"Computers at the frontiers of Organic Chemistry"


ORGLIST is the international mailing list of Organic Chemistry on the Internet, founded in March 1997. Since then, thousands of chemists from all over the world have gathered in ORGLIST to discuss Organic Chemistry. The full archive is  available at the web site and is indexed by Google. Useful content, as well as a complex thread of personal relationships, emerged from the interaction of the members through this simple technological framework. The 10 years of ORGLIST in 2007 were celebrated scientifically, with a symposium on "Computers at the frontiers of Organic Chemistry" at the 7th National Meeting of Organic Chemistry of the Portuguese Chemical Society, on July 17, 2007.


The symposium is sponsored by MestreLab.

Program (17th July 2007)

Morning (9h45 – 12h40)

Auditorium 2

Chairperson: Ana M. Lobo

9h45 – 10h30: Henry S. Rzepa: A twisted link between chemistry, maths, molecular biology (and music)

Chairperson: Henry Rzepa

11h00 – 11h10: Introduction

11h10 – 11h45: Scott Boyer: Computational models to aid safety-directed drug design

11h45 – 12h20: Valerie J. Gillet:Deriving structure-activity relationship in heterogeneous datasets

12h20 – 12h40: Bruce F. Milne: Two-parameter classifier for prediction of PKC-z modulating behaviour of xanthones


Afternoon (15h15 - 16h30)

Meeting room

Chairperson: Henry Rzepa

15h15 – 15h50: Nuno Palma: Regioselectivity of cathecol-O-methyltransferase catalyzed reaction: combined theoretical and experimental studies

15h50 – 16h25: Santiago Dominguez: From MestReC to Mnova: a revolutionary approach to NMR

16h25: Conclusion


A Twisted Link Between Chemistry, Maths, Molecular Biology (and Music)

Henry S. Rzepa

Department of Chemistry, Imperial College, London SW7 2AY, U.K.

Both chirality and aromaticity are cornerstone concepts for organic chemistry. Both had their origins in the  1840s or thereafter in the work of  Pasteur, van't  Hoff and LeBel for the former and Faraday, Loschmidt, Kekule, Armstrong for the latter, this reaching its first stage of theoretical maturity with  Huckel's quantum mechanical analysis  in the 20th Century (the famous  4n+2 rule).

For a long period, these two concepts were thought to be exclusive; after all aromaticity manifested almost entirely in flat (achiral) benzenoid rings!

Another concept, topology, also originated in the 1840s, having been coined by the mathematician Johann Listing, who also proposed fascinating topological objects such as trefoil knots, and rings now better known by their co-discoverer,  Mobius. In the 1960s, the concepts of Mobius topologies and aromaticity started merging. The chemist Heilbronner proposed aromaticity rules for  Mobius cycles, although he did not identify such cycles as being chiral (this property appears to have been gradually realised only years later, although its difficult to find this expressed in print).  The first
such Mobius molecule was only synthesized in 2003; it was not however particularly aromatic! Meanwhile, in 1978 molecular biologists had discovered the fascinating twists and knots in cyclic DNA, via James Wang's topoisomerase enzymes. This was expressed using a concept known as supercoiling, and a new generation of mathematicians formalised this into an equation expressing a so-called linking number, which is comprised of twist and writhe;

Lk = T + W     ...(1)

Applied extensively to the properties of cyclic DNA, these concepts did not migrate at all to organic chemists, who by and large dealt with much smaller molecules. Listing in 1847 had also introduced the concept of paradromic winding, which in modern language maps to imparting further twists to the basic Mobius topology. In 2005, we fused these various concepts from chemistry, topology and molecular biology, recognising that a new form of aromaticity based on double- and higher twisted conjugated, and importantly chiral, rings could be possible. We identified various interesting candidate molecules, but were surprised by how relatively stable they appeared (by computation), given they were at least twice as twisted as the classical Mobius rings. We found a resolution to this paradox in equation (1).  The (quantum mechanical) instability we realised is associated with T and not with W.  We have now computed values of T and W for a range of topologically interesting (and chiral!) systems, and approximately, those that appear the most synthetically interesting have large values of W compared to T.  So W (the writhe) can be regarded as a
fundamentally new property of cyclic conjugated molecules, and one moreover that might be associated with stability. This has led to our proposal that  eqn  (1) and  the Huckel 4n+2 rule can be combined as follows; If Lk is even (measured in units of  pi), aromaticity is implied for 4n+2 cyclic conjugated electrons  ... (2)

If Lk is odd,  aromaticity is implied for  4n cyclic conjugated electrons  ... (3)

Intriguingly both T and W are chiral indices, and they can act together or oppose to create some fascinating novel chiral isomerisms.

In a general sense, this type of aromaticity is chiral, and benzene like systems are very much the achiral exceptions (having  Lk = 0).

At the end of the talk, I will speculate on some potential real world applications of this fascinating new form of chiral aromaticity, particularly to the design of new chiral metal ligands, and perhaps even mention another interest of ours, the Semantic Web, and how this might in the future enable more efficient fusion of diverse ideas and concepts (linking is a fundamental concept there as well!).

An interactive version of the talk will be available at  and I anticipate that a podcast will also be made available.

Computational models to aid safety-directed drug design

Scott Boyer

Senior Principal Scientist and Head, Computational Toxicology

Global Safety Assessment, AstraZeneca R&D, Mölndal, Sweden

Access to metabolism and toxicology data is critical to effective decision making in early drug discovery projects.  Often in such projects little is known about the therapeutic target and usually even less is known about potential metabolism or adverse effects of the chemical series being investigated.  Simply providing unstructured metabolism- and safety-related information on targets and chemical series to project teams trying to make decisions is not adequate due to the varied nature and quality of metabolism and toxicology data.  This presentation gives examples of how relevant data can be structured, mined and in some cases modelled to enhance decision-making.  Project examples will be presented of QSAR models and their interpretation, including characterization of the underlying assay error for better interpretation of the model results, development of SAR systems that support decision-making and enhance awareness around such endpoints as metabolism/P450 activation, mutagenesis, hERG and reactive intermediates.  In general, metabolism and toxicology data should be structured depending on, 1) its intended use, 2) its overall quality and 3) its internal data structure (text vs. numerical) to assure its optimum use.  Brief examples of the varying data types and their usage in project decision making will be presented along with some strategies for hypothesis generation around adverse events using a combined approach of molecular modelling/virtual screening and text mining.  Together, these tools, built to be appropriate to the various data types, represent a basic toolkit for the toxicologist and drug metabolism scientist needing to make meaningful contributions to the myriad decisions made in early drug discovery projects.

Deriving Structure-Activity Relationships in Heterogeneous Datasets

Valerie J. Gillet

Department of Information Studies, University of Sheffield, 211
Portobello St., Regent Court, Sheffield, United Kingdom

Machine learning algorithms such as Binary Kernel Discrimination and Support Vector Machines have become popular methods for the analysis of high-throughput screening data. While they have been shown to be effective ways of deriving predictive models they suffer from the disadvantage that the models are not easily interpretable. Here we describe a new method based on genetic programming. A training set of active and inactive molecules are represented as reduced graphs and genetic programming is used to evolve reduced graph queries (subgraphs) that are best able to separate the actives from the inactives. The classification rate is determined using the F-measure which combines recall and precision into a single objective. The resulting queries are validated on datasets not used in deriving the queries, for proof of their predictive power.
As well as being useful models for prediction, the queries contain interpretable structure-activity information encoded within the reduced graph nodes. Results are presented for the well known MDDR dataset and also for GSK in-house screening data.

Regioselectivity of the Catechol-O-Methyltransferase Catalyzed Reaction: Combined Theoretical and Experimental Studies

Nuno Palma§, Maria L. Rodrigues, Margarida Archer, Maria J. Bonifácio§, Ana I. Loureiro§, David A. Learmonth§, Maria A. Carrondo, Patricio Soares-da-Silva§

§ Department of Research & Development, BIAL, Portugal

Instituto de Tecnologia Química e Biológica (ITQB), Portugal

This work presents combined theoretical and experimental studies [1,2] of the regioselectivity of O-methylation of nitrocatechol-type inhibitors of the enzyme Catechol-O-methyltransferase (COMT).

As a case study, two simple regioisomeric nitrocatechol-type inhibitors of COMT, containing a benzoyl substituent attached at the meta- or at the ortho-position, respectively, relative to the nitro group, were studied with regards to their interaction with the catalytic site of the enzyme and the in vitro regioselective formation of their mono-O-methyl ether metabolites. It is shown that the particular substitution pattern of the classical nitrocatechol pharmacophore has a profound impact on the regioselectivity of O-methylation.

In order to provide a plausible interpretation of these results, a comprehensive analysis of the protein-inhibitor interactions and of the relative chemical susceptibility to O-methylation of the catechol hydroxyl groups was performed by means of docking simulations and molecular orbital calculations. The major structural and chemical factors that determine the enzyme regioselectivity of O-methylation are identified and the X-ray structure of the complex of COMT with one of the two inhibitors (BIA 8-176) is disclosed. This is the first reported structure of COMT complexed with a nitrocatecholic inhibitor having a bulky substituent group in ortho position to the nitro group. Structural and dynamic aspects of this complex are analyzed and discussed, in the context of the present study.

[1]   Palma, P. N., Bonifacio, M. J., Loureiro, A. I., Wright, L. C., Learmonth, D. A., and Soares-Da-Silva, P. Drug Metab Dispos 2003, 31, 250-8.

[2]   Palma, P. N., Rodrigues, M. L., Archer, M., Bonifacio, M. J., Loureiro, A. I., Learmonth, D. A., Carrondo, M. A., and Soares-da-Silva, P. Mol Pharmacol 2006, 70, 143-53.

This work was partly funded by Fundação para a Ciência e Technologia/AdI trough research projects POCTI/COMT-HUM/2002 and POCTI/BME/38306/2001 and grant SFRH/BD/5228/2001 (M.L.R)

From MestReC to Mnova: A revolutionary Approach to NMR

Nikolay Larin, Stan Sykora, Santiago Domínguez, Carlos Cobas

MESTRELAB RESEARCH, Santiago de Compostela, Spain

High Resolution NMR spectroscopy is undoubtedly one of the most important methods used in organic chemistry for structure determination. Traditionally, organic chemists used to spend considerable time processing their NMR data to get the best experimental NMR as starting material for the lengthy and non trivial task of spectral analysis. Furthermore, recent years have witnessed dramatic improvements in high-throughput NMR in such a way that spectral processing and analysis have emerged as a new bottle neck due to the large amount of  spectral data available.

In this work we present Mnova, the new incarnation of MestReC as a novel software solution offering an innovative paradigm for the unattended NMR data processing and new tools such as spectral prediction, simulation and fitting algorithms to facilitate structure verification and elucidation for organic chemists.

Two-Parameter Classifier for Prediction of  PKC-ζ Modulating Behaviour of Xanthones

Bruce F. Milne §, Madalena M.M. Pinto §

§ Centro de Estudos de Química Orgânica, Fitoquímica e Farmacologia da Universidade do Porto; Laboratório de Química Orgânica, Faculdade de Farmácia da Universidade do Porto, Rua Aníbal Cunha 164, 4050-047 Porto

Protein kinase C ζ (PKC-ζ) occurs in many tissues in the body and is associated with numerous cellular processes including differentiation, mitogenesis, migration and apoptosis. PKC-ζ is implicated in the progression of a variety of disease states including colon cancer, inflammatory bowel conditions, leukaemia, melanoma and T-cell mediated hepatitis. Studies in our research group [1, 2] have identified a number of simple xanthone derivatives displaying varying levels and types of PKC- ζ modulating activity. Although structurally very similar, this group of compounds includes both potent activators and inhibitors of PKC-ζ and therefore it is desirable to have a method with which to attempt to predict which region of the activity spectrum new derivatives might fall into.

In an attempt to rationalize the behaviour of these compounds a computational QSAR study was undertaken and a two-parameter decision tree developed that successfully classifies all of the xanthones previously tested as either activators, inhibitors or inactive. In addition, a small selection of non-xanthone PKC-ζ inhibitors have been appended to this study and these are also correctly classified by the decision tree developed for the xanthones.

[1] Saraiva, L.; Fresco, P.; Pinto, E.; Sousa, E.; Pinto, M.; Gonçalves, J., Bioorganic and Medicinal Chemistry, 2002, 10, 3219-3227.

[2] Saraiva, L.; Fresco, P.; Pinto, E.; Sousa, E.; Pinto, M.; Gonçalves, J., Bioorganic and Medicinal Chemistry, 2003, 11, (7), 1215-1225.

Acknowledgments: FCT (I&D 226/94), FEDER and POCI for financial support.

BFM is funded by FCT post-doctoral research fellowship SFRH/BPD/17830/2004.