Drug discovery has a long tradition. Originating from folk medicine, for centuries empiricism played an important role in drug discovery. In the second half of the twentieth century rational approaches started to dominate drug research; the increasing understanding of biochemistry and biology supported the discovery of many drugs derived from the chemical structures of endogenous neurotransmitters and steroid hormones.
Nowadays, we interpret the application of computer-assisted methods, in first line molecular modeling and virtual screening, as "rational" drug discovery. However, taking a closer look, we realize that these approaches are also dominated by empiricism. In 1926, Erwin Schroedinger derived analytical solutions for the electronic states of the hydrogen atom. A few years later Paul Dirac (Nobel Prize 1933) formulated what we could also define as the dilemma of computer-aided drug design: "The underlying physical laws necessary for the mathematical theory of a large part of physics and the whole of chemistry are thus completely known, and the difficulty is only that the exact application of these laws leads to equations much too complicated to be soluble" . A statement that is still valid today, despite our powerful computers.
The dilemma was also elaborated by the famous Austrian philosopher Sir Karl Popper in his book "Objective Knowledge; an Evolutionary Approach": [it] "would not be very surprising ... that all the findings of chemistry can be fully explained by (that is to say, deduced from) the principles of physics ...It would not only be an exercise in unification, but a real advance in understanding the world. Let us assume that this reduction has been carried out completely. This might give us some hope that we may also reduce one day all the biological sciences to physics", and later "It is conceivable, although not yet certain, that the reduction of chemistry to physics will be completely successful. It is also conceivable, though less likely, that we may one day have good reductions of biology, including physiology, to physics, and of psychology to physiology, and thus to physics" .
With respect to biochemistry and even biology, a lot of this has been achieved by molecular modeling, at the cost of significant simplification. Force fields consider atoms as hard balls, with certain electronic and steric properties, and bonds and angles as springs with optimal lengths and angles, respectively; also torsional angles, van der Waals and electrostatic interactions are described by corresponding approximations. Nevertheless, it is surprising how well such approximations work, if the methods are applied in a proper manner.
Molecular design: is it possible, is it easy, is it difficult? Well, the novice may start by selecting a certain computer program, feeding it with his or her input, e.g. a protein 3D structure from the Protein Data Bank (PDB), and perform a docking of several potential ligands - he or she will necessarily fail. A PDB structure has to be processed in many different aspects, i.e. hydrogen addition, orientation of asparagine, glutamine, threonine, and histidine side chains, inspection (and correction) of the hydrogen bonding network, defining protonation or deprotonation states (cave pH shifts), and orienting the hydroxy groups of serine, threonine and tyrosine. In the same manner, potential ligands have to be processed. Here additional problems are the most often unknown pKa values of heterocyclic bases, the selection of the right enantiomer of a racemate, the correct or "induced" tautomers of some ligands, etc.; several low-energy conformations have to be generated, in advance or on the fly, and investigated by the docking program. Only some docking programs consider the binding site flexibility in an appropriate manner. All these actions need the skill of an experienced modeler.
The book by Gisbert Schneider and Karl-Heinz Baringhaus offers in its first two chapters a basic introduction to molecular modeling. Molecular objects and their proper presentation, drug-likeness, similarity, 3D-structure generation, receptor-ligand interactions, their thermodynamics and interaction types, QSAR, pharmacophore concepts, and docking and scoring are explained to pave the ground for the following chapters. Chapter 3 deals with computer-assisted molecular design, examples of drug targets and their ligands, binding site characterization, ligand-based design, binding modes, transition state inhibitors, de novo design, and ligand assembly, as well as multidimensional navigation in chemical space. Chapter 4 explains and compares virtual screening vs. (wet) high-throughput screening; various issues from hit to lead propagation are discussed, among them chemical diversity, filtering, shape-based similarity, and scaffold hopping. As molecular design in drug discovery is just ligand design, Chapter 5 illustrates some constraints, i.e. absorption, pharmacokinetics (distribution, metabolism and elimination), and toxicity. In addition, prodrugs and bioisosters are discussed there. The book concludes with machine learning for lead finding and optimization; PCA and PLS, SOMs (Kohonen maps), Bayesian classifiers, neural networks, and support vector machines are presented in their application for library design.
Let me come back to my favorite philosopher. Already in 1919 Karl Popper speculated about good and poor science and realized that science advances only by deductive falsification. He attempted to "distinguish between science and pseudoscience; knowing very well that science often errs, and that pseudo-science may happen to stumble on the truth". In doing this, he formulated (abstracted from the original text ):
it is easy to obtain confirmations ... if one looks for confirmations;
confirmations should count only if they are the result of risky predictions;
every "good" scientific theory is a prohibition;
a theory which is not refutable ... is non-scientific;
every genuine test of a theory is an attempt to falsify it;
confirming evidence should not count except when it is the result of the test of the theory;
some genuinely testable theories, when found to be false, are still upheld by their admirers - for example by introducing ad hoc some auxiliary assumption, or by reinterpreting the theory ad hoc in such a way that it escapes refutation.
This defines also the right spirit to perform molecular design. Especially the last comment on the temptation to adhere to false hypotheses should be a strong warning to all modelers. While there is a saying that a good physical theory should be beautiful (consider for example the Schrödinger equation or the Einstein energy-mass relationship), this does not apply to molecular modeling! Modeling results are most often beautiful, in their numbers and colors, but they can nevertheless be absolutely meaningless in their physical, chemical and biological relevance. Especially in this respect the book by Schneider and Baringhaus is an excellent guide for hitchhikers to the modeling universe. Correspondingly, it is hoped that this "Molecular Design" cookbook will get a broad distribution and that readers will follow the many recommendations from two highly experienced experts, with their strong industrial and academic background in chemoinformatics, modeling and drug design. Many representative examples demonstrate the proper application of all different modeling approaches, in their scope and limitations. Thus, this monograph is considered to be of the utmost importance, not only for the beginner and the experienced modeler, but also for all interested medicinal chemists, biochemists and biologists.
Weisenheim am Sand Hugo Kubinyi
Molecular Design: Concepts and Applications. Gisbert Schneider and Karl-Heinz Baringhaus Copyright 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31432-4
P. A. M. Dirac, Quantum mechanics of many-electron systems. Proc. R. Soc. London, Ser. A 1929. 123, 714-733.
K. Popper, Objective Knowledge; an Evolutionary Approach. Clarendon Press, Oxford, 1972.
K. Popper, Conjectures and Refutations. Routledge and Reagan Paul, London, 1963, pp. 33-39.
"History suggests drug discovery is art as well as science." 
In the course of evolution, several "designs" have emerged as preferred templates for a certain function. Just as the development of wings proved to be suited for flying, proteins and their ligands have evolved to perform particular and specific biological functions. Molecular design is grounded on an understanding of these underlying principles and using this knowledge for the creation of new molecular architecture that performs a desired function or possesses a certain property.
The development of wings during the evolution of the species represents a very successful design. The photograph shows a 49 million years old Buprestid beetle found in Grube Messet near Frankfurt, Germany. © Christa Behnke, Hessisches Landesmuseum Darmstadt. Reproduced with kind permission.
There are several definitions of the meaning of the term "design". The ones we found particularly appropriate in the context of this book "... the act of working out the form"  and "... the overall plan or strategy by which hypotheses or research questions are answered" .
Certainly, molecular design is hypothesis-driven discovery. We try to follow this concept throughout the five chapters of this book. First, we define the design objects by asking the question "What is a molecule?" Only with a profound understanding of the goals of design can the design process be successful. Chapter 2 Molecular Design: Concepts and Applications. Gisbert Schneider and Karl-Heinz Baringhaus Copyright 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31432-4
treats ligand-receptor interactions. We introduce basic concepts such as QSAR and pharmacophore modeling. Throughout, possibilities and limitations of the presented approaches are highlighted. In Chapter 3 techniques for the actual molecular design step are presented on a more conceptual level, while in Chapter 4 methodological aspects are stressed and described in greater detail. The final chapter presents modeling approaches toward ADM ET properties and lead optimization with an emphasis on machine learning methods. The use of selected case studies throughout the text helps to put the contents into a medicinal chemistry context.
The content of this book was carefully compiled over a period of almost two years and is aimed at a wide audience. We hope it will be suitable for anyone interested in molecular design, in particular advanced and graduate students of chemistry and the life sciences. Moreover, trained medicinal chemists and bioinformaticians will certainly find something useful in it, as also will novices to the field.
We would like to express our great gratitude to everyone who helped us put together the contents of this book. Most of all we thank present and previous coworkers and colleagues for constructive criticism, inspiration, continuing support and friendship, in particular (in alphabetical order): Alexander Alanine, David Banner, Jürgen Bajorath, Konrad Bleicher, Hans-Joachim Böhm, Georg Büldt, Tim Clark, Gabriele Cruciani, Swetlana Derksen, Norbert Dichter, Joachim Engels, Uli Fechner, Lutz Franke, Johann "Johnny" Gasteiger, Alireza Givehchi, Klaus Gubernator, Michael Göbel, Holger Gohlke, Eva-Maria Gutknecht, Tina Grabowski, Gerhard Hessler, Jan Hiß, Manfred Kansy, Carsten Kettner, Thomas Klabunde, Herbert Koppen, Gerhard Klebe, Björn Krüger, Paul Labute, Michael Meissner, Man-Ling Lee, Hans Matter, Lutz Müller-Kuhrt, Thorsten Naumann, Christopher Parsons, Ewgenji Proschak, Steffen Renner, Olivier Roche, Oliver Schwarz, Mark Rogers-Evans, Matthias Rupp, Brigitte Scheidemantel-Geiss, Michael Schmuker, Manfred Schubert-Zsilavecz, Andreas Schüller, Martin Stahl, Holger Stark, Christoph Steinbeck, Dieter Steinhilber, Yusuf Tanrikulu, Andreas Teckentrup, Lutz Weber, Tanja Weil, Martin Weisel, and Paul Wrede. They all contributed in one way or another, directly or indirectly, to this text.
Hugo Kubinyi was an infinite source of inspiration, and we are very happy for his willingness to contribute the Foreword.
GS gives special thanks to his wife Petra Schneider, who contributed many excellent ideas, several illustrations, and endured many weekends and long nights while he was writing and moaning.
Finally, we would like to thank the publisher, Wiley-VCH, for giving us the unique opportunity to compile this book. Frank Weinreich, Gudrun Walter, and Waltraud Wüst did a great editing job, were very patient with us, and supported us in many ways.
Molecular design is a transdisciplinary field of research. This book contains many aspects of a rapidly evolving research area, and any attempt at writing an all-embracing volume is bound to fail. Our intention is to provide an entry
point to a fascinating area of research, and hope that our view of the field finds acceptance and inspires newcomers to join us in the quest for novel molecules. Forward ho!
Frankfurt am Main Gisbert Schneider
October 2007 Karl-Heinz Baringhaus
- S. Borman, Improving efficiency, Chem. Eng. News 2006, 84, 56-78.
Word-Net definition, Princeton University, 2006.
A. E. Fortune, W. J. Reid, Research in Social Work, 3rd edn. Columbia University Press, New York, 1999.