Officers: Yvonne C. Martin (advisor to the chair), James King (treasurer), Han van de Waterbeemd (secretary/editor)

Board: John Block, Sergio Clementi, John Dearden, Bill Dunn, Marvin Charton, Ferenc Darvas, Rainer Franke, Toshio Fujita, Peter Goodford, Phil Magee, Jim McFarland, Oleg Raevsky, Joachim Seydel, Bernard Testa, Milon Tichy

Honorary chair: Corwin Hansch    Past chair: Phil Magee

Editorial

Dear colleagues and friends,

The progress in science is going on, sometimes in big steps, sometimes very smoothly. But it is going on.
In the field of quantitative structure-activity relationships, the years 1962-1964 were extremely fruitful, with the first definition of the additive lipophilicity parameter ?, the advent of Hansch analysis, and the definition of Free-Wilson analysis. These few years constituted the birth of our QSAR discipline. Another such milestone may be seen in the publication of Dick Cramer’s paper on Comparative Molecular Field Analysis, in 1988, which laid the foundation for many related 3D QSAR approaches. Also medicinal chemistry and agrochemistry developed in the past decades. If we go back to the early sixties, only animal experiments and investigations with whole plants, respectively, were used to characterize the biological actions of new compounds. Correspondingly low was the throughput. One chemist could easily keep several biologists very busy. Times have changed. Genomics (and proteomics), the proof of concept of therapeutic approaches by genetically engineered animals, high-throughput screening, combinatorial chemistry, molecular modelling, structure-based drug design, and computer-aided techniques added new dimensions to drug and agrochemical research. Nowadays thousands and ten thousands, even hundred thousands of compounds can be synthesized and tested within very short time. Whereas genomics, molecular modelling and 3D-protein structure-based approaches are rational procedures, combinatorial chemistry and high-throughput screening (seemingly) hunt for mere chance. Naturally, using combinatorial chemistry as a kind of scatter-gun technique cannot be very successful. Too unfavorable is the relationship between the number of potential leads and drug candidates, on the one hand, and inactive, toxic, unselective or otherwise uninteresting compounds, on the other hand. Many companies, especially small venture-capital organisations, have learned or are just going to learn this painful lesson.
What are the consequences? Techniques have to be developed to select ”drug-like” compounds, to look for reasonable diversity of the members of a combinatorial library and to synthesize compounds which have a good chance for oral bioavailability. Many of our colleagues are involved in developing new tools in this direction. Some approaches (e.g., the differentiation between drugs and ”non-drugs” by neural nets) may not fulfill the conditions of a scientifically sound approach; however, in industrial praxis they turn out to be extremely helpful - they just save money.
You may wonder, why I use this long introduction to come to my point. This point is that, like the years before, several colleagues complained at the QSAR Gordon Conference that this is not any longer ”their conference”. In former years they argued that too much time was wasted for the presentation of beautiful modelling pictures. In some respect, they were right because in the early years of modelling, colorful representations of proteins or some (more or less irrelevant) movements of side chains or domains were considered to be a special achievement. Sometimes free-energy perturbation calculations kept supercomputers busy without giving any reasonable results. However, molecular modelling developed further, in the direction of quantitative comparisons of molecules with respect to their biological activities. Isn’t that QSAR, too? Nowadays, the same persons who argued that there was too much modelling, complain that there is too much docking, virtual screening, data mining, and sublibrary selection stuff in the conference. But all this is QSAR, too.
There is another discussion going on: whether we should change the title of our Gordon Conference to ”QSAR and Modelling” or something like that. I was the one who, immediately after being elected as the Chair of this Society, forced the Board and the members who were present at the 1995 QSAR GRC, to change the title of our Society from ”International QSAR Society” to ”The QSAR and Modelling Society”. As you know, this proposal was accepted after some discussion. However, with respect to our Gordon Conference, I warn to follow this procedure. It is extremely difficult to ”get” a Gordon Conference and to keep it running for long time. Several years ago we were under consideration, i.e. our conference was attended by a Senior Scientist, being sent by the GRC Committee, to find out whether it still would make sense to continue with this conference. Also this year’s conference was under such observation (seemingly, we had good luck because a very experienced and able scientist was selected for this purpose). A change in the title would imply that we want to implement a new, different field. But this is not the case. QSAR, modelling, and other computer-aided techniques all have the same goal, i.e., to model the biological activities of small or large, possibly extremely large groups of compounds by appropriate techniques. Whereas regression analyses is suited for some to several hundred compounds, virtual screening and database mining can handle hundreds of thousands of data points. It is also not true that people might be afraid of the ”QSAR” in the title of our conference. First, the title still is ”Quantitative Structure-Activity Relationships” and not QSAR, and secondly, everybody can take a look into the program and learn the content of the meeting from the titles of the different lectures. Another argument is that our conference is also well attended by modellers and that it is always over-booked.
Some of our complaining colleagues consider to organize their own QSAR conference, not being spoiled by non-classical, non-regression procedures. I am happy with this and wish all success. Surely there are many interesting applications and even modifications which could produce a nice programme. But the QSAR Gordon Conference has to be open for new and diverse approaches that fit the needs of current medicinal chemistry, agrochemistry, ecotoxicity and toxicology studies. I am sure that Corwin Hansch, our honorary president, will be the first one to agree to this. He was always at the frontiers of research, where the Gordon Research Conferences have to be.

There is another important topic: we will have votes for the Chair because my time as Chair of The QSAR and Modelling Society ends in autumn 2000. The procedure will start by collecting proposals for candidates from April 1, 2000. And we will have votes for all Board Members (see below). The Secretary and Newsletter Editor, Han van de Waterbeemd, the Advisor to the Chair, Yvonne Martin, and our Treasurer, Jim King, as well as the Members of the Board, and me contributed to increase the number of members of our Society from 346 to 680 (August 1999). In addition, we started many new activities (web site, e-mailing list, funds for students, new local divisions, etc). Now I feel that somebody else should take the duty to be the Chair. My reasons for not being a candidate for a second time are simple: first of all, other colleagues should be involved and second, my retirement would fall into a possible second chairmanship period.

In 2000, we will have elections of the Chair of our Society and of all Board Members. The following time frame applies:

1)  Proposals of Candidates for the Chair and for Board Members, starting April 01, 2000  (deadline May 31, 2000):

Please send your proposals to Hugo Kubinyi, e-mail hugo.kubinyi@basf-ag.de or by fax to +49-621-60 21414. Make sure that the candidate being proposed by you will be ready to accept the position.

2)  Send the Ballots to all Members, in June 2000

3)  Return the Ballots with your selection, immediately, to Jim McFarland (deadline August 15, 2000)

The address of Jim McFarland will be provided together with the ballots.

4)  Annnouncement of the new Chair and the new Board Members at the 13th European Symposium on Quantitative Structure-Activity Relationships, in Düsseldorf, Germany, August 27-September 02, 2000).

The Secretary and Newsletter Editor, the Advisor to the Chair and the Treasurer will be selected by the elected Chair, which seems to be a good tradition of our Society. However, if you are interested to participate in one of these activities, contact me and I will bring up your proposals to the candidates for the Chair.
 

The next Gordon Research Conference on Quantitative Structure.Activity Relationships will (hopefully) take place in Summer 2001, in Tilton, NH, U.S.A.
Kate Holloway (kate_holloway@merck.com) will be the Chair of this conference. John van Drie (john.h.vandrie@am.pnu.com) was elected in summer 1999 to be the next Vice Chair.

We thank all members for their cooperation in bringing their dues up to date.  Please do not wait for a notice of dues for next year.  Please send them to the noted address and, if you desire, include multiple years. Remember also that checks must be written on banks with US offices; otherwise, the cost of collection exceeds the amount of the check.  Cash may be sent at your discretion and risk.

After some complaints by members about lost mails, etc., the Board recommended at the summer 1999 meeting to change the provider. However, as the problems seem to be resolved, we could consider to postpone or even cancel this change. We thank Pierre-Alain Carrupt at the University of Lausanne to set up and maintain this service. All members are encouraged to use this mailbox more often.

Due to sufficient funding by other organisations, the Chair of the 1999 QSAR Gordon Conference did not need our offer for financial funds. Thus we will give a fund to support a scientist or a few students, to attend the Ninth International Workshop on Quantitative Structure Activity Relationships in Environmental Sciences (QSAR in year 2000) in Bourgas, Bulgaria. Details can be provided by the organizer, Prof. Ovanes Mekenyan, mekenyan.ovanes@epamail.epa.gov.
Funds will also be provided to support the attendance of students at the 13th European Symposium on Quantitative Structure-Activity Relationships, in Duesseldorf, August 2000.

At the last meeting of the Society members it was agreed to generate an award to honor young research scientists (up to an age of 40), being active in the field of (quantitative) structure-activity relationship modelling. We are very happy that the founder of our QSAR discipline, Corwin Hansch, agreed to our proposal to call this the Corwin Hansch AWARD. Annually proposals for candidates can be made by all Society members. Sufficient scientific material (reprints of publications, reviews, etc.), an accompanying letter in favor of the candidate and a CV have to be submitted till May 31 of the corresponding year; a scientific committee (to be nominated) will then select the ”best” candidate. The name of the honored scientist will be announced at the biannual QSAR GRCs and the biannual European QSAR Symposia.
The donation associated with the award will be 1’000 US-$, a one-year free membership to our Society and a one-year free personal subscription to the QSAR journal. If the committee comes to the conclusion that no satisfactory proposals have been made, no award will be provided in the corresponding year.

If other members would like to take the responsibility for local distribution, please contact me. It would be of great help to set up a network. Many thanks to Pfizer Central Research, UK for printing of this Newsletter.
 
 

In the future we would like to send the Newsletter only to members who don’t have access to the World-Wide Web. Please notify us if you do not need the printed version. Your positive reply will help us to spend the mailing expenses for better purposes.

Please support our work, especially for the distribution of messages, by regularly updating e-mail addresses. If you detect a wrong address in the members list in our Web page, please inform han_waterbeemd@sandwich.pfizer.com
 
 

All members are invited to contribute our Newsletter and to our web site. This Newsletter shall not be a one-man show, it gains from your experience. Our publishing policy will not allow us to accept scientific contributions which better should be sent to a reviewed journal. However, tips and tricks, key references, conferences, books, shareware, even the announcement of new commercial software, are welcome. We depend on your active participation!
 

Our QSAR and Quantum Chemistry Group of Timisoara, became this year, together wih some colleagues of Bucharest, the romanian branch of your society.  Publications (sept. 98-99) - one paper in QSAR 18(3),253, one in Dyes and Pigments 40,235, one in J.Comput.-Aided Molec. Des. 12,133, one in Rev.Roumaine Chimie 43(3),241, and 9 in local journals. There was a participation at the Academic Days of Timisoara, May 27-28, when Dr. Martin Bohl (Tripos, Munchen) and Dr. Niels-Ake Bergman (Astra-Zeneca, Sweden), received theDoctor Honoris Causa title of the University of Timisoara.
Z. Simon
 
 

 Topics covered on the Symposium included: (1)  High Quality Multiple Regression Analysis (M. Randic & S. C. Basak); (2) Characterization of DNA sequences using structural invariants (M. Randic, A. Nandy, S. C. Basak & M. Vracko);  Topological indices,  hierarchical QSAR, complexity of molecules and overall connectivities for predicting properties (A. T. Balaban; S. Nikolic, N. Trinajstic; D. Bonchev).   Biodescriptors (proteomic data) and hierachical QSAR (F. Witzmann),  Hierachical QSAR, Structure - toxicity/carcinogenity studies; critical evaluation of QSAR (a case study), (S. C. Basak & coworkers;  M. Vracko and coworkers; W.C. Herndon).   D. Bahler reported on symbolic, neural and Bayesian machine learning models to predict toxicity.   Topics were not restricted solely to QSAR and toxicity as shown by a presentation on graph theory of fullerenes (K. Balasubramanian).   Finally G. W. A. Milne,  the editor of J. Chem. Inf. Comp. Sci. (JCICS) commented on "Graph Theory and JCICS" pointing to desirable and undesirable features of such contributions.  This topic was appropriate not only because of the audience but also in view of the fact that the proceedings of this
one day symposium will appear in J. Chem. Inf. Comp. Sci.   The meeting was attended by about two dozen participants,  having 13 presentations from seven countries (Bulgaria, Croatia,  England, India, Romania, Slovenia and United States).   There was neither registration fee nor deadlines for submission of abstract.   NRRI and the University of Minnesota Duluth (UMD),  were gracious host,  thanks to M. Lalich (Director,  NRRI) and V. Magnuson (Vice-Chancellor for Academic Administration, UMD).

Since 1975, this was the 13th biannual Gordon Research Conference on Quantitative Structure-Activity Relationships. About 140 scientists (this is the upper limit for a Gordon Conference), about 70% of them from pharmaceutical industry, attended this meeting. Although the fraction of non-US participants is always below 50%, this time it was below 30%. Due to the schedule (lectures in the morning and in the evening but free time in the afternoon), there was a lot of informal discussion between all participants. Many leading experts attended the meeting; the proportion of ”newcomers” was small, as usual.
All different aspects of molecular property estimations, classical and 3D QSAR, molecular modelling, structure- and computer-aided design and chemoinformatics were addressed in the lectures. There is much progress in alignment, docking, combinatorial library design, in data mining and in drug-like property estimations. On the other hand, precise scoring functions for the evaluation of docking results are still an unsolved problem, worldwide.
As usual, a large number of participants presented posters on diverse topics. A list of the posters is most probably available from the programme chair, Kate Holloway (kate_holloway@merck.com). According with the rules of the Gordon Research Conferences, no detailed report on the content of the lectures can be given here.

Session: Data Mining and Knowledge Base Development (Jon Mason)

Session: QSAR Variable Selection and Improved QSAR Methods (Alex Tropsha, University of North Carolina, Chapel Hill, NC, chair)

Session: Similarity, Diversity, and Library Design (Bob Sheridan, Merck Research Laboratories, Rahway, NJ, chair)

Session: Computing Ligand-Receptor Binding Affinity (Rebecca Wade, EMBL, Heidelberg, chair)

Session: Computer-Aided Molecular Design (Daniel Kleier, DuPont, Newark, DE, chair)

Session: The Nature of Biologically Active Molecules (Christopher Lipinski, Pfizer Central Research, Groton, CT, chair)

Session: New Developments (Ferran Sanz, I.M.I.M., Barcelona, chair)

Session: 3-D Pharmacophore Perception (Yvonne Martin, Abbott Laboratories, Abbott Park, IL)

(reported by Hugo Kubinyi)

It is 25 years ago that connectivity indices and other topological indices emerged as alternative descriptors for structure-property-activity studies.  While well received in some QSAR circles apparently they have not been welcome among some chemometricians.   A recent paper by Lahana and coworkers (Computer Assisted Rational Design of Immunosuppressive Compounds, G. Grassy, B. Calas,  A. Yasri, R. Lahana,* J. Woo, S. Iyer, M. Kaczorek, R. Floc'h and R. Buelow, Nature Biotechnology,  1998,  vol. 16, pp. 748-752) calls for revision of reservations and re-evaluation of such skepticism.   The authors describe a rational design of immunosupressive peptides without relying on information regarding their receptors or mechanism of action.   The design strategy uses a variety of topological and shape descriptors in combination with an analysis of molecular dynamic trajectories for the identification of potential drug candidates.   The researches started from 27 descriptors such as molecular volume,  lipophilicity,  12 topological indices including Kier's kappa indices, Balaban index,  Wiener index,  Randic index (the connectivity index), Kier and Hall valence connectivity indices,  Hall E-state sum,  and a number of indicator variables (number of C atoms,  H atoms,  methyl groups,  amino groups).   Among the 13 selected descriptors four were topological indices.   The strategy is based on the design and screening of virtual combinatorial libraries using rules derived from a comprehensive description of active and inactive molecules in a relevant learning set.   This strategy allows the development of potential candidates without relying on the 3-D structure of the target receptor.   The predictive capabilities of this kind of approach are improved using several filters,  defined by the range of variation of a given descriptor for all known active compounds when compared to the range of variation of the same descriptor for all known inactive compounds.
 A virtual combinatorial library  was generated for the peptides of general form:  RXXXRXXXXY,  with seven variable positions in order to create the combinatorial library.   Use of 35 amino acids (20 natural and 15 non-natural ones) leads to 357 combination (64 billion compounds),  well above computing capacity.   This number was reduced to  67 (almost 280,000 compounds) by taking into account lipophilicity distribution (considered critical  for the activity considered).   The selected amino acids were V, I, T, W, G  and nL (Nor-Leucin) The lead compounds were peptides,  derived from the heavy chain of HLA class I,  that modulate immune responses in vitro and in vivo.   The lead peptide prolonged skin and heart allograft survival in mice. The learning set consisted of 19 peptides based on the above lead peptide,  of which nine prolonged heart allograft survival in mice and ten had no significant effect.   Rapid computation using the 13 selected molecular descriptors reduced the virtual library to 26 peptides that satisfied all the constraints.   More elaborate calculations then allowed a comparison of the conformational space occupied by active and inactive peptides from the original learning set resulting in the selection of five peptides which were synthesized and tested in vivo.
The biological activity of peptide therapy with the rationally designed peptides has shown that four out of the five resulted in significant prolongation of allograft survival.   The molecule predicted to be most potent displayed an immunosuppresive activity approximately 100 times higher than the lead compound.
 This work illustrates a new avenue for using topological indices in rational design of biologically active compounds.   It is worth pointing out (1) that no alternative computational approaches are today practical for combinatorial libraries having in excess of 100,000 compounds;  (2) that topological indices along with other molecular descriptors apparently are highly efficient (in filtering two dozen compounds out of over quarter of million);  and (3) that the resulting active compound was two orders of magnitude more potent than the lead compound.   It may be argued that the paper of Lahana and coworkers represents historic step in QSAR and rational drug design and will be recognized as such.

Even a cursory review of the QSAR literature indicates the importance of the manner in which molecular structure is represented.  A set of molecular structure descriptors is at the heart of any QSAR equation or clustering technique or  database search.  Careful inspection of these descriptors reveals that they tend to fall into three broad classes:
1- counts/frequencies or presence/absence of selected structure features,
2- physical property values, and
3- continuous variables from algorithms which convert molecular structure directly.
Substructure keys and counts are members of the first class; logP and Hammett sigma values illustrate class 2; E-State indices and partial charges fall into class 3.

In recent conferences and papers several investigators have  discussed the shortcomings of classes 1 and 2, especially feature counts.  In this essay we take the position that descriptors such as the E-State indices represent molecular structure well and resolve the difficulties with feature counts and the problems of determination of logP and Hammett sigma for complex structures.

Our view is that neither classes 1 nor 2 adequately represent the content and context inherent in structure.  Chemists are experienced with the wide variation in hydrogen bond donor ability: -OH is much stronger than -NH-.  Further, -NH- is a better hydrogen bond donor with electronegative elements nearby and is less strong with isopropyl groups on alpha carbons.  Aromatic CHs vary significantly depending on the nature of the substituents.  In general, there is continuous variation in both electronic and topological character for a particular atom (or hydride group) depending upon its molecular environment.  Such variation is expressed in the range of the E-State indices.

In recent years we have developed and presented the E-State formalism, which gives expression to the wide range of atomic and molecular properties found in organic molecule.  The E-State index computed for each atom (hydride group) combines both electronic and topological information.  These two molecule attributes, long recognized as important in QSAR, arise from the same aspect of molecules, the electron distribution.  The E-State index for an atom is composed of an intrinsic state plus perturbation by all other atoms in the molecule. The intrinsic state is a fundamentally new contribution to the arena of molecular structure representation.  It combines directly the electronic character with the steric/topological character without resorting to the approximation device of linear combination.  The intrinsic state is the ratio of the valence state electronegativity to the number of bonded skeletal neighbors, that is, the potential for gain/loss of electron density (upon bonding) divided by the number of (bond) avenues over which electron density is redistributed as the result of bonding.  The intrinsic state value also implicitly identifies each atom (hydride group) without need of some classification scheme.

The second key aspect of the E-State is that the index computed for each atom is determined as a perturbation of the intrinsic state value by every other atom in the molecule, perturbation being diminished as the square of the number of bonds separating each pair of atoms.  In this manner the E-State encodes electron accessibility at each atom, an essential aspect for relating structure to property for phenomena arising from non-covalent interactions.  Each atom is represented in its molecular context in a manner which encodes is actual electrotopological character rather than merely by a crude count.

The E-State indices provide the basis for chemically meaningful description of molecular structure, making possible rational discrimination of even closely related structures and meaningful organization of structures in a chemical structure space.  A clear example is given in our recent book (Molecular Structure Description: The Electrotopological State, Academic Press, 1999).  The set of PCBs is represented in a two-dimensional structure space based on atom type E-State indices.  One dimension is the atom type E-State index for chlorine, that is, the sum of the atom E-State values for all chlorines in the molecule: ST(-Cl).  The second atom type index used here is the sum of the atom E-State values for all aromatic CHs in the molecule: ST(..CH..).  When the twelve dichlorobiphenyls are displayed in this two-dimensional space, they are arranged in a chemically meaningful way as shown in the figure.  For example, at the top of the plot, both chlorines are in ortho positions whereas at the bottom they are both in para positions.  In the middle one chlorine is ortho and the other is either meta or para.  On the right of the plot the two chlorines are both on the same ring whereas on the left one is on each ring. For the other isomeric sets of PCBs a similarly meaningful organization is revealed.  As this example shows the chemist can recognize a structurally meaningful organization of the molecules in E-State space. 

Based on this structurally meaningful organization, it is clear that similarity searching can be effectively done in a space defined by the atom type E-State indices.  Structures in the neighborhood of a given structure are structurally similar.  One can pose hypothetical questions of a database so that sets of related structures can be retrieved.

This kind of structure representation, then, provides a strong basis for the wide range of problems encountered in QSAR type analyses.  The computed E-State value is sensitive to electronic and topological environment within the molecule, being interpreted as electron accessibility which forms the basis for modeling properties which arise from non-covalent interactions.  Finally, the E-State indices represent structure in a familiar and chemically intuitive manner.
 
 

Software
 
SciPolymer 3.0

A new software called SciPolymer 3.0 that allows one to build, visualize, and predict the properties of polymers. The basis of the algorithms in SciPolymer are based on the establishment of quantitative structure-property relationships (QSPR) based on low order connectivity indices and atom and group correction contributions. Formulations for bulk property predictions are easy to assess and understand by students and researchers alike.

"In addition to SciPolymer's proven value to the research community, it has become the number one teaching tool for our students who need to understand basic organic chemistry as it applies to the structure and properties of polymers.Students can rapidly get up to speed by using its powerful visualization capabilities, but where it really excels is that by using the SciPolymer-Alchemy interface, new or existing polymer structures can be examined fast by a click of the mouse, and their properties are available at any time". Mary G. Chisholm, Penn State University.
 

Evolutionary Molecular Design ™ (EMD) has been able to address known limitations of existing commercially available computational chemistry software. EMD more accurately describes the interaction between compounds and their biological targets with the simultaneous use of an extensive set of parameters not considered by other computational methods.  Identification of molecular features important for activity and subsequent de novo design occur in one fully integrated computer-based process. Additionally, EMD fully explores multiple binding modes by not requiring pre-alignment of molecules. This along with other key features of EMD enables Nanodesign® to quickly and precisely generate a small, focused set of qualified drug candidates, using a small amount of data as input.

There are several ways in which the EMD approach can enhance your drug discovery activities.  EMD can design optimized leads using a small number of structurally diverse, moderately active hits from combinatorial chemistry, high throughput screening or natural product screening. Additionally, EMD can design optimized, novel drug candidates utilizing existing proprietary molecules that have been abandoned as a result of toxicity or other unsuitable characteristics.  Nanodesign can rejuvenate inactive/shelved projects and help recapture investments made in early R&D.  In these and other ways, Nanodesign is able to work with other drug discovery technologies that you employ.  Nanodesign is actively seeking R & D
collaborative arrangements to capitalize on the design capabilities of its powerful, proprietary technology.  Nanodesign is confident that our drug design capabilities will strengthen your discovery efforts.

Balaban, A. T. (ed.). From Chemical Topology to Three-Dimensional Geometry. Plenum Press, New York, 1997.
This book has half of its chapters written by well-known authors who devised new molecular and shape descriptors for WSAR/QSPR, based both on graph-theoretical and on geometrical approaches. The second half contains chapters dealing with the theory of cage compounds, including fullerenes, nanotubes, and tori.

Devillers, J. and Balaban, A.T. (eds.) Topological Indices and Related Descriptors in QSAR and QSPR. Gordon and Breach, UK.
Ca 800 pages appearing in August 1999.

Lemont B. Kier and Lowell H. Hall, Molecular Structure Description: The Electrotopological State, Academic Press  1999, San Diego, ISBN Number 0-12-406555-4
The electrotopological state is a new approach to defining key structural features of a molecule in drug design.  Combining both electronic and topological attributes the E-State index facilitates the development of a structure-activity model, the definition of a pharmacophore, and the search through a database for similar or dissimilar compounds.  The background for the method, the concept of the intrinsic state, and the E-State index as a function of the atom or group within the field of all atoms within a molecule are all described in this primer for a new strucural paradigm.
Software on the bundled CD ROM allows the reader to compute and display the E-State indices for molecules, while examples in the book illustrate strategies that can be used in drug research.
KEY FEATURES
  * Reflects the rich experience of the authors as developers of the E-State method
  * Provides a first-hand account of the origins and use of this strategy
  * Includes E-Calc software on CD ROM for hands-on experience with the method
  * Offers strategies for drug design.
 

Molecular Diversity in Drug Design, ed. P.M. Dean, Hardbound, ISBN 0-7923-5980-1, November 1999, 268 pp., NLG 245.00 / USD 150.00 / GBP 85.00, http://www.wkap.nl/book.htm/0-7923-5980-1

This book focuses on the theoretical problems associated with molecular diversity as it is being applied in the pharmaceutical industry. Therefore, this book deals with algorithms that are involved in understanding chemical space and selection of diverse sets of structures. The algorithms also deal with the problem of focused diversity where chemical libraries are being created within a structured physical volume.

Diversity is necessarily connected to combinational chemistry, although this book is limited to the application of diversity methods to combinational chemistry and does not deal with synthetic methods. It is this focus on algorithms and strategies for exploiting molecular diversity that makes it different from books on combinational chemistry.

  Molecular Structure Description: the Electrotopological State by Lemont B. Kier and Lowell H. Hall Academic Press, San Diego, 1999,  ca. 290 pp., $99-95 (including CD)

The chemical literature, and especially that relating to QSAR, is not short of publications concerning new ways in which to describe molecules, and the ways in which such descriptors can be applied. So do we need yet another set of descriptors? The question can be answered by understanding what they mean, how easily they are derived, how readily they can be applied, what they can tell us about the system under investigation, and (most importantly) whether they are any better than the ones already in use.

Strictly speaking, E-state indices are not all that new; Kier and Hall first introduced them almost ten years ago (Pharm. Res. 7 (1990) 801-807). However, they have not been very widely used, partly perhaps because they have been thought, like molecular connectivities, to be difficult of interpretation. Kier and Hall’s new book on the E-state aims to dispel such ideas, and to show the power of E-state indices in describing molecular structure; the enthusiastic foreword by Gerry Maggiora of Pharmacia Upjohn sets the tone for the book.

The book comprises ten chapters, which are headed: introduction; the electrotopological state; significance and interpretations of the E-state indices; extended forms of the E-state; strategies for use of the E-state; database applications – molecular similarity and diversity; structure-activity studies – atom-level indices; structure-activity studies – mixed indices; structure-activity studies – atom-type indices; future directions of E-state indices. Included also is a CD for the calculation of E-state indices, and a 41-page user’s guide which is bound into the book.

What are E-state indices? Kier and Hall first define a d value of an atom i as the number of its s electrons, whilst d^v is defined as the number of its s,p and lone-pair electrons (excluding bonded hydrogens). The intrinsic or I-state is then defined as (d^v - d)/d; this characterises the local environment of the atom. The E-state of the atom is then obtained by incorporating the influence of the rest of the atoms in the molecule. This is done by firstly taking the difference in I-state values between atom i and  another atom j, and then dividing by the square of the number of bonds rij in the shortest path between the two atoms: DI_ij = (I_I – I_j)/r_ij². The E-state of atom i (Si) is given by the sum of its I-state value and the values of all the perturbing terms from the contributions of the rest of the atoms in the molecule: S_i = I_i + EjDIij. The total E-state of the molecule is given by ESj. Although the method of calculating perturbations due to other atoms is clearly an approximation, the approach seems intuitively satisfying. In chapter 4 the authors deal with various extended forms of the E-state, in particular for dealing with polar hydrogen atoms and with heterogeneous sets of compounds without a common skeleton; the latter are known as atom-type indices, in which atoms and hydride groups are classified according to a valence state scheme (the most commonly used atom-type E-state index is the sum of the E-state indices for a specific atom type).

In chapter 3 Kier and Hall discuss the significance and interpretation of E-state indices. They point out that the indices are a composite of electronic and topological attributes, and thus might not correlate well with, for example, atomic charges. They found also only a “moderate” correlation with Taft s* values. Excellent correlations have, however, been found with NMR chemical shifts, since the latter reflect both electronic and steric contributions. It is reasonable to assume, if the E-state index for a particular atom is found to be significant in a QSAR correlation, that the atom concerned is involved in some way (e.g. receptor binding) in the process under investigation, and Kier and Hall discuss the interpretation of many of the E-state QSARs that they report.

Are E-state indices any better than other descriptors used in QSAR? They are certainly quickly and easily calculated, and of course there is no need for energy minimisation or other molecular manipulation prior to calculation. Whether they are “better” in the sense of giving improved correlations is just about impossible to judge from the book. Numerous examples are given of QSARs involving E-state indices, both alone and in combination with other descriptors, and many of these are excellent correlations; for example, a study of 15 benzimidazoles as influenza virus inhibitors yielded a QSAR containing two E-state indices with r2 = 0.91, s = 0.16 and F = 63. However, I could find only one example where E-state indices were compared with other descriptors (atomic charges). This study involved the inhibition of MAO by a series of 24 hydrazides, and the best three-parameter models were: with E-state indices, r2 = 0.90, s = 0.19, F = 61, and with atomic charges, r2 = 0.80, s = 0.27, F = 26. The authors should have provided many more comparisons to enable the reader to judge the relative ability of E-state indices to model biological and chemical effects.

In chapter 6, Kier and Hall discuss the application of E-state indices to the problem of molecular similarity and diversity, and show that atom-type indices can provide a measure of similarity. Their final chapter, on future directions of E-state studies, examines a number of interesting possibilities such as group E-state indices, iterative procedures for improving the calculation of perturbation effects on an atom, handling cis-trans isomers and gauche-trans conformers, and E-state indices for bonds and bond types.

The CD-ROM, called E-Calc, that comes with the book enables one to calculate all E-state indices for a given compound. Structures can be inputted only by drawing them using the Sketcher Window. Whilst this is simple and quick to use, it is a pity that SMILES and other input formats cannot be used. Neither can files be imported into E-Calc, so structure input for a large data-set can be tedious. The lack of this facility is of particular concern in view of Kier and Hall’s comments in chapter 6 of the application of E-state indices to molecular similarity and diversity. (It should perhaps be mentioned here that Hall’s MOLCONN-Z software, which calculates E-state indices as well as molecular connectivities, does allow various input formats to be used, and does allow files to be imported. Why could this not have been done for E-Calc?). I was puzzled at first to find that the layout of the Sketcher Window was different from that shown on page E-19 of the book. However, I soon discovered that the Element Table (shown on page E-20) is brought up by clicking on button M, and the atom-label menu by clicking on the right-hand mouse button. That sorted, the program was easy to use, and there is a good help facility. There is more help available, with comments on many topological indices, by clicking on the E-Calc Help? file, which is on the CD but which is accessed externally to E-Calc. I found one small bug in the software, namely that, when using the zoom facility, zooming in or out  sometimes did not delete previous smaller or larger structures, with the result that the screen resembled a spider’s web!

Maggiora, in his foreword to the book, says that “the use of E-state indices is in its infancy. Nevertheless, based on the type and variety of studies carried out to date, I believe that they have a promising future”. I concur with this. In Liverpool we have already found E-state indices to be of value in modelling genotoxicity, and I anticipate that they will find wider and wider acceptance and application in a wide range of QSAR and modelling studies. I commend the use of E-state indices, and of this book, to all who are interested in research involving molecular structure description. At $99-95, the book is not cheap, but nevertheless is good value for money.

John Dearden
School of Pharmacy and Chemistry
Liverpool John Moores University
UK

ESSENTIAL JOB FUNCTIONS/DESCRIPTIONS:
Interaction with medicinal chemists, combinatorial chemists, the compound repository, and the HTS group to decrease the time to a clinical candidate.

The work involves:
analyzing large datasets for molecular diversity; and developing computational models for biological and physical properties; and and building, searching, and analyzing databases of reactions, small molecules, and proteins.

SKILL/EXPERIENCES REQUIREMENTS
Ph.D. in chemistry (organic or medicinal preferred) or a related area; and training or experience in biological or medicinal chemistry; and ability to interact scientifically with medicinal and combinatorial chemists.

Computer skills required:
experience with generating biologically relevant molecular properties of large datasets of chemical structures; and experience with automated quantitative structure-activity recognition; and facility with multivariate statistical/chemometric techniques; and experience in applying data mining software to medicinal chemistry related problems.

We will be looking for these particular skills and experiences and hope to identify someone who will bring unique skills to the group.

YVONNE C. MARTIN
D-47E AP10/2
ABBOTT LABORATORIES
100 ABBOTT PARK RD
ABBOTT PARK IL  60064-6100

Fax: 847 937 2625
e-mail: yvonne.c.martin@abbott.com

This VCH journal is considered to be the "home" journal of THE QSAR AND MODELLING SOCIETY. Editors are Prof. Michael Wiese, University of Halle, and Prof. Gerd Folkers, ETH Zurich.

and consider that a publication in this journal will reach your audience of QSAR and modelling colleagues much better than a publication in JACS, JCICS, JMC, Biochemistry, etc.
Of course, Ferenc Darvas remains the Editor of the Abstracts Section. Please consider also to subscribe personally to the QSAR journal. It's good and it's cheap, extremely cheap for members of our Society (call VCH, phone +49-6201-6060, for the current price). Of course, the publisher Wiley-VCH would also like to encourage you to order a personal copy of this important journal. First of all, it has a relatively high impact factor, as compared to many other journals, and second, an incredibly low price for personal subscriptions is offered to our members:

When ordering a subscription (e-mail subservice@wiley-vch.de), please identify yourself as a member of our Society and don’t forget to ask for the special price.
 

R. Benigni, L. Passerini, A. Pino and A. Giuliani: The information content of the eigenvalues from modified adjacency matrices: large scale and small scale correlations
M. Feher, K. Kanai, I. Hermecz, A. Lopata and I. Novak: A 3D-QSAR analysis of prolyl endopeptidase inhibitors
R. Cercos-del-Pozo, F. Perez-Gimenez, M. Salabert-Salvador, F. Garcia-March and M. Murcia-Soler: New hypolipaemic agents designed by molecular topology: Pharmacological Studies of 2,6-Di-tert-butyl-4-methylpyridine and 2,6-Di-tert-butylpyridine
V. Segarran, M. Lopez, H. Ryder and J. Palacios: Prediction of drug permeability based on GRID calculations
A. Tsantili-Kakoulidou, A. Varvaresou, T. Siatra-Papastaikoudi and O. Raevsky: A comprehensive investigation of the partitioning and hydrogen bonding behavior of indole containing derivatives of 1,3,4-thiadiazole and 1,2,4-triazole by means of experimental and calculative approaches

Kluwer Academic Publishers is pleased to offer members of The QSAR and Modelling Society a special individual subscription rate to the Journal of Computer-Aided Molecular Design. Further information, including  Instructions for Authors and Tables of Contents can be found at the Journal's homepage:

 
http://www.wkap.nl/journals/jcamd

Subscribers to the Journal of Computer-Aided Molecular Design will automatically receive Perspectives in Drug Discovery and Design as a supplement.

Please send your order, indicating your prefered form of delivery (Normal, Express, Registered, Airmail), to:
Kluwer Academic Publishers
Order Department
PO Box 322
NL-3300 AH  Dordrecht
The Netherlands

Tel: +31-78-639-2392 Fax: +31-78-654-6474
 E-mail: orderdept@wkap.nl

Please contact the Publishing Editor if you have any queries.
Peter Butler
tel: +31-78-639-2312
fax: +31-78-639-2377
E-mail: peter.butler@wkap.nl
http://www.wkap.nl/journals/jcamd

Meetings and Courses
 
ACS Short Course on Bioinformatics

A 2-day ACS short course will be held on bioinformatics.  The  course is on December 16 - 17, 1999, at Indiana University-Purdue University at Indianapolis (IUPUI).  The course instructor is Professor Shankar Subramaniam (University of California, San Diego), developer of the widely-used Biology Workbench, a web-based environment.  Besides the classroom activities, participants will be able to get hands-on experience with modern workstations.

For more information about attending the course, see
http://chem.iupui.edu/acs/bioinformatics/
Or contact Professor Jack Breen at breen@chem.iupui.edu
or tel. 317-274-0052.

1999

 
 

LogP2000: Lipophilicity in Drug Disposition - Practical and Computational Approaches to Molecular Properties Related to Drug Permeation, Disposition and Metabolism

March 5-9, 2000, Lausanne, Switzerland.
Contact: Prof. B. Testa (bernard.testa@ict.unil.ch)
http://www-ict.unil.ch/logP2000.

Second Indo-US Workshop on Mathematical Chemistry,
May 30 - June 3, 2000, Duluth, Minnesota, USA

The Second Indo-US Workshop on Mathematical Chemistry (with applications to drug discovery, environmental toxicology, chemoinformatics and bioinformatics) will be held during May 30 - June 3, 2000 at the University of Minnesota Duluth, Duluth, Minnesota, USA. Dilip K. Sinha, Visva Bharati University (India) and Subhash C. Basak, Natural Resources Research Institute, University of Minnesota Duluth (USA), are the Co-Chairpersons of the Workshop from India and United States, respectively.
The First Indo-US Workshop on Mathematical Chemistry was held during January 9-13, 1998, at Visva Bharati University, Santiniketan, West Bengal, India.  More than one hundred scientists and scholars from four continents (America, Asia, Australia and Europe) participated in the conference.The Second Workshop will bring together leading researchers in the field of Mathematical and Computational Chemistry.  The results of latest research in the field will be presented along with important applications of mathematical and computational chemistry in drug discovery, environmental toxicology, quantitative structure-activity relationships (QSAR), quantitative molecular similarity analysis (QMSA), chemoinformatics and bioinformatics.

For further information please contact:
Subhash C. Basak, Workshop Co-Chairperson
Natural Resources Research Institute
University of Minnesota Duluth
5013 Miller Trunk Highway
Duluth, MN  55811, USA
Tel: (218) 720-4230
Fax: (218) 720-4328
Email: sbasak@wyle.nrri.umn.edu
For further details see the Workshop web site at:
http://wyle.nrri.umn.edu/IndoUS2

2000

 
 
  13th European QSAR Symposium, Contact: Hans-Dieter Höltje

The Institute for Pharmaceutical Chemistry at the Heinrich-Heine-University Düsseldorf is proud to announce the 13th European Symposium on Quantitative Structure-Activity Relationships to be held in Düsseldorf  27 August - 1 September 2000. The congress will be organized by the local organizing committee under the patronage of the Ministry for Education and Science of the state of Nordrhein-Westfalen.

 

The city of Düsseldorf is located in the North-West of Germany,  at short distance from Köln, Bonn and the Ruhrgebiet. It is the seat of the government of the state of Nordrhein-Westfalen.

Chairman of the Congress:

Prof. Dr. H.-D. Höltje
Institut für Pharmazeutische Chemie
Heinrich-Heine-Universität Düsseldorf
Universitätsstr. 1
D-40225 Düsseldorf, Germany
e-mail:  hoeltje@pharm.uni-duesseldorf.de
 

Scientific Programme
The scientific programme will include an opening lecture, plenary lectures, selected communications and poster sessions.
 

Provisional Application
Those interested in attending the Symposium are invited to return the Preliminary Registration Form. Provisional application does not imply any
obligation, but will greatly assist the Organizing Committee in its preparatory work and ensure receipt of the Second Announcement which will be
distributed end 1999.

Correspondence
All correspondence concerning the Symposium should be addressed to:

Congress Secretariat
13th European Symposium on Quantitative-Activity Relationships
Institut für Pharmazeutische Chemie
Heinrich-Heine-Universität Düsseldorf
Universitätsstr. 1
D-40225 Düsseldorf, Germany
Tel: +49 211 811 3835 - Fax +49 211 811 3847
e-mail: qsar2000@pharm.uni-duesseldorf.de
Web: http://www.pharm.uni-duesseldorf.de/qsar2000

 
          QSARs in Environmental Sciences: Crossroads to the XXI Century. The Ninth International Workshop on QSARS in Environmental Sciences) will be held in Bourgas, Bulgaria from September 16-20, 2000. The conference will continue the tradition of previous workshops held in Burlington (Canada), Knoxville (USA), Veldhoven (The Netherlands), Duluth (USA), Belgirate (Italy), Elsinore (Denmark) and Baltimore (USA). For the first time the International Workshop will take place in an eastern European country, with Bulgaria’s rich history as a cultural and religious crossroad serving as a unique backdrop to bring together scientists from North America, Europe, the countries of the former Soviet Union and Asia.
The Workshop is being organized to provide participants opportunities to learn more about the latest achievements and challenges in the following topics:

· Computational Methods in Chemical Parameter Development and Similarity Analysis
· Fate, Biodegradation and Metabolism
· Toxic Mechanisms and Effects Modeling
· 3D QSAR Methods.  Evaluation and Applicability.
· Artificial Intelligence and Decision Support Systems

Professor Ovanes Mekenyan
Chair of the Organizing Committee
 

Miscellaneous

IUPAC Glossaries
Not all new, but good to know!

Medicinal Chemistry, C.G. Wermuth et al., Ann.Rep.Med.Chem. 33 (1998) 385-398.
Computational Drug Design, H. van de Waterbeemd et al., Ann.Rep.Med.Chem. 33 (1998) 397-409.

The IUPAC home page is on http://chemistry.rsc.org/rsc/iupac.htm
 

Annual Fees

Please send annual fees of $10 to Dr. J.W. King, The QSAR and Modelling Society, PO Box 116, Balsam, NC 28707-0016, USA.
Or to Han van de Waterbeemd in the UK (Eurocheques should be drafted in £).
 

 
This page is free for publicity. Interested?
Contact:  Han van de Waterbeemd at fax +44-1304-656433

 Last Updated: June 27, 2001