Advanced Computer-Assisted Techniques in Drug Discovery edited by Han van de Waterbeemd Methods and Principles in Medicinal Chemistry Edited by R. Mannhold P. Krogsgaard-Larsen H. Timmerman Volume 1 Hugo Kubinyi, QSAR: Hansch Analysis and Related Approaches Volume 2 Han van de Waterbeemd (ed.), Chemometric Methods in Molecular Design Volume 3 Han van de Waterbeemd (ed.), Advanced Computer- Assisted Techniques in Drug Discovery Methods and Principles in Medicinal Chemistry edited by R. Mannhold, P. Krogsgaard-Larsen, H. Timmerman This practice-oriented series of handbooks and monographs introduces the reader to basic principles and state-of-the-art methods in medicinal chemibtry. Topics treated in-depth include W chemical propertiesofdrugs W characterization of biological activity W advanced techniques in QSAR W physiological and biochemical understanding of diseases Volume I In prepamtion: Kubinyi,H. H.-D. Holtje, G. Folkers QSAR: Hansch Analysis and Molecular Modeling and Related Approaches Drug Design An Introductory Handbook 1993. XI], 240 pages with 60 figures and 32 tables. Hardcover. - Winter l995/6 - DM 164.00. ISBN 3-527-30035-X (VCH, Weinheim) V. Pliska, B. Testa, H. van de Waterbeemd (eds.) Lipophilicity in Drug Research Volume 2 and Toxicology van de Waterbeemd, H. (ed.) -Winter 1995/6 - Chemometric Methods in Molecular Design 1995. Ca 300 pages. Hardcover. Ca DM 178.00. 4b ISBN 3-527-30044-9 (VCH, Weinheim) VCH Advanced Computer- Assisted Techniques in Drug Discovery edited by Han van de Waterbeemd Weinheim New York Base1 Cambridge Tokyo Volume editor: Prof. Povl Krogsgaard-Larsen Dr. Han van de Waterbeemd Dept. of Organic Chemistry F. Hoffmann - La Roche Ltd. Royal Danish School of Pharmacy Pharma Research New Technologies DK-2100 Copenhagen CH-4002 Basel Denmark Switzerland Editors: Prof. Raimund Mannhold Prof. Hendrik Timmerman Biomedical Research Center Faculty of Chemistry Molecular Drug Research Group Dept. of Pharmacochemistry Heinrich- Heine-Universitat Free University of Amsterdam UniversitatsstraRe 1 De Boelelaan 1083 D-40225 Diisseldorf NL-1081 HV Amsterdam Germany The Netherlands I I This book was carefully produced. Nevertheless, authors, editors and publisher do not warrant the information contained therein to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Published jointly by VCH Verlagsgesellschaft mbH, Weinheim (Federal Republic of Germany) VCH Publishers, Inc., New York, NY (USA) Editorial Director: Dr. Thomas Mager Production Manager: Dip1.-Ing. (FH) Hans Jorg Maier Library of Congress Card No. applied for. British Library Cataloguing-in-Publication Data: A catalogue record for this book is available from the British Library. Deutsche Bibliothek Cataloguing-in-Publication Data: Advanced computer assisted techniques in drug discovery I ed. by Han van de Waterbeemd. - Weinheim ; New York ; Basel ; Cambridge ; Tokyo : VCH, 1994 (Methods and principles in medicinal chemistry ;V ol. 3) ISBN 3-527-29248-9 NE: Waterbeemd, Han van de [Hrsg.]; GT 0 VCH Verlagsgesellschaft mbH. D-6945 I Weinheim (Federal Republic of Germany), 1995 Printed on acid-free and chlorine-free paper. All rights reserved (including those of translation in other languages). No part of this book may be reproduced in any form - by photoprinting, microfilm, or any other means - nor transmitted or translated into machine language without written permission from the publishers. Registered na- mes, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Composition:K+V Fotosatz GmbH, D- 64743 Beerfelden. Printing: betz-druck gmbh, D-64291 Darmstadt. Printed in the Federal Republic of Germany. Distribution: VCH, P.O. Box 10 11 61, D-69451 Weinheim (Federal Republic of Germany) Switzerland: VCH, P.O. Box, CH-4020 Basel (Switzerland) United Kingdom and Ireland: VCH (UK) Ltd., 8 Wellington Court, Cambridge CBI 1HZ (England) USA and Canada: VCH, 220 East 23rd Street, New York, NY 10010-4606 (USA) Japan: VCH, Eikow Building, 10-9 Hongo 1-chome, Bunkyo-ku, Tokyo 113 (Japan) Preface The main objective of this series is to offer a practice-oriented survey of techniques currently used in Medicinal Chemistry. Following the volumes on Hansch analysis and related approaches (Vol. 1) and multivariate analyses (Vol. 2), the present hand- book focuses on some new, emerging techniques in drug discovery; emphasis is plat- ed on showing users how to apply these methods and to avoid time-consuming and costly errors. Four major topics are covered. The first centers on three-dimensional QSAR, and some of the enormous progress achieved in this field is summarized. Both the various 3D-QSAR methods available as well as the chemometric tools for handling the statis- tical problems involved in 3D-QSAR studies are covered. Intimately coupled with 3D-QSAR is the current trend in pharmaceutical industry to establish chemical structure databases as a tool for identifying new leads. Cor- respondingly, in the second section, problems encountered in our understanding of molecular similarity and aspects of compound selection by clustering databases are treated. The third section covers advanced statistical techniques in drug discovery. Inter alia the approach of Svante Wold to apply PLS to non-linear structure-activity rela- tions deserves to be mentioned here. Last but not least, the use of neural networks for data analysis in QSAR problems is discussed. Advantages and disadvantages are critically analysed by comparing net- works versus statistics. The editors would like to thank all contributors and VCH publishers for their fruitful cooperation. Summer 1994 Diisseldorf Raimund Mannhold Kopenhagen Povl Krogsgaard-Larsen Amsterdam Hendrik Timmerman A Personal Foreword It is no coincidence that the first three volumes of Methods and Principles in Medicinal Chemistry deal with computer-assisted medicinal chemistry. After the classical Hansch method in Volume 1 and applications of chemometric methods in Volume 2, the present volume of the series contains a number of emerging new tech- niques. Of course, all approaches using molecular modeling techniques, such as structure-based design and de novo design, rely on computers as well. These will be treated separately in a forthcoming volume. This volume is a logical continuation of Volume 2. In fact, after analyzing the methods that have been developed following the Hansch method, we came to the conclusion that a number of these techniques have now matured, while others still require further developments. This criterion was used to select the chapters for Volumes 2 and 3. In reviewing the contents of the first three volumes in this series, it is evident that highly specialized tools have become available for the analysis of complex biological and chemical data sets in order to unravel quantitative structure-activity relation- ships. It has not become easier for the bench chemist to select the ideal method for dealing with the analysis of structure-activity relationships using chemical and bio- logical data. Specialist support is required to validate and apply statistical or chemo- metric and other computer-assisted tools. Volume 3 focusses very much on the newest methods employed by the chemometrician. We hope that, in an indirect way, some of the methods discussed will be of use to molecular design on a day to day basis. I am grateful to, and would like to thank all the contributing authors for their efforts in compiling this volume. February 1994, Base1 Han van de Waterbeemd Contents Preface ........................................................... V A Personal Foreword ............................................... VI 1 Introduction ............................................... 1 H . van de Waierbeemd 1.1 3D QSAR .................................................. 1 1.2 Databases .................................................. 4 1.3 Progress in Multivariate Data Analysis ......................... 4 1.4 Scope of this Book .......................................... 5 References ......................................................... 6 2 3D QSAR: The Integration of QSAR with Molecular Modeling .................................................. 9 2.1 Chemometrics and Molecular Modeling ........................ 9 D. Piiea. c! Cosenfino. G. Moro. L . Bonaii. E . Fraschini. M . Lasagni and R . Todeschini 2.1 .1 Introduction ................................................ 10 2.1.2 QSAR Methodology using Molecular Modeling and Chemometrics 11 2.1.2.1 Search for the Geometric Pharmacophore ...................... 13 2.1.2.2 Quantitative Correlation between Molecular Properties and Activity ................................................ 16 2.1.2.3 Computer Programs ......................................... 18 2.1.3 Illustrative Examples ......................................... 18 2.1.3.1 Amnesia-Reversal Compounds ................................. 18 2.1.3.2 Non-Peptide Angiotensin I1 Receptor Antagonists ............... 21 2.1.3.3 HMG-CoA Reductase Inhibitors ............................... 25 2.1.3.4 Antagonists at the 5-HT3 Receptor ............................. 28 2.1.3.5 Polychlorinated Dibenzo-p-dioxins ............................. 32 2.1.4 Conclusions ................................................. 35 References ......................................................... 36 X Contents 2.2 3D QSAR Methods .......................................... 39 A.M. Davis 2.2.1 Introduction ................................................ 39 2.2.2 3D QSAR of a Series of Calcium Channel Agonists ............. 41 2.2.2.1 Molecular Alignment ......................................... 43 2.2.2.2 Charges .................................................... 45 2.2.2.3 Generating 3D Fields ......................................... 45 2.2.2.4 Compilation of GRID Maps .................................. 47 2.2.2.5 Inclusion of Macroscopic Descriptors with 3D Field Data ........ 48 2.2.3 Statistical Analysis ........................................... 49 2.2.3.1 Results of the Analysis ....................................... 51 2.2.3.2 Testing the Model ........................................... 56 2.2.4 Conclusions ................................................. 57 References ......................................................... 59 2.3 GOLPE Philosophy and Applications in 3D QSAR ............. 61 G. Cruciani and S . Clementi 2.3.1 Introduction ................................................ 61 2.3.1.1 3D Molecular Descriptors and Chemometric Tools ............... 63 2.3.1.2 Unfolding Three-way Matrices ................................. 64 2.3.2 The GOLPE Philosophy ...................................... 65 2.3.2.1 Variable Selection ............................................ 68 2.3.3 Applications ................................................ 70 2.3.3.1 PCA on the Target Matrix .................................... 71 2.3.3.2 PCA on the Probe Matrix .................................... 73 2.3.3.3 PLS Analysis on the Target Matrix ............................ 76 2.3.3.4 PLS on Target Matrix as a Strategy to Ascertain the Active Conformation .......................... 78 2.3.3.5 GOLPE with Different 3D Descriptors ......................... 81 2.3.4 Conclusions and Perspectives .................................. 82 References ......................................................... 87 3 Rational Use of Chemical and Sequence Databases .......... 89 3.1 Molecular Similarity Analysis: Applications in Drug Discovery .... 89 M . A . Johnson. G . M . Muggiora. M . S . Lujiness. J. B. Moon. J D . Petke and D. C. Rohrer 3.1.1 Introduction ................................................ 89 3.1.2 Similarity-Based Compound Selection .......................... 91 3.1.2.1 Similarity Measures and Neighborhoods ........................ 91 Contents XI 3.1.2.2 Application of 2D and 3D Similarity Measures .................. 94 3.1.2.3 Application of Dissimilarity-Based Compound Selection for Broad Screening ................................................... 95 3.1.3 Structure-Activity Maps (SAMs) ............................... 96 3.1.3.1 A Visual Analogy ........................................... 96 3.1.3.2 Representing Inter-Structure Distances .......................... 97 3.1.3.3 Structure Maps .............................................. 99 3.1.3.4 Coloring a Structure Map .................................... 101 3.1.4 Field-Based Similarity Methods ................................ 102 3.1.4.1 Field-Based Similarity Measures ............................... 103 3.1.4.2 Field-Based Molecular Superpositions .......................... 104 3.1.4.3 An Example of Field-Based Fitting: Morphine and Clonidine ..... 105 3.1.5 Conclusions ................................................. 108 References ......................................................... 109 3.2 Clustering of Chemical Structure Databases for Compound Selection ................................................... 111 G.M. Downs and I? Willett 3.2.1 Introduction ................................................ 111 3.2.2 Review of Clustering Methods ................................. 114 3.2.2.1 Hierarchical Clustering Methods ............................... 115 3.2.2.2 Non-Hierarchical Clustering Methods .......................... 118 3.2.3 Choice of Clustering Method ................................. 121 3.2.3.1 Computational Requirements .................................. 121 3.2.3.2 Cluster Shapes .............................................. 122 3.2.3.3 Comparative Studies ......................................... 123 3.2.4 Examples of the Selection of Compounds from Databases by Clustering Techniques ........................................ 125 3.2.4.1 The Jarvis-Patrick Method .................................... 125 3.2.4.2 The Leader Method .......................................... 126 3.2.5 Conclusions ................................................. 127 References ......................................................... 128 3.3 Receptor Mapping and Phylogenetic Clustering .................. 131 I?J Lewi and H . Moereels 3.3.1 G-protein Coupled Receptors .................................. 132 3.3.2 Principal Coordinates Analysis of 71 Receptor Sequences ......... 135 3.3.3 Principle Coordinates Analysis of 26 Receptor Subtypes .......... 144 3.3.4 Phylogenetic Clustering ....................................... 148 3.3.5 Discussion .................................................. 157 References ......................................................... 161