Chemometrics and Cheminformatics in Aquatic Toxicology

Description

CHEMOMETRICS AND CHEMINFORMATICS IN AQUATIC TOXICOLOGY Explore chemometric and cheminformatic techniques and tools in aquatic toxicology Chemometrics and Cheminformatics in Aquatic Toxicology delivers an exploration of the existing and emerging problems of contamination of the aquatic environment through various metal and organic pollutants, including industrial chemicals, pharmaceuticals, cosmetics, biocides, nanomaterials, pesticides, surfactants, dyes, and more. The book discusses different chemometric and cheminformatic tools for non-experts and their application to the analysis and modeling of toxicity data of chemicals to various aquatic organisms. You’ll learn about a variety of aquatic toxicity databases and chemometric software tools and webservers as well as practical examples of model development, including illustrations. You’ll also find case studies and literature reports to round out your understanding of the subject. Finally, you’ll learn about tools and protocols including machine learning, data mining, and QSAR and ligand-based chemical design methods. Readers will also benefit from the inclusion of: A thorough introduction to chemometric and cheminformatic tools and techniques, including machine learning and data miningAn exploration of aquatic toxicity databases, chemometric software tools, and webserversPractical examples and case studies to highlight and illustrate the concepts contained within the bookA concise treatment of chemometric and cheminformatic tools and their application to the analysis and modeling of toxicity data Perfect for researchers and students in chemistry and the environmental and pharmaceutical sciences, Chemometrics and Cheminformatics in Aquatic Toxicology will also earn a place in the libraries of professionals in the chemical industry and regulators whose work involves chemometrics. Cover Half Title Chemometrics and Cheminformatics in Aquatic Toxicology Copyright Contents Preface Part I. Introduction 1. Water Quality and Contaminants of Emerging Concern (CECs) 1.1 Introduction: Water Quality and Emerging Contaminants 1.2 Contaminants of Emerging Concern 1.2.1 Pharmaceuticals 1.2.2 Personal Care Products 1.2.3 Nanomaterials 1.2.4 Plasticizers 1.2.5 Surfactants and Metabolites 1.2.6 Flame Retardants 1.2.7 Industrial Additives and Agents 1.2.8 Anticorrosives and Antifouling Agents 1.2.9 Natural Emerging Contaminants: Mycotoxins and Phytotoxins 1.3 Summary and Recommendations for Future Research References 2. The Effects of Contaminants of Emerging Concern on Water Quality 2.1 Introduction 2.1.1 Sources of CECs to the Aquatic Ecosystem 2.1.2 Fate of CECs in Aquatic Environments 2.2 Assessing the Effects of CECs in Aquatic Life 2.2.1 Pharmaceuticals 2.2.2 Personal Care Products 2.2.3 Agricultural Pesticides 2.2.4 Industrial Chemicals 2.3 Multiple Stressors 2.3.1 Mixtures of CECs 2.3.2 Interactions of CECs and Other Environmental Stressors 2.3.3 Climate Change 2.4 Conclusions Acknowledgments References 3. Chemometrics: Multivariate Statistical Analysis of Analytical Chemical and Biomolecular Data 3.1 Introduction 3.2 Historic Origins 3.3 Applied Statistics 3.4 Analytical and Physical Chemistry 3.5 Scientific Computing 3.6 Development from the 1980s 3.7 A Review of the Main Methods 3.8 Experimental Design 3.9 Principal Components Analysis and Pattern Recognition 3.10 Multivariate Signal Analysis 3.11 Multivariate Calibration 3.12 Digital Signal Processing and Time Series Analysis 3.13 Multiway Methods 3.14 Conclusion References 4. An Introduction to Chemometrics and Cheminformatics 4.1 Brief History of Chemometrics/Cheminformatics 4.2 Current State of Cheminformatics 4.3 Common Cheminformatics Tasks 4.4 Cheminformatics Toolbox 4.5 Conclusion References Part II. Chemometric and Cheminformatic Tools and Protocols 5. An Introduction to Some Basic Chemometric Tools 5.1 Introduction 5.2 Example Datasets 5.2.1 Example 1 – The Mono-Substituted Nitrobenzenes Dataset 5.2.2 Example 2 – The Oil Offshore Production Emission Dataset 5.3 Data Analytical Methods 5.3.1 Pretreatment Methods 5.3.2 Principal Components Analysis (PCA) 5.3.3 Partial Least Squares Projections to Latent Structures (PLS) 5.3.4 Orthogonal Partial Least Squares (OPLS®) 5.3.5 Cross-Validation 5.4 Results 5.4.1 Results for Example 1 5.4.2 Results for Example 2 5.5 Discussion References 6. From Data to Models: Mining Experimental Values with Machine Learning Tools 6.1 Introduction 6.2 Data and Models 6.2.1 Data 6.2.2 Models 6.3 Basic Methods in Model Development with ML 6.3.1 Inputs to the Model 6.3.2 Output of the Model 6.3.3 Basic Algorithms 6.3.4 Evaluating What the Model Has Learned from Data 6.3.5 Model Interpretability 6.4 More Advanced ML Methodologies 6.4.1 Classifiers: from Decision Trees to Ensemble 6.4.2 Mining Datasets to Extract Frequent Subgroups 6.4.3 Kernel Methods and Support Vector Machine (SVM) 6.4.4 From Perceptron to Neural Nets 6.5 Deep Learning 6.5.1 Main DNN Architectures 6.5.2 Interpretation of DNN Models 6.5.3 Consequences of Deep Learning for QSAR 6.6 Conclusions References 7. Machine Learning Approaches in Computational Toxicology Studies 7.1 Introduction 7.1.1 Computer-Based Toxicity Prediction 7.1.2 Brief History of QSAR and Modern Machine Learning Techniques 7.2 Toxicity Data Set Preparation 7.2.1 Data Collection and Chemical Structure Representation 7.2.2 Descriptors and Fingerprints 7.3 Machine-Learning Techniques 7.3.1 Unsupervised Learning 7.3.2 Supervised Learning 7.3.3 Semi-Supervised Learning 7.4 Model Evaluation 7.5 Freely Available Software Tools and Open-Source Libraries Relevant to Machine Learning 7.6 Concluding Remarks Acknowledgment References 8. Counter-Propagation Neural Networks for Modeling and Read Across in Aquatic (Fish) Toxicity 8.1 Introduction 8.2 Examples of Counter-Propagation Artificial Neural Networks in Fish Toxicity Modeling 8.3 Counter-Propagation Artificial Neural Networks 8.4 Conclusions References 9. Aiming High versus Aiming All: Aquatic Toxicology and QSAR Multitarget Models 9.1 Introduction 9.2 Multitarget QSARS and Aquatic Toxicology 9.2.1 Multitarget QSARS: Basics Overview 9.2.2 Mt-QSAR and the Biotarget Perspective: A Review from Selected Works 9.3 Biotargets and Aqueous Environmental Assessment: Solutions and Recommendations 9.4 Future Perspectives and Conclusion References 10. Chemometric Approaches to Evaluate Interspecies Relationships and Extrapolation in Aquatic Toxicity 10.1 Introduction 10.2 Acute Toxicity Estimation 10.2.1 Quantitative Structure–Activity Relationship (QSAR) Models 10.2.2 Interspecies Correlation Estimation (ICE) Models 10.2.3 Species Sensitivity Distributions (SSDs) 10.2.4 Linking Acute Toxicity Models 10.3 Sublethal Toxicity Extrapolation 10.3.1 Genomics and Sequence-Based Relationships 10.3.2 Chemical Proteomics 10.3.3 Differential Expression and Pathway Analysis 10.4 Discussion 10.5 Conclusions Disclaimer References Part III. Case Studies and Literature Reports 11. The QSAR Paradigm to Explore and Predict Aquatic Toxicity 11.1 Introduction 11.2 Application of QSAR Methodology to Predict Aquatic Toxicity 11.2.1 Overview 11.2.2 Aquatic Toxicity Endpoints and Relevant Databases 11.2.3 Criteria for Robust QSAR Models 11.2.4 MOA-Based Aquatic Toxicity QSAR (QSTR) 11.2.5 Software Tools for Ecotoxicological Endpoints 11.3 QSAR for Narcosis – The Impact of Hydrophobicity 11.3.1 Linear Solvation Energy Relationships for Narcosis 11.3.2 Application of Chromatographic Systems for Building Narcotic Models 11.4 Excess Toxicity – Overview 11.4.1 QSAR (QSTR) Models for Reactive and Specific Acting Chemicals 11.5 Predictions of Bioconcentration Factor 11.6 Conclusions References 12. Application of Cheminformatics to Model Fish Toxicity 12.1 Introduction 12.2 Fish Toxicities 12.3 Toxicity in Fish Families and Species 12.4 The Fathead Minnow, the Rainbow Trout, and the Bluegill 12.5 Toxicity Variations in FIT Compounds 12.6 Modeling Wide-Range Toxicity Compounds 12.7 Further Evaluations 12.8 Alternative Approaches 12.9 Mechanisms of Action 12.10 Conclusions Acknowledgments References 13. Chemometric Modeling of Algal and Daphnia Toxicity 13.1 Introduction 13.2 Algae Class 13.2.1 Short Characterization of Algae Class 13.2.2 QSAR Models Developed Using the Algae 13.3 Daphniidae Family 13.3.1 Short Characterization of Daphniidae Family 13.3.2 QSAR Models Developed Using Daphnia magna 13.4 Interspecies Correlation Estimation for Algal and Daphnia Aquatic Toxicity 13.4.1 Algal and Daphnia Toxicity Correlation 13.4.2 Algal, Daphnia and Other Species Toxicity Correlation 13.4.3 Daphnia and Other Species Toxicity Correlation 13.4.4 Algae Species Toxicity Correlations 13.4.5 Algal and Other Species Toxicity Correlation 13.5 Conclusions Abbreviations List References 14. Chemometric Modeling of Algal Toxicity 14.1 Introduction 14.1.1 Environmental Importance of Algae 14.1.2 OECD Principles 14.1.3 Brief Summary of Algal QSAR Models 14.2 Criteria Set for the Comparison of Selected QSAR Models 14.2.1 The Modeled Endpoints 14.2.2 Descriptors 14.2.3 Model Performance 14.2.4 Applicability Domain 14.2.5 Software Used for QSAR Modeling 14.3 Literature MLR Studies on Algae 14.4 Conclusion References 15. Chemometric Modeling of Daphnia Toxicity 15.1 Introduction 15.2 QSTR and QSTTR Analyses 15.3 QSTR/QSTT/QSTTR Modeling of Daphnia Toxicity 15.3.1 Categorized Chemicals 15.3.2 Non-categorized Chemicals 15.4 Mechanistic Interpretations of Chemometric Models 15.5 Conclusive Remarks and Future Directions Acknowledgment References 16. Chemometric Modeling of Daphnia Toxicity: Quantum-Mechanical Insights 16.1 Introduction 16.2 Quantum-Mechanical Methods 16.3 Quantum-Mechanical Descriptors for Daphnia Toxicity 16.4 Concluding Remarks and Future Outlook References 17. Chemometric Modeling of Toxicity of Chemicals to Tadpoles 17.1 Introduction 17.2 Overview and Morphology of Tadpoles 17.2.1 Tadpole as a Target for Ecotoxicity Testing 17.3 Reports of Tadpole Toxicity Due Various Environmental Contaminants: What Do We Know So Far? 17.4 In silico Models Reported for Tadpole Ecotoxicity: A Literature Review 17.5 Application of QSARs or Related Approaches in Modeling Tadpole Toxicity: A Future Perspective 17.6 Conclusion Acknowledgment References 18. Chemometric Modeling of Toxicity of Chemicals to Marine Bacteria 18.1 Introduction 18.1.1 Marine Bacteria: A Source of Ocean’s Wealth 18.1.2 Morphology of Marine Bacteria 18.1.3 Marine Bacteria in Symbiotic Association with Other Species 18.1.4 Marine Bacteria as Nitrogen Fixers 18.2 Marine Bacteria and Their Role in Nitrogen Fixing 18.2.1 Marine Bacteria That Actually “Fix” Nitrogen 18.2.2 Marine Bacteria Which Are Involved in Nitrification 18.2.3 Primary Producers Marine Bacteria Those Who Do Not Fix Nitrogen 18.3 Marine Bacteria as Biomarkers for Ecotoxicity Estimation 18.4 Chemometric Tools Applied in Ecotoxicity Evaluation of Marine Bacteria 18.4.1 Ecotoxicity Evaluations of Organic Compounds 18.4.2 Ecotoxicity Evaluations Using Capacity Factors (k) 18.4.3 Ecotoxicity Evaluations of Shale Oil Components 18.4.4 Ecotoxicity Evaluations of Human Pharmaceuticals 18.4.5 Ecotoxicity Evaluations of Ionic Liquids (ILs) 18.5 Conclusion Acknowledgment References 19. Chemometric Modeling of Pesticide Aquatic Toxicity 19.1 Introduction 19.2 QSARs Models 19.2.1 QSAR Models Developed Using Fish Species 19.2.2 QSAR Models Developed Using Zebrafish Embryos 19.2.3 QSAR Models Developed Using Algae Species 19.2.4 QSAR Models Developed Using Americamysis bahia Species 19.2.5 QSAR Models Developed Using Daphnia magna 19.2.6 QSAR/QAAR Models Developed Using Interspecies Correlations 19.3 Conclusions References 20. Contribution of Chemometric Modeling to Chemical Risks Assessment for Aquatic Plants: State-of-the-Art 20.1 Introduction 20.2 Definition and Classification 20.3 Advantage of Aquatic Plants 20.3.1 Ecosystems Benefits 20.3.2 Economic Benefits 20.3.3 Phytoremediation Using Aquatic Plants 20.4 Contaminants and Their Toxicity 20.5 Chemometrics for Aquatic Plants Toxicity 20.6 Review of Literature on Chemometrics for Aquatic Plants Toxicity 20.6.1 Toxicity of Pharmaceuticals 20.6.2 Toxicity of Pesticides 20.6.3 Toxicity of Nanoparticles 20.6.4 Toxicity of Heavy Metal and Metalloids 20.6.5 Toxicity of Others Pollutants 20.7 Conclusions References 21. Application of 3D-QSAR Approaches to Classification and Predictionof Aquatic Toxicity 21.1 Introduction 21.1.1 Environmental Risk Assessment of Chemicals 21.1.2 In silico Models in Environmental Risk Assessment 21.1.3 Introduction and Limitation of the Previous QSAR Approaches 21.1.4 Challenges and Improvement Through 3D-QSAR 21.2 Principles of CAPLI 3D-QSAR 21.2.1 Docking Protocols 21.2.2 Data Preparation 21.2.3 Structure-based Pharmacophore and 3D-fingerprint Descriptors 21.2.4 CAPLI 3D-QSAR Development and Validation 21.2.5 Prediction of Binding Mode and Affinity 21.3 Applications in Chemical Classification and Toxicity Prediction 21.3.1 Mechanism-Based Classification of OP Inhibitors 21.3.2 Species Susceptibility Prediction 21.3.3 Structure–Toxicity Relationship Analysis 21.4 Limitation and Potential Improvement 21.4.1 Convolutional Neural Network 21.5 Conclusions and Recommendations Acknowledgments References 22. QSAR Modeling of Aquatic Toxicity of Cationic Polymers 22.1 Introduction 22.2 Materials and Methods 22.2.1 Polymers 22.2.2 Dataset 22.2.3 Descriptor Calculation 22.2.4 Dataset Division 22.2.5 Model Development 22.2.6 Model Validation 22.3 Results and Discussion 22.3.1 QSTR Modeling for Fish Toxicity 96 h Dataset 22.3.2 QSTR Modeling for Daphnia magna Toxicity 48 h Dataset 22.3.3 QSTR Modeling for Green Algae Toxicity 96 h Dataset 22.3.4 QSTR Modeling for Chronic Toxicity Against Green Algae 22.3.5 Interspecies Modeling of Polymers 22.4 Conclusions Acknowledgments References Part IV. Tools and Databases 23. In Silico Platforms for Predictive Ecotoxicology: From Machine Learning to Deep Learning 23.1 Introduction 23.2 Machine Learning and Deep Learning 23.2.1 Support Vector Machines 23.2.2 Random Forest 23.2.3 Deep Neural Networks 23.3 Toxicity Prediction Modeling 23.3.1 General Procedure of Modeling 23.3.2 QSAR 23.3.3 Molecular Docking 23.3.4 Read-Across 23.3.5 Structural Alerts 23.3.6 Adverse Outcome Pathway 23.4 Challenges and Future Directions References 24. The Use and Evolution of Web Tools for Aquatic Toxicology Studies 24.1 Introduction 24.2 Methodologies Used in Aquatic Toxicology Tests 24.2.1 Database 24.2.2 Toxicity 24.2.3 Quantitative Structure–Activity Relationships (QSARs) Between Chemical Structures and Biological Activity in Aquatic Toxicity Studies 24.3 Web Tools Used in Aquatic Toxicology 24.3.1 Aggregated Computational Toxicology Online Resource (ACToR) 24.3.2 ECOTOXicology (ECOTOX) 24.3.3 OASIS 24.3.4 TOXMATCH 24.3.5 OSIRIS 24.3.6 BIOWIN Models 24.3.7 AdmetSar 24.3.8 Chembench 24.3.9 Ecological Structure–Activity Relationships (ECOSAR) 24.3.10 OECD QSAR Toolbox 24.3.11 PASS 24.3.12 Applications of in silico Techniques to Aquatic Toxicology Tests 24.4 Perspectives References 25. The Tools for Aquatic Toxicology within the VEGAHUB System 25.1 Introduction 25.2 The VEGA Models 25.2.1 The VEGA Models for Aquatic Toxicity 25.2.2 The Example of the Fish Acute Toxicity Model Developed Using Neural Networks 25.2.3 The Differences Between the Aquatic Toxicity Models 25.2.4 The Components of the Applicability Domain Index 25.2.5 The Evaluation of the Results of the VEGA Models 25.3 ToxRead and Read-Across Within VEGAHUB 25.4 Prometheus and JANUS 25.5 The Future Developments 25.5.1 The VERMEER Project 25.5.2 The toDIVINE Project 25.6 Conclusions References 26. Aquatic Toxicology Databases 26.1 Introduction 26.2 Aquatic Toxicity 26.2.1 Aquatic Toxicity Test 26.2.2 Aquatic Test Species 26.3 Importance of Aquatic Toxicity Databases 26.4 Characteristic of an Ideal Aquatic Toxicity Database 26.5 Aquatic Toxicology Databases 26.5.1 Acute Toxicity Database 26.5.2 Aquatic Toxicity Information Retrieval (AQUIRE) 26.5.3 Ecotoxicology Database (ECOTOX) 26.5.4 Environmental Residue Effects Database (ERED) 26.5.5 EnviroTox 26.5.6 MOAtox 26.5.7 Toxicity/Residue Database 26.6 Overview and Conclusion Acknowledgments Conflicts of Interest References 27. Computational Tools for the Assessment and Substitution of Biocidal Active Substances of Ecotoxicological Concern: The LIFE-COMBASE Project 27.1 Introduction 27.1.1 Biocides Regulation 27.1.2 Alternative Methods 27.1.3 Computational Approaches on Biocides: State of the Art 27.1.4 The LIFE-COMBASE Project 27.2 Database Compilation 27.2.1 Criteria Definition for the Selection of Biocidal Active Substances 27.2.2 Sources of Data 27.3 Development of the QSAR Models 27.3.1 Preparation of the Data Sets 27.3.2 QSAR Models for Microorganisms 27.3.3 QSAR Models for Algae 27.3.4 QSAR Models for Daphnia magna 27.3.5 QSAR Models on Fish 27.4 Prediction of Metabolites and their Associated Toxicity 27.5 Implementation of the In Silico QSARs Within VEGA and Integration with Read Across Models in ToxRead 27.5.1 Implementation of the QSAR Models Within VEGA 27.5.2 Implementation of the Rules for Read-Across and Grouping Within ToxRead 27.5.3 Integration of QSARs and Read-Across Within a Weight-of-evidence Strategy 27.6 Implementation of the LIFE-COMBASE Decision Support System 27.6.1 Database Search Engine 27.6.2 Biocides’ Chemical Space 27.6.3 Metabolites Prediction 27.6.4 Calculation of Aquatic Ecotoxicity 27.6.5 Generation of Alternative Biocide Structures 27.7 Implementation of the LIFE-COMBASE Mobile App 27.8 Concluding Remarks Acknowledgments References 28. Image Analysis and Deep Learning Web Services for Nano informatics 28.1 Introduction 28.2 NanoXtract 28.2.1 NanoXtract Environment and Image Uploading 28.2.2 Computational Workflow and Available Settings 28.2.3 Produced Results 28.3 DeepDaph 28.3.1 DeepDaph Environment 28.3.2 Produced Results 28.4 Conclusions Acknowledgments References Index