Data Mining in Drug Discovery -- Contents -- List of Contributors -- Preface -- A Personal Foreword -- Part One: Data Sources -- 1 Protein Structural Databases in Drug Discovery -- 1.1 The Protein Data Bank: The Unique Public Archive of Protein Structures -- 1.1.1 History and Background: AWealthy Resource for Structure-Based Computer-Aided Drug Design -- 1.1.2 Content, Format, and Quality of Data: Pitfalls and Challenges When Using PDB Files -- 1.1.2.1 The Content -- 1.1.2.2 The Format -- 1.1.2.3 The Quality and Uniformity of Data -- 1.2 PDB-Related Databases for Exploring Ligand-Protein Recognition -- 1.2.1 Databases in Parallel to the PDB -- 1.2.2 Collection of Binding Affinity Data -- 1.2.3 Focus on Protein-Ligand Binding Sites -- 1.3 The sc-PDB, a Collection of Pharmacologically Relevant Protein-Ligand Complexes -- 1.3.1 Database Setup and Content -- 1.3.2 Applications to Drug Design -- 1.3.2.1 Protein-Ligand Docking -- 1.3.2.2 Binding Site Detection and Comparisons -- 1.3.2.3 Prediction of Protein Hot Spots -- 1.3.2.4 Relationships between Ligands and Their Targets -- 1.3.2.5 Chemogenomic Screening for Protein-Ligand Fingerprints -- 1.4 Conclusions -- References -- 2 Public Domain Databases for Medicinal Chemistry -- 2.1 Introduction -- 2.2 Databases of Small Molecule Binding and Bioactivity -- 2.2.1 BindingDB -- 2.2.1.1 History, Focus, and Content -- 2.2.1.2 Browsing, Querying, and Downloading Capabilities -- 2.2.1.3 Linking with Other Databases -- 2.2.1.4 Special Tools and Data Sets -- 2.2.2 ChEMBL -- 2.2.2.1 History, Focus, and Content -- 2.2.2.2 Browsing, Querying, and Downloading Capabilities -- 2.2.2.3 Linking with Other Databases -- 2.2.2.4 Special Tools and Data Sets -- 2.2.3 PubChem -- 2.2.3.1 History, Focus, and Content -- 2.2.3.2 Browsing, Querying, and Downloading Capabilities -- 2.2.3.3 Linking with Other Databases.

2.2.3.4 Special Tools and Data Sets -- 2.2.4 Other Small Molecule Databases of Interest -- 2.3 Trends in Medicinal Chemistry Data -- 2.4 Directions -- 2.4.1 Strengthening the Databases -- 2.4.1.1 Coordination among Databases -- 2.4.1.2 Data Quality -- 2.4.1.3 Linking Journals and Databases -- 2.4.2 Next-Generation Capabilities -- 2.5 Summary -- References -- 3 Chemical Ontologies for Standardization, Knowledge Discovery, and Data Mining -- 3.1 Introduction -- 3.2 Background -- 3.2.1 The OBO Foundry: Ontologies in Biology and Medicine -- 3.2.2 Ontology Languages and Logical Expressivity -- 3.2.3 Ontology Interoperability and Upper-Level Ontologies -- 3.3 Chemical Ontologies -- 3.4 Standardization -- 3.5 Knowledge Discovery -- 3.6 Data Mining -- 3.7 Conclusions -- References -- 4 Building a Corporate Chemical Database Toward Systems Biology -- 4.1 Introduction -- 4.2 Setting the Scene -- 4.2.1 Concept of Molecule, Substance, and Batch -- 4.2.2 Challenge of Registering Diverse Data -- 4.3 Dealing with Chemical Structures -- 4.3.1 Chemical Cartridges -- 4.3.2 Uniqueness of Records -- 4.3.3 Use of Enhanced Stereochemistry -- 4.4 Increased Accuracy of the Registration of Data -- 4.4.1 Establishing Drawing Rules for Scientists -- 4.4.2 Standardization of Compound Representation -- 4.4.3 Three Roles and Two Staging Areas -- 4.4.4 Batch Reassignment -- 4.4.4.1 Unknown Compounds Management -- 4.4.5 Automatic Processes -- 4.5 Implementation of the Platform -- 4.5.1 Database -- 4.5.2 Software -- 4.5.3 Data Migration and Transformation of Names into Structures -- 4.6 Linking Chemical Information to Analytical Data -- 4.7 Linking Chemicals to Bioactivity Data -- 4.8 Conclusions -- References -- Part Two: Analysis and Enrichment -- 5 Data Mining of Plant Metabolic Pathways -- 5.1 Introduction -- 5.1.1 The Importance of Understanding Plant Metabolic Pathways.

5.1.2 Pathway Modeling and Its Prerequisites -- 5.2 Pathway Representation -- 5.2.1 Compounds -- 5.2.1.1 The Importance of Having Uniquely Defined Molecules -- 5.2.1.2 Representation Formats -- 5.2.1.3 Key Chemical Compound Databases -- 5.2.2 Reactions -- 5.2.2.1 Definitions of Reactions -- 5.2.2.2 Importance of Stoichiometry and Mass Balance -- 5.2.2.3 Atom Tracing -- 5.2.2.4 Storing Enzyme Information: EC Numbers and Their Limitations -- 5.2.3 Pathways -- 5.2.3.1 How Are Pathways Defined? -- 5.2.3.2 Typical Size and Distinction between Pathways and Superpathways -- 5.3 Pathway Management Platforms -- 5.3.1 Kyoto Encyclopedia of Genes and Genomes (KEGG) -- 5.3.1.1 Database Structure in KEGG -- 5.3.1.2 Navigation through KEGG -- 5.3.2 The Pathway Tools Platform -- 5.3.2.1 Database Management in Pathway Tools -- 5.3.2.2 Content Creation and Management with Pathway Tools -- 5.3.2.3 Pathway Tools' Visualization Capability -- 5.4 Obtaining Pathway Information -- 5.4.1 "Ready-Made" Reference Pathway Databases and Their Contents -- 5.4.1.1 KEGG -- 5.4.1.2 MetaCyc and PlantCyc -- 5.4.1.3 MetaCrop -- 5.4.2 Integrating Databases and Issues Involved -- 5.4.2.1 Compound Ambiguity -- 5.4.2.2 Reaction Redundancy -- 5.4.2.3 Formats for Exchanging Pathway Data -- 5.4.3 Adding Information to Pathway Databases -- 5.4.3.1 Manual Curation -- 5.4.3.2 Automated Methods for Literature Mining -- 5.5 Constructing Organism-Specific Pathway Databases -- 5.5.1 Enzyme Identification -- 5.5.1.1 Reference Enzyme Databases -- 5.5.1.2 Enzyme Function Prediction Using Protein Sequence Information -- 5.5.1.3 Enzyme Function Inference Using 3D Protein Structure Information -- 5.5.2 Pathway Prediction from Available Enzyme Information -- 5.5.2.1 Pathway "Painting" Using KEGG Reference Maps -- 5.5.2.2 Pathway Reconstruction with Pathway Tools -- 5.5.3 Examples of Pathway Reconstruction.

5.6 Conclusions -- References -- 6 The Role of Data Mining in the Identification of Bioactive Compounds via High-Throughput Screening -- 6.1 Introduction to the HTS Process: the Role of Data Mining -- 6.2 Relevant Data Architectures for the Analysis of HTS Data -- 6.2.1 Conditions (Parameters) for Analysis of HTS Screens -- 6.2.1.1 Purity -- 6.2.1.2 Assay Conditions -- 6.2.1.3 Previous Performance of Samples -- 6.2.2 Data Aggregation System -- 6.3 Analysis of HTS Data -- 6.3.1 Analysis of Frequent Hitters and Undesirable Compounds in Hit Lists -- 6.3.2 Analysis of Cell-Based Screening Data Leading to Mode of Mechanism Hypotheses -- 6.4 Identification of New Compounds via Compound Set Enrichment and Docking -- 6.4.1 Identification of Hit Series and SAR from Primary Screening Data by Compound Set Enrichment -- 6.4.2 Molecular Docking -- 6.5 Conclusions -- References -- 7 The Value of Interactive Visual Analytics in Drug Discovery: An Overview -- 7.1 Creating Informative Visualizations -- 7.2 Lead Discovery and Optimization -- 7.2.1 Common Visualizations -- 7.2.1.1 SAR Tables -- 7.2.1.2 Scatter Plots -- 7.2.1.3 Histograms -- 7.2.2 Advanced Visualizations -- 7.2.2.1 Profile Charts -- 7.2.2.2 Dose-Response Curves -- 7.2.2.3 Heat Maps -- 7.2.3 Interactive Analysis -- 7.3 Genomics -- 7.3.1 Common Visualizations -- 7.3.1.1 Hierarchical Clustered Heat Map -- 7.3.1.2 Scatter Plot in Log Scale -- 7.3.1.3 Histograms and Box Plots for Quality Control -- 7.3.1.4 Karyogram (Chromosomal Map) -- 7.3.2 Advanced Visualizations -- 7.3.2.1 Metabolic Pathways -- 7.3.2.2 Gene Ontology Tree Maps -- 7.3.2.3 Clustered All to All "Heat Maps" (Triangular Heat Map) -- 7.3.3 Applications -- 7.3.3.1 Understanding Diseases by Comparing Healthy with Unhealthy Tissue or Patients -- 7.3.3.2 Measure Effects of Drug Treatment on a Cellular Level -- References.

8 Using Chemoinformatics Tools from R -- 8.1 Introduction -- 8.2 System Call -- 8.2.1 Prerequisite -- 8.2.2 The Command System() -- 8.2.3 Example, Command Edition, and Outputs -- 8.3 Shared Library Call -- 8.3.1 Shared Library -- 8.3.2 Name Mangling and Calling Convention -- 8.3.3 dyn.load and dyn.unload -- 8.3.4 .C and .Fortran -- 8.3.5 Example -- 8.3.6 Compilation -- 8.4 Wrapping -- 8.4.1 Why Wrapping -- 8.4.2 Using R Internals -- 8.4.3 How to Keep an SEXP Alive -- 8.4.4 Binding to C/C++ Libraries -- 8.5 Java Archives -- 8.5.1 The Package rJava -- 8.5.2 The Package rcdk -- 8.6 Conclusions -- References -- Part Three: Applications to Polypharmacology -- 9 Content Development Strategies for the Successful Implementation of Data Mining Technologies -- 9.1 Introduction -- 9.2 Knowledge Challenges in Drug Discovery -- 9.3 Case Studies -- 9.3.1 Thomson Reuters Integrity -- 9.3.1.1 Knowledge Areas -- 9.3.1.2 Search Fields -- 9.3.1.3 Data Management Features -- 9.3.1.4 Use of Integrity in the Industry and Academia -- 9.3.2 ChemBioBank -- 9.3.3 Molecular Libraries Program -- 9.4 Knowledge-Based Data Mining Technologies -- 9.4.1 Problem Transformation Methods -- 9.4.2 Algorithm Adaptation Methods -- 9.4.3 Training a Mechanism of Action Model -- 9.5 Future Trends and Outlook -- References -- 10 Applications of Rule-Based Methods to Data Mining of Polypharmacology Data Sets -- 10.1 Introduction -- 10.2 Materials and Methods -- 10.2.1 Data Set Preparation -- 10.2.2 Preparation of the σ-1 Binders Data Set -- 10.2.3 Association Rules -- 10.2.4 Novel Hybrid Structures by Fragment Swapping -- 10.3 Results -- 10.3.1 Rules Generation and Extraction -- 10.3.1.1 Rules Describing the Polypharmacology Space -- 10.3.1.2 Optimization of σ-1 with Selectivity Over D2 -- 10.3.1.3 Optimization of σ-1 with Selectivity over D2 and 5HT2 -- 10.4 Discussion -- 10.5 Conclusions.

References.