clustering of co‐expression data using Cytoscape. Source: Dataset from Prieto et al. [121]. In the network, genes are mapped as nodes, while edges represent the co‐expression relation. Identified clusters (isolated groups of nodes highlighted with ellipses) represent different cellular functions, for example, A, mitochondrial metabolism; B, ribosome; C, nuclear related metabolism; and D, immune response.
4. Figure 15.4 A cellular overview of the metabolic network of yeast generated with Pathway Tools. In this visualization, metabolic reactions are subdivided in pathways (grey boxes), exchange reactions are placed across the border of a rectangle representing the cellular membrane, while reactions not assigned to a specific pathway are grouped on the right side. The metabolic map has been enriched with data (colored nodes) on reporter metabolites (see section “Integrating Transcriptomics and Metabolic Pathways” for a reference on reporter metabolites), the color of the node is set accordingly to the confidence value for the reporter metabolite identification (top left color bar), the distribution of nodes having a given confidence value is shown on a histogram in the bottom left corner.
5. Figure 15.5 Workflow of constraint‐based data integration approaches to cancer. Workflow from the extraction and curation of a model from the generic human metabolic map (top‐right box), to its customization according to ‐omics data (following the arrows downstream of top boxes), to its flux balance analysis to estimate fluxes and other approaches (such as the ensemble approach) to explore the space of possible phenotypes, and to the simulation of possible drug targets (bottom box). For each process some references are reported as an example.
Chapter 16
1. Figure 16.1 Modeling regulatory network at the intersection of the top‐down and the bottom‐up methodologies to describe biological systems and predict their behavior/outcome.
2. Figure 16.2 Curation workflow for prior knowledge network (PKN) construction and the different steps necessary to build a contextualized PKN.
3. Figure 16.3 Modeling workflow. Network construction and iterative optimization against experimental datasets. (a) Toy example of a Boolean network with five nodes A, B, C, D, and E and one AND logic gate; arrow edges are activations and T edges are inhibitions. (b) Network attractors: grey nodes are active (1) and white nodes are inactive (0). (c) Test the correctness of the model compared with experimental data: expression data, proteomic data, or other data relevant for the biological question. (d) In silico simulations and experimental prediction. (e) Experimental validation of simulation outcomes. (f) Final model.
4. Figure 16.4 Example of metabolic network with three metabolites—A, B, and C. The system contains three reactions that are internal fluxes with one reversible reaction (v3, v4) and three exchanges fluxes (b1, b2, and b3). S is the stoichiometric matrix, V the matrix of fluxes, at steady state S∙V = 0, A: −v1 − v2 + b1 = 0, B: v1 + v3 − v4 − b2 = 0, C: v2 − v3 + v4 − b3 = 0.
Chapter 17
1. Figure 17.1 Data life cycle. Concept supporting the management of the data and the flow of an information system’s data throughout its life cycle: from creation, collection, and discovery process until storage or disposal.
2. Figure 17.2 Simplified representation of database system (DBS) and components. End users communicate with the software systems/applications, which, in turn, communicate (through the programming interface) with the DBMS. The DBMS communicates with the operating system to store data in and/or extract data from the database.
3. Figure 17.3 Hierarchy of data. Database consists of integrated files, which are a collection of related records. Each record is a collection of related fields, where single facts or attributes are stored.
4. Figure 17.4 Schematic representation of hierarchical, network, and relational database models.
5. Figure 17.5 Basic concepts of relational model. (a) Database entities (tables) are matched based on data in key columns, forming a relationship. (b) Relational model organizes data into tables (relations) of columns and rows, with a unique identifier of each row (primary key).
6. Figure 17.6 Non‐SQL database types. (a) Key‐value stores—key is a unique attribute of the content (value). (b) Graph databases—storage of interconnected data represented by nodes and edges. (c) Column family stores—key‐value pair, where the key is mapped to a value that is a set of columns. (d) Document database—data is stored in the form of multiple documents (e.g., JSON, XML), which contain more complex grouping of key‐value pairs.
7. Figure 17.7 Database design process.
Chapter 18
1. Figure 18.1 In this cartoon, the schematic structure of a normal glomerulus can be appreciated. The capillary stalk is composed of mesangial cells (M) from which the branching of capillary lumen starts. The capillary wall is composed, from the inside to the outside, by fenestrated endothelial cells (E), the glomerular basement membrane (GBM), and from the foot processes of podocytes (PO).
2. Figure 18.2 MPGN—a simple classification.
3. Figure 18.3 Definition of endocapillary hypercellularity.
4. Figure 18.4 In figure a particular pattern of glomerular injury is represented, known as “endocapillary hypercellularity,” as the main manifestation of four different glomerular diseases. In Figure 18.3 there is the definition of endocapillary hypercellularity. Figure 18.4 shows the light microscopy images of four glomeruli with a similar endocapillary hypercellularity but affected by different pathology ((a), (b), (d) stained with trichrome stain, (c) stained with Hematoxylin and Eosin, ×20).
5. Figure 18.5 In this figure a particular pattern of glomerular injury is represented, known as “endocapillary hypercellularity,” as the main manifestation of four different glomerular diseases. The same glomeruli was stained with immunofluorescence technique for different antisera. In (a) there are sparse subepithelial deposits of C3, in (b) mesangioparietal deposits of C3, in (c) mesangioparietal deposits of C1q, and in (d) parietal deposits of IgM.
6. Figure 18.6 In this figure a particular pattern of glomerular injury is represented, known as “endocapillary hypercellularity,” as the main manifestation of four different glomerular diseases. It depicts the electron microscopy feature of each case. In (a) the presence of subepithelial deposits, in (b) the double contour aspect of basement membrane, in (c) the so‐called wire loops, and (d) reveal the formation of structured cilindric deposits.
7. Figure 18.7 In this figure a particular pattern of glomerular injury is represented, known as “endocapillary hypercellularity,” as the main manifestation of four different glomerular diseases. The diagnosis for each case is reported.
8. Figure 18.8 In this figure a particular pattern of glomerular injury is represented, called “nodular,” as the main manifestation of four different glomerular diseases. There is the definition of nodular pattern.
9. Figure 18.9 In this figure a particular pattern of glomerular injury is represented, called “nodular,” as the main manifestation of four different glomerular diseases. The light microscopy images of four glomeruli with a similar nodular pattern but affected by different pathology ((a) stained with PAS; (b), (c), and (d) stained with trichrome stain ×20).
10. Figure 18.10 In this figure a particular pattern of glomerular injury is represented, called “nodular,” as the main manifestation of four different glomerular diseases. The same glomeruli is stained with immunofluorescence technique for different antisera. In (a) there are continuous subendothelial deposits with lobular appearance of C3, in (b) linear deposits of IgG along the glomerular basement membrane (GBM), in (c) mesangioparietal deposits of lambda light chain, and in (d) the negativity to any antiserum tested.
11. Figure 18.11 In this figure a particular pattern of glomerular injury is represented, called “nodular,” as the main manifestation of four different glomerular diseases. Depicts the electron microscopy feature of each case. In (a) the presence of subendothelial deposits, in (b) the absence of immunodeposits in the GBM, in (c) the finely granular subendothelial deposits, and (d) reveal the formation of fibrillary deposits.
12. Figure 18.12 In this figure a particular pattern of glomerular injury is represented, called “nodular,” as the main manifestation of four different glomerular diseases. The diagnosis for each case is reported.
13. Figure 18.13 An example of the WSI viewer.
Chapter 19
1. Figure 19.1 The evolving concept of chronic kidney disease (CKD).
2. Figure 19.2 CKD: an aging‐associated disease.
3. Figure 19.3 CKD: an aging‐associated disease that increases the risk of death. Risk for individuals in the 18–54 years age range depicted in black and for individuals 75 years of age or older in red.
4. Figure 19.4 CKD is the nontransmissible cause of death that increased the most in the past 20 years according to the global burden of disease 2013 (GBD2013) study [16]. For comparison other key causes of death from chronic disease are shown.
Chapter 20
1. Figure 20.1 Purposed workflow for a data‐driven approach in biomarker discovery, starting from data acquisition till clinical implementation. db, database.
2. Figure 20.2 Molecular overlap of genes/proteins and microRNAs in CRI: Chronic Renal Insufficiency, DN: Diabetic Nephropathy, AKI: Acute Kidney Failure, and GN: Glomerulonephritis. Venn diagram generated in http://bioinfogp.cnb.csic.es/tools/venny/.
3. Figure 20.3 Molecular network associations of the 18 overlapping molecules from the four selected kidney disorders. The search parameters in the online based tool GeneMania were: organism Homo sapiens; Genes input: IGF1, HGF, CCL2, VEGFA, SOS1, IL1B, APOE, PPARG, AGT, IL6, TLR4, PARP1, REN, AGTR1, ACE, ALB, IL8, and TNF; Network weighting was based on the Biological process. GeneMania online application version 3.5.0.
4. Figure 20.4 The 20 most deregulated proteins from the Ghent study are compared across several published studies related with chronic kidney disease. The Heatmap displays the hierarchical clustering by average linkage of the Euclidean distances. Morpheus web‐tool available at https://software.broadinstitute.org/morpheus. The datasets comparisons were pulled from CKDdb (http://www.padb.org/ckddb/). The prefix “Exp” is related with the PubMed ID of the experimental study.
5. Figure 20.5 Variables (disorders) view of the Principal Component analysis (PCA) analysis. Both PC1 and PC2 explain 33.9% of the cumulative data variation. Two main clusters are formed the Chronic Renal Insufficiency (CRI) highlighted in the box on the PC1 and the Diabetes cluster on PC2.
6. Figure 20.6 Loadings (molecules) view of the Principal Component analysis (PCA) analysis with the top 11 molecules more accountable for discriminating within samples.
7. Figure 20.7 Using the KEGG [17] mapper tool for mapping the molecules into the human pathway for Complement and coagulation cascades—available at http://www.genome.jp/kegg‐bin/show_pathway?hsa04610. The binary color rearrangement denotes molecular regulation from the proteomic study, which compares progressors with stable patients. Red objects: upregulation (fold‐change, FC ≥ 1.5); Green objects: downregulation (fold‐change, FC ≤ −1.5).
8. Figure 20.8 Protein–protein interactions (PPIs), 51 nodes, 8 clusters, no enrichment, plus miRNAs as regulatory elements. GeneMania app [29] for Cytoscape. The pink color stroke of the network edges represents PPIs and the black stroke color is linking regulatory elements, such as miRs. The molecular fold‐change (FC) interval established was [−10, 10].
9. Figure 20.9 Reconstruction of metabolic pathways using network‐based analysis in MetScape [32] app for Cystoscape. The following thresholding was applied metabolite FC ≥ 1.10 and p value < 0.05, gene–protein FC ≥ 1.5 and p value < 0.05. Association between gene–enzyme–reaction–metabolite is shown. The direction of the regulation (up or down) is represented by node size, in which smaller to greater nodes denote down‐ and upregulation, respectively. Statistically significant (p value < 0.05) molecules are surrounded by a border green paint.
Chapter 21
1. Figure 21.1 Clinical needs for bladder cancer. BCG, Bacillus Calmette–Guerin; BM, biomarker; CIS, carcinoma in situ; COU, context of use; metBC, metastatic BC; N(MIBC), (non)–muscle‐invasive BC; UUT, upper urinary tract.
2. Figure 21.2 Graphical summary of the systems medicine‐driven approach.

WILEY SERIES ON MASS SPECTROMETRY

Integration of Omics Approaches and Systems Biology for Clinical Applications