Skip to main content
main-content

Über dieses Buch

Bringing together scientists from the various disciplines of chemistry who are actively engaged in developing software and using computers to solve their problems was the main objective of the 4th workshop 'Computers in Chemistry' (November 22-24, 1989) held in Hochfilzen, Tyrol. Fields covered include molecular modelling, chemometrics, synthesis planning, computer science.

Inhaltsverzeichnis

Frontmatter

Representation of Chemical Structures; Factual Databases

Chemical Substructure Search on CD-ROM

Does the CD-ROM with its high storage capacitiy but rather slow random access times offer the opportunity to search large scale chemical structure data bases with a PC in acceptable search times? This article briefly describes the problems and introduces the S4 substructure search system which is a positive answer to that question.

Helmut Maier, Dirk Walkowiak

Documentation of Special Bond Types and Evaluation of Factual Data from “B-Base”

Computer assisted handling of chemical structures — the international language of chemistry — has become very comfortably within the last 30 years. Problems arise by leaving the field of classic valence bond type structures and incorporating metal complexes or cluster compounds for instance. This report shows the concept for documentation of Lewis adduct compounds, transition metal complexes with π bonds including d- or f-orbitals and cluster compounds within the factual database “B-Base”. Furthermore it describes the evaluation of factual data for visualisation via graphical comparison. A new feature of ChemBase software, now applicated to “B-Base”, enables the linkage of substructure and graphs of shift ranges.

H. Nöth, E. Striedl

AUTONOM: Automatic Generation of IUPAC-Names from Structural Input

An algorithm has been developed for the computer generation of organic compound names from structure diagram input. The structure diagrams of compounds are input via a graphic interface and the names are generated purely on the basis of the resulting molecular connection tables. The system, when fully implemented and installed, will be used in the production of both the Beilstein Database and the Beilstein Handbook. It is also expected that a PC-version of AUTONOM will provide the chemical community with a new valuable tool. This paper describes the general design of AUTONOM and presents a detailed analysis of software solutions adopted during the work on the system.

J. L. Wisniewski, L. Goebels, A. Lawson

Keys to the Beilstein Database

A ring searching algorithm

The experiences with a retrieval algorithm for ring systems are described. The program was tested with 3 million structures of the Beilstein database. The ring searching is a fast and accurate replacement for name fragment searching and gives additional possibilities for structure retrieval from large structure files.

László Domokos

The Integration of the Cambridge Crystallographic Data Files into the Relational Information Network of the German Cancer Research Center

The Cambridge Crystallographic Data Files (CCDF) represent the largest collection of published crystallographic data. The CCDF, originally implemented in a hierarchical data base design, have been converted to a relational architecture in order to make data access compatible with the concepts developed for the spectroscopic information system, SPEKTREN II, of the German Cancer Research Center (DKFZ), i.e.: the procedures for handling chemical structures and related user-interface software. In addition, the 3-D data of the CCDF can be stored, modified and used to generate starting conformations for Molecular Modeling calculations.

Elke Lang, Thomas Förster, Claus-Wilhelm von der Lieth

The Gmelin Information System the Connection between Handbook and Database

The Gmelin Information System consists of the Gmelin Handbook of Inorganic Chemistry and the Gmelin Database. Currently, the Gmelin Handbook is actually created without computer data processing; the procedure of entering data into the database started at the beginning of 1989. Both components will be integrated into the Gmelin Information System. Therefore, the primary information for the handbook production (information about compounds, facts, and bibliographic data) will be delivered from the Gmelin Database. This information and careful study of the original literature forms the basis of the handbook. The distinctive characteristic of the handbook is the evaluation from the view of a scientist of published data concerning a special compound. Within this process, the data from the Gmelin Database will be edited in some cases and after completing a handbook project these data will be uploaded from the handbook department. From this point of view, the connection between the handbook and the database is a process of changing information from one component of the Gmelin Information System to the other. The main qualities of both the handbook, which are described above, and of the database (availability of on-line factual information) should be able to be retained. As an interface between the handbook and the database we have developed the Gmelin Electronic Card Index. It is the card file for the handbook production and employs methods of electronic data processing. The development was based on the hypertext software product CAMS4. This system is a means of presenting information of all kinds as a sequence of pages. Pages are organized into a logical structure called a knowledge base with the program. The pages are connected in the form of a tree structure. The system provides a number of functions in order to access a dBase III Plus compatible data file. Further developments from the CAMS4 hypertext system for the Gmelin Electronic Card Index include a flexible generator of templates and functions for direct access to an ADABAS database under VAX/VMS. The flexible generator for templates is necessary to present the contents of the Gmelin Database (815 fields) in individual form for special handbook projects. This template acts as a filter on the database for downloading, working, and uploading.

A. Nebel, G. Olbrich, R. Deplanque

The WODCA System

An Integrating Environment for the Chemist

The WODCA system was developed to serve as the central module of an integrating environment for the interactive handling of programs for synthesis planning and reaction prediction. The program combines both flexibility and user-friendliness by providing fully mouse-and menu-driven input as well as a powerful command language. The program contains a data base management subsystem for access to starting materials catalogs, reads and writes various structure formats and interfaces smoothly to a set of display programs, input facilities and other utilities.

J. Gasteiger, W. D. Ihlenfeldt

Structure and Properties; Bridging the Gap

Computer — Assisted Analysis of Qualitative Structure/Activity Relations of Organic Molecules

The topic is a theoretical approach to the search for activity — carrying substructures. For this the theoretical foundations of structure/activity relations are exploited. A data structure is presented that affords structure/activity correlations, and its realisation in CORREL II, a modified version of CORREL. The upper bound of performance that can be achieved by present day computers and the corresponding presequisites are discussed. Practical examples are given.

K. Bley, I. Ugi

Prediction of the Threshold Soot Index for Hydrocarbon Fuels with Randic’s Topological Indices

Topological indices and simple structural descriptors were used to calculate the threshold soot indices of 87 hydrocarbons of different types. Regression analysis yielded for the topological indices an only slightly better result than for the simple structure parameters. Better equations could be achieved by dividing the dataset into two appropriate cluster and correlating each cluster seperately to the TSI — values.

Degen

A Graph Theoretical Method to Estimate Substance Data

The lack of data is a major problem for predicting the fate of chemicals in the environment. Data sets taken from handbooks or online databases are often incomplete and inaccurate. We have developed two computer tools to complete such data sets via estimation and to test the quality of the data set. Both programs use methods from graph theory to make optimal use of a complex system of relations while maintaining a very simple user interface.

R. Brüggemann, B. Münzer

PIMO, a program visualizing HMO-results by producing transferable graphics output

The program PIMO presented here uses either simple topological input matrices or CHEMTEXTR-connectivity tables to produce graphics output of the results of HMO- or Ω-calculations.

Beate Heinrich, Erich Schreiner, Dieter Spielbauer, Markus Wolperdinger, Hans-Ulrich Wagner

The Calculation of Electrolyte Solution Properties with the Help of the ELDAR Data and Method Bank Exemplified by Electrolyte Conductance

Chemical models based on mean force potentials are generally used to calculate properties of electrolyte solutions from measurements with the help of the Gauss-Marquardt method for least squares fits. For these calculation processes ELDAR offers a data bank managing fact knowledge, a method bank managing algorithmic knowledge and a communication manager supplying the input of the modules. The interaction of data and method bank for the automatic calculation of electrolyte solution properties and the production of basic data under the control of the communication manager is exemplified for electrolyte conductance.

J. Barthel, H. Popp, G. Schmeer

Molecular Modelling

Triangulation of Molecular Surfaces

An algorithm for the generation of a solid molecular surface from a given number of surface points in 3D space is presented. This surface is represented by a triangle mesh. Such a mesh of triangles is not uniquely defined by the position vectors of the surface points, and there is no straight forward algorithm for triangulation. In this paper a hierarchy of algorithms is described which successfully works for the automatic triangulation of molecular surfaces of all sizes. Some examples are given for the application of this solid surface approach in molecular modelling of proteins.

W. Heiden, M. Schlenkrich, C.-D. Zachmann, J. Brickmann

MOBY: Molecular Modeling on the PC

The molecular Modeling programme MOBY is introduced that makes the standard features of high level modeling tools available to PC users, without the limitations on speed and the restrictions on the molecular systems found for other PC programmes. Several extensions to the standard features like a detailed energy analysis and quantum chemical methods have been implemented. Some examples for the applicability of MOBY are shown on the poster.

Udo Höweler

ROCOCO: Reference Online Library for the Computer Aided Construction of Molecular Geometries Computer Aided Construction of Realistic Molecular Models Using a Knowledge Base (I)

This article describes a general concept for the construction of molecular models which accurately mirror the structure of the real molecule without large scale optimization. Imitating the procedure of an experienced chemist the computer programme system analyses the topological data of a molecule -presented as a 2D structural formula — and determines the wanted geometric parameters using an internal data base, which contains the values of the respective fragments of the molecule. The results can be displayed in tables or printed onto a plotter. In the current stage of development the programme together with the data base is a “comfortable, electronically readable handbook” for bond lengths and bond angles of aliphatic and heteroaliphatic compounds as well as strain-free aliphatic cyclic and aliphatic heterocyclic compounds containing the elements C, H, O, N, F, Cl, Br and I.

Benno Krieg, Peter Keller

MMGEO A Versatile Tool for MMX Users

MMGEO is a versatile tool for MMX users, it is a so called software interface between MM2/MMX and other programs like SCHAKAL, CHEMTEXT and CDCHEM. It has many other build-in functions for calculating atomic distances, bond angles and dihedral angles.

N. Reuter

Spectral Data; Acquisition and Interpretation

JCAMP — DX, A Standard?

The implementation of the JCAMP-DX standard as transport format for Infrared Spectra has been a great step forward in the direction of laboratory communication. The standard has however been implemented by different manufacturers with varying degrees of success. Despite the explicit and clear language used in the publication of the standard, the various data system manufacturers have generally so far failed to produce software capable of complying with the regulations laid down by McDonald and Wilks. The types of errors found in the present generation of JCAMP-DX files and handling software will be discussed with a view to assisting people writing there own JCAMP-DX software. A programme capable of checking the transport file format, diagnosing errors, and where possible correcting errors will be presented. Future developments in the field of JCAMP-DX type standard formats, with particular reference to the structure transport format JCAMP-CS will also be presented.

A. N. Davies, H. Hillig, M. Linscheid

Automatic Interpretation of 2D-NMR-Spectra [1]

During the past decade two-dimensional NMR-spectroscopy [2] has developed into one of the most powerful tools for organic structure determination. On spectrometers equipped with a sample changer 2D-NMR-spectroscopy can be completely done under computer control producing a vaste amount of data, thus shifting the bottle-neck from accumulation to data interpretation. These sophisticated experimental techniques can only be fully utilized by appropriate computer software supporting the spectroscopist during the whole process of data interpretation. Our recently published approach for automatic interpretation of NMR-spectra [3] has been expanded and is now able to use distance-constraints derived from different types of 2D-NMR experiments. The main advantage of this algorithm is its capability to work only with one-dimensional data thus leading to the possibility to interpret incomplete 2D-NMR data describing only a part of the structure. Some examples selected from natural product chemistry will be given proofing the power of this concept.

Peter Haas, Wolfgang Robien

CDDS — A Personal Computer Based System for Automated Interpretation of GC/MS-Analyses

The application of chemometric detectors in gaschromatography/mass spectrometry has been shown to be useful in recent years [1,2]. The most important drawbacks of the systems developed so far are the lack of user friendliness and the complicated development procedure to obtain new chemometric detectors with a new selectivity. The system presented in this paper has a clearly defined and graphically oriented user interface and makes it possible to develop new detectors in a convenient and efficient way.

H. Lohninger

EDAS — MS Exploratory Data Analysis of Mass Spectra

EDAS-MS is a software-tool, developed for the mass spectroscopic part of the spectroscopic information system SPECINFO. EDAS-MS uses methods from multivariate statistics, especially linear mapping procedures belonging to the field of exploratory data analysis. It supports the chemist in the search for relationships between spectral features und substructures and can be a help to determine the chemical class membership of an unknown compound.Essential parts of EDAS-MS are: Input routines for mass spectra, procedures for the generation of spectral features, scaling and feature selection methods, principal component analysis and similar mapping methods and statistical substructure analysis. The user interface supports interactive working with the two-dimensional representations of data including access to spectra and chemical structures.

W. Werther, K. Varmuza

Towards the Automatic Generation of a Mass Spectrum from the Structure of a Compound

A system is developed that automatically predicts for a given organic structure the main fragmentation steps and the corresponding peaks in its mass spectrum. This is accomplished by interpretation of fragmentation rules stored in an external file. Physicochemical parameters calculated by empirical methods are used to decide which fragmentations and rearrangements occur preferentially.

W. Hanebeck, K. Rafeiner, K.-P. Schulz, P. Röse, J. Gasteiger

Factor analysis of spectral data from chemical reactions

Factor analysis (FA) is a mathematical method, which was originally developped for examination of psychological tests. It was introduced by Thurstone [1], who also showed, that this method could help the psychologists in evaluation of a large amount of data more objectively.

G. Gauglitz, S. Weiß

Normalization of in Situ-Spectra in Thin Layer Chromatography

Normalization is a necessary step in order to compare two spectra- Mostly point normalization at the maximum of absorbance is used. Other methods are area normalization or the use of regression techniques. Reflectance spectra in TLC are not obeying Bouguer-Lambert-Beer’s law, therefore normalization has to be done by two-dimensional linear regression or nonlinear regression using a saturation function. Both methods are leading to good results. Nonlinear regression needs more computing time and more sophisticated computer programming.

S. Ebel, J. S. Kang, W. Windmann

Algorithms for Use in Purity Control of Drugs Using UV-Spectroscopy

UV-spectroscopy is a nonspecific analytical method, but in contrast to this it is possible to develop methods for purity control of drugs using this type of spectroscopy. Two algorithms will be described in this paper: a differential method subtracting the spectra of the analysis and a reference followed by nonparametric tests and the rank of a matrix containing both spectra. Base of both methods is a sophisticated comparison of both spectra. With two examples will be confirmed the proposed method. In some cases it is possible to detect impurities below 0.1%.

S. Ebel, C. Urban, S. Windmann

DSYM-PC a Novel Program to Simulate HR-NMR Spectra for Spins I=1/2 on IBM-Compatible Computers of PC/XT/AT Type

DSYM-PC is a novel program to simulate HR-NMR-spectra. Modern menue-driven techniques are used to present screen- and graphical output on HP-Laser Printer in HPGL-standard. DSYM-PC is designed for research and teaching.

S. Goudetsidis, G. Hägele

Chemical Reactions and Synthesis Planning

IGOR2: A Program System for Generating Chemical Reactions and Structures

IGOR2 is a computer program that generates chemical reactions and structures. The development of IGOR2 took place 1986–1989 in several steps. The present status is a PC version with a menu guided user surface and graphic output.

J. Bauer

Reaction Databases in a University Chemistry Department — Online or in-House?

Chemical reactions [1] are described by chemists in a variety of ways, with graphical descriptions using structural formulae dominating. Figure 1 shows the information elements commonly used in such descriptions (graphics enclosed in boxes): the main elements are chemical structures with their roles. Reaction information sources should permit queries for any of the elements given in fig. 1. singly or in combination, because questions addressed to them show a great variety which cannot really be accomodated by anything less

Engelbert Zass

Towards Synthesis Planning Aids Through Databank Analysis

Since the late sixties Synthesis Planning Programs have been a major field of research in computational chemistry. Usually one finds two opposite approaches: A combinatorial or an empirical approach. The Chemical Synthetic Planning System currently being developed at RISC-LINZ, with a design philosphy based on the powerful methods offered by higher mathematics and artificial intelligence, is of semiem pineal nature, and therefore requires the analysis of reaction databanks. The system in its present prototype form is made up of more than 30,000 CommonLISP lines of code. The system is built around the philosophy of flexibility in which, through, for example, its preprocessing stage, the essential chemical information is abstracted from specific information in existing databanks. The ORGSYN© database has been analysed and a pool of transforms automatically generated. The performance of the algorithm is described and examples given. The set of transforms is then used in a Synthesis Planning Aid called “RETROSYN”, which is described.

Edward Blurock, Thomas Strelow

Implementation of Synthesis Strategies in PROLOG

The program STRATOS is presented, that generates synthesis plans for aliphatic compounds. They are made without the help of a reaction database, but solely by examination of the relationships between functional groups in the molecules. Implementation of the program in PROLOG and FORTRAN77 is discussed.

M. Wagener, J. Gasteiger

EROS 6.0, a Knowledge Based System for Reaction Prediction — Application to the Regioselectivity of the Diels-Alder Reaction

The EROS 6.0 system is used to analyse the effects determining reactivity in organic reactions and this knowledge can be applied to the prediction of still unknown reactions. Reactions are predicted by modelling the mechanism, the simulation of elementary reaction steps. Each reaction type is described by a reaction rule. Deduction or induction can be used for the derivation of reaction rules. In the inductive way, methods of machine learning and statistics are used to derive rules from databases of example reactions. Application of the system is demonstrated by modelling the regioselectivity in the Diels-Alder reaction.

Peter Röse, Johann Gasteiger

CARSA (Computer-Assisted Research in Synthesis and Application)

CKB’s computer chemistry project CARSA (Computer-Assisted Research in Synthesis and Application) is a modular structured system, which at present consists of the following parts: WIFODATA (substance — and finding — documentation; database)QSAR (programs for quantitative structure-activity-relationships)RDSS (Reaction Design by Synthon Substitution; program package for synthesis planning) The system is devised as a tool for synthesis chemistry and of special importance in industrial research.

R. Moll

Simulations of Structures, Reactions and Properties

Modelling of Polymer Gel Formation and Gel Reactions with Monte Carlo Methods for 3-dimensional Networks

The thermal treatment of poly-(p-methyl-styrene) was investigated experimentally. In contrast to polystyrene, gels are formed and decomposed again, depending on temperature. Based on the experimental results a reaction model for polymer decomposition, polymer gel formation and polymer gel degradation was developed. A Monte Carlo simulation program using methods from percolation theory is presented which covers thermal gel formation, gel formation caused by radiation, gel formation during polymerization, thermal, mechanical and ultrasonic decomposition of polymer gels and highly branched polymers.

Johannes Kinkel, Hanns J. Ederer, Klaus H. Ebert

Molecular Dynamic Simulation of the Interface Aqueous Ionic Solution / Lipid Membrane

The interface between an ionic solution and a membrane, modelled by an ensemble of COO− groups with translational and rotational degrees of freedom, is studied by molecular dynamics (MD) computer simulations. The charged membrane leads to a layering of the ions and the water molecules. Several water layers can be distinguished with structural properties very different from those found in the bulk phase.

K. Nicklas, J. Böcker, M. Schlenkrich, P. Bopp, J. Brickmann

Electroanalytical simulations. 10. The simulation of fast second order reactions in electrochemical systems

Two methods for the simulation of fast chemical second order reactions coupled to an electron transfer at an electrode are used in conjunction with orthogonal collocation to calculate theoretical cyclic voltammograms: spline collocation and the technique of the heterogeneous equivalent. The techniques prove to produce reliable numerical results for the model of an electro-chemically induced irreversible dimerization reaction after the application of a non-linear space coordinate transformation. Also, simulations of the ece/disp model (two electron transfers coupled by a chemical equilibrium and a disproportionation reaction) with the spline collocation method are discussed. It is shown in this case that the assumption about the DISP equilibrium usually required for the solution of the model equations can be dropped and the spline algorithm can be used throughout the parameter space.

Bernd Speiser

The Rashevsky-Turing System: Two Coupled Oscillators as a Generic Reaction-Diffusion Model

A two-compartment reaction-diffusion system is investigated both analytically and numerically. Owing to the system’s symmetry, a complete characterization of the three fixed points and their respective stability is achieved. Dependent on the coupling strength, different types of chaos can be observed in computer simulations. Some implications of the model are pointed out.

M. Klein, G. Baier, O. E. Rössler

Photoreactions in Solids Experiment and Simulation

To reduce the problems in evalution of photokinetic rate equations, the liquid reaction system has to be stirred in order to get closed solutions. In the case of the examination of photoreactions in solids, the evaluation requires the application of numerical methods. In the special case of azobenzene in silgel 604A a further problem shows up, the photolysis product does not stay fixed at its site of production, but exhibits diffusion. It is demonstrated that even in this case an evalution is possible using modified rate equation, which take into consideration time and space dependencies as well.

D. Fröhlich, G. Gauglitz

Computer Science; New Methods and Their Applications

The Role of Machine Learning in Knowledge Acquisition

Acquiring the knowledge for a knowledge-based system has proven to be a difficult task. Machine learning techniques are one possible approach to tackle this problem. Three case studies are described which show that machine learning techniques can produce superior results than more traditional knowledge acquisition techniques. Finally, some conclusion drawn from these examples are presented.

Kai Zercher, Bernd Radig

Neural Networks

The basic principles of neural networks (weight vector, layers, connections, feed-back corrections, etc.) are described and discussed. In particular, Hopfleld and Hamming net, Kohonen learning, and back-propagation algorithm are described more in detail. Besides the basic architectures of these nets some of the problems that may arise in the actual applications are pointed out.

Marjan Tušar, Jure Zupan

Application of Fuzzy Neural Network to Spectrum Identification

The main concept of using neural network models for pattern association and pattern classification is outlined. Applications for handling spectroscopic patterns are demonstrated with storing UV-spectra recorded under different experimental conditions in a neural network and identifying unknown spectra by presenting them to the network. Essentially Kosko’s Adaptive Bidirectional Associative Memory [2] is explored that enables to learn similar spectra of one chemical compound as well as spectra of different chemical identity. The final comparison of sample and reference spectra is carried out by fuzzy set operations.

Matthias Otto, Uwe Hörchner

Tools for Automatic Program Generation

The efficiency of program development can be improved by employing appropriate software tools. In this paper we discuss the compiler generator LEX-YACC and the GENTRAN module of the computer algebra (C. A.) system REDUCE. Their use is demonstrated for examples from physical chemistry. The LEX-YACC program, a UNIX tool, is employed to generate the user interface of a spectral analysis program. REDUCE is used to modify REDUCE subroutines in a symbolic manner. The resulting code is then translated into “C” code by the GENTRAN module and the resulting subroutine inserted into a numerical program.

H. Armitage, A. Khuen, D. Ziessow

Automatic Translation from Fortran to “C”

Programs written in FORTRAN can be translated to “C” with an automatic translator. It is shown, which techniques are used for this translation.

Alfred Mechsner

Miscellaneous Subjects

LABORG Laboratory optimisation by the elimination of weaknesses and the implementation of LIMS

In the past few years the possibilities for quantitative improvements in laboratory services, as well as faster determination of results based on the introduction of information technology, have grown rapidly.However, after the initial investemnt has been made, the expected increase in efficiency often does not occur to the extent planned. This is most often caused by the optimisation of isolated laboratory techniques. A comprehensive investigation of the company’s laboratories, and an associated system and development plan, create the efficient laboratory organisation which is a necessary condition for implementing of the Laboratory-Information-Management-System (LIMS).

Andreas Gasteiger

C - LIT

C-LIT is a literature data base program which helps chemists to organize their collection of literature. The features of this Shareware program are described.

Jürgen Kammerer

Possibilities and Limitations of Combined Database Enquiries Modelled on “Environmental Significance of Aluminum”

Environmental problems are discussed more and more, therefore it is obvious to use databases as comprehensive sources of mankind’s knowledge in order to achieve unemotionally answers on holistic questions. The usefullness of databases in the area of research and development is well known, especially their importance as tools for rapid access to the demanded information. Nevertheless the possibility to act against an almost emotionally handled discussion concerning problems like “mankind and environment” remained to be seen until the present. A search for literary citations dealing with the “environmental significance of aluminum”, a wide selection of primary references should be found. By starting an OnLine-search within the relevant databases, it could be expected to achieve a comprehensive point of view, because it is possible to avoid a biased representation of facts a priori.

A. Müller, E. Striedl

Backmatter

Weitere Informationen