QSPR/QSAR Analysis Using SMILES and Quasi-SMILES

Description

Preface Contents Contributors Abbreviations Part I Theoretical Conceptions 1 Fundamentals of Mathematical Modeling of Chemicals Through QSPR/QSAR 1.1 Introduction 1.2 QSPR/QSAR: Tools and Tasks 1.3 Five OECD Principles 1.4 Praxis of the QSPR/QSAR Development 1.5 Molecular Descriptors are the Basis for the QSPR/QSAR 1.5.1 Principal Component Analysis 1.5.2 Multiple Linear Regressions 1.5.3 Partial Least Squares 1.5.4 K-Nearest Neighbor Classification 1.5.5 Artificial Neural Network 1.5.6 Support Vector Machine 1.5.7 Random Forest 1.5.8 Monte Carlo Method 1.5.9 Data Curation 1.6 Reproducibility 1.6.1 Applicability Domain 1.6.2 Model Validation 1.7 Recommendations for Building Robust QSPR/QSAR Models 1.8 Is It Possible to Obtain Correlations Suitable for QSPR/QSAR Using SMILES? 1.9 The Main Quality of a Descriptor Is to Indicate the Differences Between Molecules 1.10 Significant Notes 1.11 Conclusions References 2 Molecular Descriptors in QSPR/QSAR Modeling 2.1 Introduction 2.1.1 History 2.1.2 QSPR/QSAR Modeling 2.1.3 Molecular Descriptors 2.2 Descriptors for Nano-QSPR/QSAR 2.3 SMILES and Quasi-SMILES Descriptors 2.3.1 Quasi-SMILES Examples in Peer-Reviewed Papers 2.4 Software for Generation of Molecular Descriptors 2.5 Conclusion and Future Direction References 3 Application of SMILES to Cheminformatics and Generation of Optimum SMILES Descriptors Using CORAL Software 3.1 Introduction 3.1.1 The CORAL software description 3.1.2 An Example of Model Training and Validation (Graphically) 3.2 Conclusions  References Part II SMILES Based Descriptors 4 All SMILES Variational Autoencoder for Molecular Property Prediction and Optimization 4.1 Introduction 4.1.1 Summary of Novel Contributions 4.2 Efficient Molecular Encoding with Multiple SMILES Strings 4.3 Review of Recurrent Neural Networks 4.4 All SMILES VAE Architecture 4.4.1 Computational Complexity 4.4.2 Latent Space Optimization 4.5 Datasets 4.5.1 ZINC 4.5.2 Tox21 4.6 Results 4.6.1 Reconstruction Accuracy and Validity 4.6.2 Property Prediction 4.6.3 Molecular Optimization 4.6.4 Ablation of Model Components 4.7 SMILES Grammar Can Be Enforced with a Pushdown Automaton 4.7.1 Ringbond and Valence Shell Semantic Constraints 4.7.2 Redundancy in Graph-Based and SMILES Representations of Molecules 4.8 Conclusion References 5 SMILES-Based Bioactivity Descriptors to Model the Anti-dengue Virus Activity: A Case Study 5.1 Introduction 5.2 Materials and Methods 5.2.1 Importance of Bioactivity Descriptors 5.2.2 Dataset Collection 5.2.3 Calculation of Molecular Descriptors 5.2.4 Development of Linear 2D-QSAR Models 5.2.5 Statistical Analysis of Models 5.2.6 Applicability Domain of the Models 5.2.7 Non-linear Model Development 5.3 Results and Discussion 5.4 Conclusions References Part III SMILES for QSPR/QSAR with Optimal Descriptors 6 QSPR Models for Prediction of Redox Potentials Using Optimal Descriptors 6.1 Introduction, Redox Potential, and Its Significance 6.2 Relationship Between Redox Potential and Structure 6.3 Optimal Descriptors in QSPR of Redox Potential 6.3.1 Basic Principles of Employing Optimal Descriptors in QSPR 6.3.2 Published Studies on SMILES-Based QSPR for Redox Potential 6.3.3 Case Study of Two Large Data Sets 6.4 Conclusions References 7 Building Up QSPR for Polymers Endpoints by Using SMILES-Based Optimal Descriptors 7.1 Introduction 7.1.1 The General Scheme of QSPR/QSAR Analysis of Endpoints Related to Polymers 7.1.2 QSPR Analysis of Endpoints Related to Polymers with MLR 7.1.3 QSPR/QSAR Analysis of Endpoints Related to Polymers with PLS 7.1.4 QSPR Analysis of Endpoints Related to Polymers with ANN 7.1.5 QSPR Analysis of Endpoints Related to Polymers with SVM 7.2 Significant Notes 7.3 Building Up Models of Polymers Endpoints Using SMILES 7.3.1 SMILES 7.3.2 Optimal SMILES-Based Descriptors 7.3.3 The Monte Carlo Optimization Procedure 7.3.4 The Classic Scheme of Building Up the QSPR/QSAR Model Using the Optimal Descriptors 7.3.5 The Balance of Correlations for the QSPR/QSAR Model Using the Optimal Descriptors 7.3.6 Search and Use for Reliable Criteria of the Predictive Potential of QSPR/QSAR Models Based on the Optimal Descriptors 7.3.7 Hybrid Optimal Descriptors 7.3.8 Model Complication 7.4 Examples of Improving Models Built Up with Optimal Descriptors 7.4.1 Development of a New Conception to Building Up a Model 7.4.2 QSPR Models for the Glass Transition Temperature 7.4.3 QSPR Models for the Refractive Index 7.5 Comparison QSPR-Models 7.6 Possible Ways of Evolution of the QSPR for Polymers 7.7 Quasi-SMILES Can Be a Tool for the Discussion of Experimentalists and Model Developers 7.8 Conclusions References Part IV Quasi-SMILES for QSPR/QSAR 8 Quasi-SMILES-Based QSPR/QSAR Modeling 8.1 Introduction 8.2 Principals of QSPR/QSAR Models 8.3 Monte Carlo Technique for Nano-QSPR/QSAR 8.3.1 SMILES and Quasi-SMILES 8.3.2 The Main Step for QSPR/QSAR Modeling by SMILES or Quasi-SMILES 8.4 Examples of Quasi-SMILES-Based QSPR/QSAR Models 8.5 Conclusion and Future Direction References 9 Quasi-SMILES-Based Mathematical Model for the Prediction of Percolation Threshold for Conductive Polymer Composites 9.1 Introduction 9.2 Theoretical Background of the Percolation Threshold 9.2.1 Effect of the Conductive Fillers 9.2.2 Effect of the Host Polymers 9.3 Methods for the Synthesis of Conductive Polymers 9.3.1 Chemical Method 9.3.2 Metathesis Method 9.3.3 Photochemical Method 9.3.4 Electro-Chemical Method 9.3.5 Plasma Polymerisation 9.3.6 Solid-State Method 9.3.7 Inclusion Method 9.4 Various Properties of Conducting Polymers 9.4.1 Magnetic Properties 9.4.2 Optical Properties 9.4.3 Electrical Properties 9.5 Applications of Conductive Polymers 9.5.1 Sensors 9.5.2 Solar Cells 9.5.3 Supercapacitors 9.5.4 Data Storage Transistors 9.5.5 Batteries 9.6 Mathematical Models for the Prediction of Percolation Threshold 9.6.1 Data and Building the Quasi-SMILES Codes 9.6.2 Optimal Descriptor 9.7 Results and Discussion 9.8 Conclusion References 10 On the Possibility to Build up the QSAR Model of Different Kinds of Inhibitory Activity for a Large List of Human Intestinal Transporter Using Quasi-SMILES 10.1 Introduction 10.1.1 Literature Review on Various QSAR Models for Human Intestinal Transporter 10.1.2 An Overview of Computer Simulations Study of Human Intestinal Transporter 10.2 Materials and Methods 10.2.1 Experimental Data Curation 10.2.2 Development of the Models 10.3 Result and Discussion 10.4 Conclusion References 11 Quasi-SMILES as a Tool for Peptide QSAR Modelling 11.1 Introduction 11.2 A Brief Overview of QSAR 11.3 Peptide QSAR Modelling 11.4 SMILES-Based Descriptors for QSAR Model Development 11.5 Quasi-SMILES 11.5.1 Development of QSAR Model by Quasi-SMILES 11.5.2 Optimal Descriptor Approach 11.6 Different Application of SMILES/Quasi-SMILES in Peptide QSPR/QSAR Modelling 11.6.1 Antimicrobial Peptides 11.6.2 Epitope Peptides with Class I Major Histocompatibility Complex (MHC) 11.7 Mathematical Approaches Used for Peptide QSAR Modelling 11.7.1 Multiple Linear Regressions (MLR) 11.7.2 Partial Least Square (PLS) 11.7.3 Principal Component Analysis (PCA) 11.7.4 Genetic Algorithm (GA)-Based Peptide QSAR 11.7.5 Particle Swarm Optimization Algorithm (PSO) 11.7.6 Artificial Neural Network (ANN) 11.7.7 Support Vector Machine (SVM) 11.7.8 Other Methods 11.8 Conclusions References Part V SMILES and Quasi-SMILES for QSPR/QSAR 12 SMILES and Quasi-SMILES Descriptors in QSAR/QSPR Modeling of Diverse Materials Properties in Safety and Environment Application 12.1 Introduction 12.1.1 QSAR/QSPR Methods 12.1.2 Brief Description of the QSAR/QSPR Methodology 12.2 SMILES and Quasi-SMILES Descriptors 12.3 Study of Several Important Properties/Activities in Safety and Environmental Applications 12.3.1 The Cytotoxicity of Metal Oxide Nanoparticles 12.3.2 Flammability Properties of Chemicals and Their Mixtures 12.3.3 Thermal Hazard Properties of Ionic Liquids and Their Mixtures 12.3.4 Toxicity of Ionic Liquids and Their Mixtures 12.4 Limitations and Outlook in Safety and Environmental Applications 12.4.1 Limitations 12.4.2 Outlook References 13 SMILES and Quasi-SMILES in QSAR Modeling for Prediction of Physicochemical and Biochemical Properties 13.1 Introduction 13.2 Fundamentals of SMILES and Quasi-SMILES 13.3 Application of SMILES and Quasi-SMILES-Based QSAR Model 13.3.1 Nanoparticles Toxicity and Property Prediction 13.3.2 Toxicity Predictions and Risk Assessment of Organic Chemicals 13.3.3 Miscellaneous Physicochemical and Biochemical Property Predictions of Organic Chemicals 13.4 Conclusion References Part VI Possible Ways of Nano-QSPR/Nano-QSAR Evolution 14 The CORAL Software as a Tool to Develop Models for Nanomaterials’ Endpoints 14.1 Introduction 14.2 Theory and Practices of QSPR/QSAR 14.3 SMILES and Nanomaterials 14.4 Quasi-SMILES and Nanomaterials 14.5 Optimal SMILES-Based Descriptor 14.6 The Monte Carlo Optimization 14.7 Conclusions References 15 Employing Quasi-SMILES Notation in Development of Nano-QSPR Models for Nanofluids 15.1 Introduction 15.1.1 Nanofluids 15.1.2 Theoretical Methods Applied for Study of Nanofluids’ Properties 15.1.3 The Importance of QSPR Study for Nanofluids 15.2 Methodology of CORAL-Based Models Generation 15.2.1 Collection of a Valid Data Set 15.2.2 Quasi-SMILES for Nanofluids 15.2.3 Optimal Descriptors, Predictability Criteria, and Optimization 15.3 Successful Nano-QSPR Studies on Nanofluids 15.4 Conclusion and Perspective Outlook References Part VII Possible Ways of QSPR/QSAR Evolution in the Future 16 On Complementary Approaches of Assessing the Predictive Potential of QSPR/QSAR Models 16.1 Introduction 16.2 Software for Building Up QSPR/QSAR Models 16.3 The Critical Analysis of Existing Approaches to Assessing the Predictive Potential 16.4 Convenience and Inconvenience of Correlation 16.5 Convenience and Inconvenience of Causation 16.6 Note on “Secrets of QSPR/QSAR” 16.7 Index Ideality of Correlation (IIC) 16.8 Correlation Intensity Index (CII) 16.9 Can IIC and CII Be Useful? 16.10 Is It Possible to Improve the Predictive Potential of Such Models Using IIC? 16.11 Is It Possible to Improve the Predictive Potential of Such Models Using CII? 16.12 Testing Assumptions About the Significance of IIC and CII 16.13 The Comparison of Criteria of the Predictive Potential of QSPR/QSAR 16.14 The System of Self-consistent Models 16.14.1 Examples of Successful Applications of Self-consistent Models 16.15 Conclusions References 17 CORAL: Predictions of Quality of Rice Based on Retention Index Using a Combination of Correlation Intensity Index and Consensus Modelling 17.1 Introduction 17.2 Materials and Method 17.2.1 Data 17.2.2 Model 17.2.3 Optimal Descriptor 17.2.4 Monte Carlo Optimization 17.2.5 Applicability Domain 17.2.6 Validation 17.2.7 Consensus Modelling 17.3 Results and Discussion 17.3.1 QSRR Modelling and Validation 17.3.2 Mechanistic Interpretation 17.3.3 Consensus Modelling 17.4 Conclusions References Index