Recent Advances in QSAR Studies (eBook)

Preface 6
Contents 7
Part I Theory of QSAR 15
1 Quantitative Structure--Activity Relationships (QSARs) -- Applications and Methodology 16
1.1 Introduction 16
1.2 Purpose of QSAR 17
1.3 Applications of QSAR 17
1.4 Methods 18
1.5 The Cornerstones of Successful Predictive Models 20
1.6 A Validated (Q)SAR or a Valid Prediction? 22
1.7 Using in Silico Techniques 22
1.8 New Areas for in Silico Models 24
1.9 Conclusions 24
References 24
2 The Use of Quantum Mechanics Derived Descriptors in Computational Toxicology 25
2.1 Introduction 25
2.2 The Schrdinger Equation 27
2.3 HartreeFock Theory 29
2.4 Semi-Empirical Methods: AM1 and RM1 30
2.5 AB Initio: Density Functional Theory 31
2.6 QSAR for Non-Reactive Mechanisms of Acute (Aquatic) Toxicity 31
2.7 QSAR s for Reactive Toxicity Mechanisms 33
2.7.1.Aquatic Toxicity and Skin Sensitisation 33
2.7.2.QSARs for Mutagenicity 36
2.8 Future Directions and Outlook 37
2.9 Conclusions 38
Acknowledgement 38
References 38
3 Molecular Descriptors 41
3.1 Introduction 41
3.1.1.Definitions 41
3.1.2.History 43
3.1.3.Theoretical vs. Experimental Descriptors 45
3.2 Molecular Representation 47
3.3 Topological Indexes 50
3.3.1.Molecular Graphs 50
3.3.2.Definition and Calculation of Topological Indexes (TIs) 51
3.3.3.Graph-Theoretical Matrixes 54
3.3.4.Connectivity Indexes 60
3.3.5.Characteristic Polynomial 62
3.3.6.Spectral Indexes 65
3.4 Autocorrelation Descriptors 67
3.4.1.Introduction 67
3.4.2.Moreau--Broto Autocorrelation Descriptors 69
3.4.3.Moran and Geary Coefficients 71
3.4.4.Auto-cross-covariance Transforms 72
3.4.5.Autocorrelation of Molecular Surface Properties 75
3.4.6.Atom Pairs 75
3.4.7.Estrada Generalized Topological Index 78
3.5 Geometrical Descriptors 80
3.5.1.Introduction 80
3.5.2.Indexes from the Geometry Matrix 81
3.5.3.WHIM Descriptors 89
3.5.4.GETAWAY Descriptors 93
3.5.5.Molecular Transforms 102
3.6 Conclusions 105
References 106
4 3D-QSAR -- Applications, Recent Advances, and Limitations 115
4.1 Introduction 115
4.2 Why is 3D-QSAR so Attractive? 116
4.3 Ligand Alignment 117
4.4 CMFA and Related Methods 119
4.4.1.CoMFA 119
4.4.2.CoMSIA 120
4.4.3.GRID/GOLPE 121
4.4.4.4D-QSAR and 5D-QSAR 121
4.4.5.AFMoC 121
4.5 Reliability of 3D-QSAR Models 122
4.6 Receptor-Based 3D-QSAR 124
4.7 Conclusion 131
References 131
5 Virtual Screening and Molecular Design Based on Hierarchical QSAR Technology 138
5.1 Introduction 139
5.2 Multi-Hierarchical Strategy of QSAR Investigation 141
5.2.1.HiT QSAR Concept 141
5.2.2.Hierarchy of Molecular Models 143
5.3.2.1 Simplex Representation of Molecular Structure (SiRMS) 143
5.3.2.2 Lattice Model 148
5.3.2.3 Whole-Molecule Descriptors and Fourier Transform of Local Parameters 150
5.2.3.Hierarchy of Statistical Methods 151
5.3.3.1 Classification Trees 151
5.3.3.2 Trend-Vector 152
5.3.3.3 Multiple Linear Regression 153
5.3.3.4 Partial Least Squares or Projection to Latent Structures (PLS) 154
5.2.4.Data Cleaning and Mining 154
5.3.4.1 Automatic Variable Selection (AVS) Strategy in PLS 155
5.3.4.2 Genetic Algorithms 155
5.3.4.3 Enumerative Techniques 155
5.2.5.Validation of QSAR Models 156
5.2.6.Hierarchy of Aims of QSAR Investigation 158
5.3.6.1 Virtual Screening (Including Consensus Modeling and DA) 159
5.3.6.2 Inverse Task Solution and Interpretation of QSAR Models 162
5.3.6.3 Molecular Design 163
5.2.7.HiT QSAR Software 164
5.3 Comparative Analysis of HT QSAR Efficiency 165
5.3.1.Angiotensin Converting Enzyme (ACE) Inhibitors 166
5.3.2.Acetylcholinesterase (AChE) Inhibitors 167
5.4 HT QSAR Applications 169
5.4.1.Antiviral Activity 169
5.5.1.1 Antiherpetic Activity of N , N 0-(bis-5-nitropyrimidyl)Dispirotripiperazine Derivatives The authors express sincere gratitude to Dr. M. Schmidtke, Prof. P. Wutzler, Dr. V. Makarov, Dr. O. Riabova, Mr. N. Kovdienko and Mr. A. Hromov for fruitful cooperation that made the development of this task possible. (2D) 169
5.5.1.2 Antiherpetic Activity of Macrocyclic Pyridinophanes Anti-influenza and antiherpetic investigations described below were carried out as a result of fruitful cooperation with Dr. V.P. Lozitsky, Dr. R.N. Lozytska, Dr. A.S. Fedtchouk, Dr. T.L. Gridina, Dr. S. Basok, Dr. D. Chikhichin, Mr. V. Chelombitko and Dr. J.-J. Vanden Eynde. The authors express sincere gratitude for all mentioned above colleagues. 170
5.5.1.3 [(Biphenyloxy)propyl]isoxazole Derivatives 0 Human Rhinovirus 2 Replication Inhibitors The authors express sincere gratitude to Dr. M. Schmidtke, Prof. P. Wutzler, Dr. V. Makarov, Dr. O. Riabova and Ms. Volineckaya for fruitful cooperation that made possible the development of this task. (2D) 172
5.5.1.4 Anti-influenza Activity of Macrocyclic Pyridinophanes 4 (2D--4D) 173
5.4.2.Anticancer Activity of MacroCyclic Schiff Bases 8 The authors express sincere gratitude to Dr. V.P. Lozitsky, Dr. R.N. Lozytska and Dr. A.S. Fedtchouk for fruitful cooperation during the development of this task. (2D and 4D) 175
5.4.3.Acute Toxicity of Nitroaromatics 176
5.5.3.1 Toxicity to Rats The authors express sincere gratitude to Prof. J. Leszczynski, Dr. L. Gorb and Dr. M. Quasim for fruitful cooperation during the development of this task. (1D--2D) 176
5.5.3.2 Toxicity to Tetrahymena Pyriformis The authors express sincere gratitude to Prof. J. Leszczynski, Dr. L. Gorb, Dr. M. Quasim and Prof. A. Tropsha for fruitful cooperation during the development of this task. (2D) 177
5.4.4.AChE Inhibition The authors express sincere gratitude to Prof. J. Leszczynski, Dr. L. Gorb and Dr. J. Wang for fruitful cooperation during the development of this task. (2.5D, Double 2.5D, and 3D) 178
5.4.5.5-HT 1A Affinity (1D04D) The authors express sincere gratitude to Academician S.A. Andronati and Dr. S.Yu. Makan for fruitful cooperation during the development of this task. 179
5.4.6.Pharmacokinetic Properties of Substituted Benzodiazepines (2D) 180
5.4.7.Catalytic Activity of Crown Ethers The authors express sincere gratitude to Prof. G.L. Kamalov, Dr. S.A. Kotlyar and Dr. G.N. Chuprin for fruitful cooperation during the development of this task. (3D) 181
5.4.8.Aqueous Solubility The authors express sincere gratitude to Prof. J. Leszczynski, Dr. L. Gorb and Dr. M. Quasim for fruitful cooperation during the development of this task. (2D) 182
5.5 Conclusions 183
References 183
6 Robust Methods in QSAR 188
6.1 Introduction 188
6.2 Outliers and their genesis in the QSAR Studies 190
6.3 Major Concepts of Robustness 192
6.3.1.The Breakdown Point of an Estimator 192
6.3.2.Influence Function of an Estimator 192
6.3.3.Efficiency of an Estimator 192
6.3.4.Equivariance Properties of an Estimator 193
6.4 Robust Estimators 194
6.4.1.Robust Estimators of Data Location and Scatter 194
6.4.2.Robust Estimators for Multivariate Data Location and Covariance 196
6.5 Exploring the Space of Molecular Descriptors 198
6.5.1.Classic Principal Component Analysis 198
6.5.2.Robust Variants of Principal Component Analysis 199
6.6.2.1 Spherical and Elliptical PCA 200
6.6.2.2 Projection Pursuit with the Qn Scale 202
6.6.2.3 ROBPCA -- A Robust Variant of PCA 202
6.6 Construction of Multivariate QSAR Models 203
6.6.1.Classic Partial Least Squares Regression 203
6.6.2.Robust Variants of the Partial Least Squares Regression 204
6.7.2.1 Partial Robust M-Regression 204
6.7.2.2 Robust Version of PLS via the Spatial Sign Preprocessing 206
6.7.2.3 RSIMPLS and RSIMCD -- Robust Variants of SIMPLS 207
6.6.3.Outlier Diagnostics Using Robust Approaches 207
6.7 Examples of Applications 209
6.7.1.Description of the Data Sets Used to Illustrate Performance of Robust Methods 209
6.7.2.Identification of Outlying Molecules Using the Robust PCA Model 210
6.7.3.Construction of the Robust QSAR Model with the PRM Approach 213
6.8 Concluding Remarks and Further Readings 216
References 216
7 Chemical Category Formation and Read-Across for the Prediction of Toxicity 220
7.1 Introduction 220
7.2 Benefits of the Category Formation 221
7.3 Chemical Similarity 221
7.4 General Approach to Chemical Category Formation 222
7.5 Examples of Category Formation and Read-Across 223
7.5.1.Chemical Class-Based Categories 224
7.5.2.Mechanism-Based Categories 224
7.5.3.Chemoinformatics-Based Categories 227
7.6 Conclusions 228
Acknowledgement 228
References 228
Part II Practical Application 231
8 QSAR in Chromatography: Quantitative Structure--Retention Relationships (QSRRs) 232
8.1 Introduction 232
8.1.1.Methodology of QSRR Studies 232
8.1.2.Intermolecular Interactions and Structural Descriptors of Analytes 235
8.2 Chromatographic Retention Predictions 239
8.2.1.Retention Predictions in View of Optimization of HPLC Separations 240
8.2.2.Retention Predictions in Proteomics Research 248
8.3 Characterization of Stationary Phases 249
8.4 Assessment of Lipophilicity by QSRR 252
8.5 QSRR in Affinity Chromatography 257
8.6 Conclusions 259
Acknowledgement 260
References 260
9 The Use of QSAR and Computational Methods in Drug Design 269
9.1 Introduction 269
9.2 From New Chemical Entities (Nce) To Drug Candidates: Preclinical Phases 270
9.2.1.Stage 1: Hit Finding 270
9.2.2.Stage 2: Lead Finding 271
9.2.3.Stage 3: Lead Optimization 272
9.3 Failure in Drug Candidates Development 272
9.4 Classic Qsar in Drug Design 273
9.4.1.Hansch Analysis 273
9.4.2.Non-parametric Methods: Free-Wilson and Fujita-Ban 274
9.4.3.Linear Solvation Free-Energy Relationships (LSERs) 274
9.5 QSAR Methods in Modern Drug Design 276
9.5.1.Tools for QSAR 277
9.6.1.1 Data and Databases 277
9.6.1.2 Novel Molecular Descriptors 278
9.6.1.3 3D-QSAR 279
9.6.1.4 Applicability Domain in QSARs 281
9.6 QSAR in Modern Drug Design: Examples 282
9.6.1.Example 1: Application of QSAR to Predict hERG Inhibition 282
9.6.2.Example 2: Application of QSAR to Predict Blood--Brain Barrier Permeation 283
9.6.3.Example 3: Application of QSAR to Predict COX-2 Inhibition 284
9.7 Conclusion and Perspectives 286
References 286
10 In Silico Approaches for Predicting ADME Properties 291
10.1 Introduction 291
10.1.1.Overview of Key ADME Properties 292
10.1.2.Data for Generation of in Silico Models 297
10.2 Models for the Prediction of ADME Properties 299
10.3 Software Developments 302
10.4 Selecting the Most Appropriate Modeling Approach 305
10.5 Future Direction 306
10.6 Conclusion 309
Acknowledgement 309
References 309
11 Prediction of Harmful Human Health Effects of Chemicals from Structure 313
11.1 Introduction 313
11.1.1.Prediction of Harmful Effects to Man? 314
11.1.2.Relevant Toxicity Endpoints Where QSAR Can Make a Significant Contribution 315
11.2 In Silico Tools for Toxicity Prediction 316
11.2.1.Databases 316
11.2.2.QSARs 319
11.3.2.1 Skin Sensitization Data for Modeling 319
11.3.2.2 SAR (Qualitative) Models for Skin Sensitization 319
11.3.2.3 QSAR Models for Skin Sensitization 320
11.3.2.4 General Comments of the Use of QSAR Models for Predicting Human Health Effects 322
11.2.3.Expert Systems 323
11.2.4.Grouping Approaches 326
11.3 The Future of in Silico Toxicity Prediction 328
11.3.1.Consensus (Q)SAR Models 328
11.3.2.Integrated Testing Strategies (ITS) 329
11.4 Conclusions 330
Acknowledgement 330
References 330
12 Chemometric Methods and Theoretical Molecular Descriptors in Predictive QSAR Modeling of the Environmental Behavior of Organic Pollutants 334
12.1 Introduction 334
12.2 A Defined Endpoint (OECD Principle 1) 336
12.3 An Unambiguous Algorithm (OECD Principle 2) 336
12.3.1.Chemometric Methods 337
12.4.1.1 Regression Models 337
12.4.1.2 Classification Models 337
12.3.2.Theoretical Molecular Descriptors 339
12.3.3.Variable Selection and Reduction. The Genetic Algorithm Strategy for Variable Selection 340
12.4 Applicability Domain (OECD Principle 3) 341
12.5 Model Validation for Predictivity (OECD Principle 4) 343
12.5.1.Splitting of the Data Set for the Construction of an External Prediction Set 344
12.5.2.Internal and External Validation 345
12.5.3.Validation of Classification Models 346
12.6 Molecular Descriptor Interpretation, If Possible (OECD Principle 5) 347
12.7 Environmental Single Endpoints 347
12.7.1.Physico-chemical Properties 347
12.8.1.1 Soil Sorption of Pesticides 348
12.7.2.Tropospheric Reactivity of Volatile Organic Compounds with Oxidants 350
12.7.3.Biological Endpoints 352
12.8.3.1 Bioconcentration Factor 352
12.8.3.2 Toxicity 353
12.8 Modeling More than a Single Endpoint 357
12.8.1.PC Scores as New Endpoints: Ranking Indexes 357
12.8.2.Multivariate Explorative Methods 357
12.9.2.1 Principal Component Analysis 358
12.9.2.2 QSAR Modeling of Ranking Indexes 358
12.9 Conclusions 366
Acknowledgement 366
References 366
13 The Role of QSAR Methodology in the Regulatory Assessment of Chemicals 374
13.1 Introduction 374
13.2 Basic Concepts 375
13.3 The Regulatory Use of (Q)SAR Methods 376
13.4 The Validity, Applicability, and Adequacy of (Q)SARs 377
13.4.1.Demonstrating Validity 378
13.4.2.Demonstrating Applicability 380
13.4.3.Demonstrating Adequacy 381
13.5 The Integrated Use of (Q)SARs 383
13.5.1.Stepwise Approach for Using (Q)SAR Methods 383
13.5.2.Use of (Q)SARs in Chemical Categories 384
13.5.3.Use of (Q)SARs in Integrated Testing Strategies 385
13.6 Conclusions 386
References 386
14 Nanomaterials -- the Next Great Challenge for QSAR Modelers 390
14.1 Increasing Role of Nanomaterials 390
14.2 Their Incredible Physical and Chemical Properties 391
14.3 Nanomaterials can be Toxic 392
14.3.1.Specific Properties Cause Specific Toxicity 393
14.3.2.Oxidative Stress 394
14.3.3.Cytotoxicity and Genotoxicity 394
14.3.4.Neurotoxicity 394
14.3.5.Immunotoxicity 395
14.3.6.Ecotoxicity 395
14.4 NANO-QSAR Advances and Challenges 396
14.4.1.Description of Structure 397
14.4.2.Nanostructure -- Electronic Properties Relationships 402
14.4.3.Nano-QSAR Models 403
14.5 Summary 409
Acknowledgement 410
References 410
Appendix A 417
Index 420