摘要:

The three-dimensional (3D)-pattern search problem can be summarized as finding, in a molecule, the subset of atoms that have the most similar spatial arrangement as those of a given 3D pattern. For this NP-complete combinatorial optimization problem we propose, by analogy to the travelling salesman problem, a new method taking advantage of the capability of Hopfield-like neural networks to carry out combinatorial optimization of an objective function. This objective function is built from the sum of the differences of interatomic distances in the pattern and the molecule. Here we present the implementation we have found of the 3D-pattern search problem on Hopfield-like neural networks. Initial tests indicate that this approach not only successfully retrieves a given pattern, but can also suggest partial solutions having one or two atoms less than the given pattern, an interesting feature in the case of local conformational flexibility of the molecule. The distributed representation of the problem on Hopfield-like neural networks offers a good perspective for parallel implementation.

ISSN:1062-936X    

DOI:10.1080/10629369308028822

出版商:Taylor & Francis Group    

年代:1993

数据来源: Taylor

摘要:

In addition to its outstanding abilities in both classification and fitting, the neural network can also accurately predict the values of the untrained region. To rationalize this ability of prediction, the authors mathematically discussed the valid region of prediction. Based on such a background, the authors proposed “descriptor mapping” in the QSAR analysis, which visualizes the nonlinear dependencies between structural parameters. A variable of the linear multiple regression analysis in the QSAR study is supposed to be linear to the biological intensity and is independent of other variables. Analysis by the descriptor mapping method discloses the reality.

ISSN:1062-936X    

DOI:10.1080/10629369308028823

出版商:Taylor & Francis Group    

年代:1993

数据来源: Taylor

摘要:

A new method for predicting the impact sensitivity of explosive molecules is presented. This method makes use of a network of formal neurons. The experiment uses 124 molecules belonging to different families. The molecular descriptors taken into account are the molecule's oxygen balance and the enumeration of certain groups. The results obtained are satisfactory: 80% of the molecules are correctly classed on a scale of four sensitivities. Comparison with a multivariate linear regression analysis gives a slight advantage to the neural network method.

ISSN:1062-936X    

DOI:10.1080/10629369308028824

出版商:Taylor & Francis Group    

年代:1993

数据来源: Taylor

摘要:

Artificial neural networks (ANN) can be used for the direct QSAR analysis of percent effect biological data, thus avoiding the bias introduced by arbitrarily chosen classes and the loss of information due to prior classification. For two data sets the ANN results are compared with those obtained by adaptive least squares and nonlinear regression analyses. In comparison with the other methods the neural network shows higher predictive power and does not require an explicit equation relating the observed effect to physico-chemical descriptors.

ISSN:1062-936X    

DOI:10.1080/10629369308028825

出版商:Taylor & Francis Group    

年代:1993

数据来源: Taylor

摘要:

Partially recurrent neural networks with different topologies are applied for secondary structure prediction of proteins. The state of some activations in the network is available after a pattern presentation via feedback connections as additional input during the processing of the next pattern in a sequence. A reference data set containing 91 proteins in the training set and 15 non-homologous proteins in the test set is used for training and testing a network with a modified, hierarchical Elman architecture. The network predicts the secondary structures α-helix, β-sheet, and “coil” for each amino acid. The percentage of correctly classified amino acids is 67.83% on the training set and 63.98% on the test set. The best performance of a three-layer feedforward network is 62.7% on the same test set. A cascaded network, where the outputs of the recurrent network are processed by a second net with 13 × 3 inputs, four hidden and three output units has a predictive performance of 64.49%. The best corresponding feedforward net has a performance of 64.3%.

ISSN:1062-936X    

DOI:10.1080/10629369308028826

出版商:Taylor & Francis Group    

年代:1993

数据来源: Taylor

摘要:

The aim of this paper was to explore the usefulness of a backpropagation neural network (BNN) to estimate the biodegradability of benzene derivatives. 127 chemicals selected from the BIODEG data bank (Syracuse Research Corporation, 1992) were described by means of 20 structural descriptors taking into account the nature and position of the substituents on the benzene ring. Three classes of biodegradability were selected and modelled from the BNN. A 20/5/3 BNN (α = 0.8 and η = 0.5) correctly classified 92% (104/113) of the training and 86% (12/14) of the testing sets. The results were compared to those produced by the BIODEG probability program (Syracuse Research Corporation, Version 2.13).

ISSN:1062-936X    

DOI:10.1080/10629369308028827

出版商:Taylor & Francis Group    

年代:1993

数据来源: Taylor

摘要:

A neural network was applied to a large, structurally heterogeneous data set of mutagens and nonmutagens to investigate structure-property relationships. Substructural data comprising a total of 1280 fragments were used as inputs. The training of the back-propagation networks was directed by an algorithm which selected an optimal subset of fragments in order to maximize their discriminating power, and a good predictive network.

ISSN:1062-936X    

DOI:10.1080/10629369308028828

出版商:Taylor & Francis Group    

年代:1993

数据来源: Taylor

摘要:

The field half-lives of 110 pesticides were modelled using a backpropagation neural network (NN). The molecules were described by means of the frequency of 17 structural fragments. Before training the NN, different scaling transformations were assayed. Best results were obtained with correspondence factor analysis which also allowed a reduction of dimensionality. The training and testing sets of the NN analysis gave 95.5% and 84.6% of good classifications, respectively. Comparison with discriminant factor analysis showed that a backpropagation NN was more appropriate to model the field half-lives of pesticides.

ISSN:1062-936X    

DOI:10.1080/10629369308028829

出版商:Taylor & Francis Group    

年代:1993

数据来源: Taylor

摘要:

Two types of neural networks were used to establish relationships between chemical structure and musk odour of 79 nitrobenzenic compounds. Substituents on the five free sites of the benzene ring (one position was always occupied by at−butyl group) were described using three volume descriptors and three electronegativity descriptors. Musk odour was coded by a binary variable. First a classical network with two hidden layers containing six and three neurons was used. This network gave a better classification (94%) than that obtained by linear discriminant analysis (81%). The odour was men predicted using a leave-ten-out procedure, with 77% of correct prediction for the whole sample. Then a dual two-way network was built to mimic the symmetry of the problem (two sides on a molecule, two muskophore patterns). This network recognized both patterns already known to chemists and gave 99% of correct classifications by taking into account substitution in all positions. As a side benefit of the modified network structure it was possible to evaluate the influence of each of 19 substituents in each of the five possible positions.

ISSN:1062-936X    

DOI:10.1080/10629369308028830

出版商:Taylor & Francis Group    

年代:1993

数据来源: Taylor

摘要:

The present study was performed on a group of 27 derivatives of phenylsuccinimides, of which only 12 were active against maximal electrical shock in spite of the structural similarities of these compounds. The work consisted of four main parts: 1. crystallographic investigations of a subset of chosen compounds; 2. conformational analysis of characteristic molecules from the investigated series, performed by means of molecular mechanics calculations; 3. molecular orbital optimization of all the molecules using the MNDO method starting with conformations obtained in 2; 4. molecular electrostatic potential (MEP) analysis which was performed on the semiempirical (MNDO) and ab initio levels. This research showed that MEP maps provide a signature that distinguishes between active and inactive compounds. There are MEP minima close to the two carbonyl oxygens of the imide ring, and although the magnitude of the difference between the two minima is approximately constant, the sign of the difference provides an activity index. The initial orientations of phenylsuccinimide molecules in relation to the receptor are not equivalent and they depend on the potential distribution around both the succinimide molecules and around the receptor. In the active compounds the negative potential difference at the discussed points most probably influences the initial set-up of the molecules in relation to the receptor and results in a considerably higher probability of the molecules being bound at the right place on the receptor.

ISSN:1062-936X    

DOI:10.1080/10629369308028831

出版商:Taylor & Francis Group    

年代:1993

数据来源: Taylor