Prediction of drug solubility from molecular structure using a drug-like training set.

Using a training set of 191 drug-like compounds extracted from the AQUASOL database a quantitative structure-property relationship (QSPR) study was conducted employing a set of simple structural and physicochemical properties to predict aqueous solubility. The resultant regression model comprised five parameters (ClogP, molecular weight, indicator variable for aliphatic amine groups, number of rotatable bonds and number of aromatic rings) and demonstrated acceptable statistics (r2 = 0.87, s = 0.51, F = 243.6, n = 191). The model was applied to two test sets consisting of a drug-like set of compounds (r2 = 0.80, s = 0.68, n = 174) and a set of agrochemicals (r2 = 0.88, s = 0.65, n = 200). Using the established general solubility equation (GSE) on the training and drug-like test set gave poorer results than the current study. The agrochemical test set was predicted with equal accuracy using the GSE and the QSPR equation. The results of this study suggest that increasing molecular size, rigidity and lipophilicity decrease solubility whereas increasing conformational flexibility and the presence of a non-conjugated amine group increase the solubility of drug-like compounds. Indeed, the proposed structural parameters make physical sense and provide simple guidelines for modifying solubility during lead optimisation.

QSAR studies using the parashift system.

A novel way of describing molecules in terms of their surfaces and local properties at the surfaces is described. The use of these surfaces and properties to explain chemical reactivity and model simple molecular properties has already been demonstrated. This study reports an examination of the use of these descriptions of molecules to model a simple chemical interaction (complex formation) and a diverse set of mutagens. Both of these systems have been modelled successfully and the results are discussed.

Modelling mutagenicity using properties calculated by computational chemistry.

The recent advances in combinatorial chemistry and high throughput screening technologies have led to an explosion in the numbers of possible therapeutic candidates being produced at the early stages of drug discovery. This rapid increase in the number of chemicals to be classified results in a greater need for alternative methods for the prediction of toxicity. Most QSAR models for mutagenicity have been constructed for congeneric series. The prediction requirements of the pharmaceutical industry, however, cover quite diverse chemical structures. This paper reports a study of mutagenicity data for a diverse set of 90 compounds. Good discriminant models have been built for this data set using properties calculated by the techniques of computational chemistry. Jack-knifed (leave one out) predictions for these models are of the order of 85%.

Simultaneous prediction of aqueous solubility and octanol/water partition coefficient based on descriptors derived from molecular structure.

It has been shown that water solubility and octanol/water partition coefficient for a large diverse set of compounds can be predicted simultaneously using molecular descriptors derived solely from a two dimensional representation of molecular structure. These properties have been modelled using multiple linear regression, artificial neural networks and a statistical method known as canonical correlation analysis. The neural networks give slightly better models both in terms of fitting and prediction presumably due to the fact that they include non-linear terms. The statistical methods, on the other hand, provide information concerning the explanation of variance and allow easy interrogation of the models. Models were fitted using a training set of 552 compounds, a validation set and test set each containing 68 molecules and two separate literature test sets for solubility and partition.

[Application of neural networks using the volume learning algorithm for the study of structure-activity relationship in chemical compounds]. / Primenenie neironnykh setei, ispol'zuiushchikh algoritm prostranstvennogo obucheniia dlia issledovaniia sviazei struktura-aktivnost' khimicheskikh soedinenii.

A volume learning algorithm for artificial neural networks was developed to quantitatively describe the three-dimensional structure-activity relationships using as an example N-benzylpiperidine derivatives. The new algorithm combines two types of neural networks, the Kohonen and the feed-forward artificial neural networks, which are used to analyze the input grid data generated by the comparative molecular field approach. Selection of the most informative parameters using the algorithm helped to reveal the most important spatial properties of the molecules, which affect their biological activities. Cluster regions determined using the new algorithm adequately predicted the activity of molecules from a control data set.

Volume learning algorithm artificial neural networks for 3D QSAR studies.

The current study introduces a new method, the volume learning algorithm (VLA), for the investigation of three-dimensional quantitative structure-activity relationships (QSAR) of chemical compounds. This method incorporates the advantages of comparative molecular field analysis (CoMFA) and artificial neural network approaches. VLA is a combination of supervised and unsupervised neural networks applied to solve the same problem. The supervised algorithm is a feed-forward neural network trained with a back-propagation algorithm while the unsupervised network is a self-organizing map of Kohonen. The use of both of these algorithms makes it possible to cluster the input CoMFA field variables and to use only a small number of the most relevant parameters to correlate spatial properties of the molecules with their activity. The statistical coefficients calculated by the proposed algorithm for cannabimimetic aminoalkyl indoles were comparable to, or improved, in comparison to the original study using the partial least squares algorithm. The results of the algorithm can be visualized and easily interpreted. Thus, VLA is a new convenient tool for three-dimensional QSAR studies.

Quantitative measurements of edge-to-face aromatic interactions by using chemical double-mutant cycles.

Synthetic H-bonded zipper complexes have been used to quantify the magnitude of an edge-to-face aromatic interaction between a benzoyl group and an aniline ring. Four chemical double-mutant cycles were constructed by using a matrix of nine closely related complexes in which the aromatic rings were sequentially substituted for alkyl substituents. The stability constants and three-dimensional structures of the complexes were determined by using 1H NMR titrations in deuterochloroform at room temperature. The value of the interaction energy is similar in all cases, the average is -1.4 +/- 0.5 kJ mol(-1). The scope and limitations of the double-mutant approach are explored, and the consequences of conformational equilibria are discussed.

Unsupervised forward selection: a method for eliminating redundant variables.

An unsupervised learning method is proposed for variable selection and its performance assessed using three typical QSAR data sets. The aims of this procedure are to generate a subset of descriptors from any given data set in which the resultant variables are relevant, redundancy is eliminated, and multicollinearity is reduced. Continuum regression, an algorithm encompassing ordinary least squares regression, regression on principal components, and partial least squares regression, was used to construct models from the selected variables. The variable selection routine is shown to produce simple, robust, and easily interpreted models for the chosen data sets.

Polynomial neural network for linear and non-linear model selection in quantitative-structure activity relationship studies on the internet.

This article presents a self-organising multilayered iterative algorithm that provides linear and non-linear polynomial regression models thus allowing the user to control the number and the power of the terms in the models. The accuracy of the algorithm is compared to the partial least squares (PLS) algorithm using fourteen data sets in quantitative-structure activity relationship studies. The calculated data show that the proposed method is able to select simple models characterized by a high prediction ability and thus provides a considerable interest in quantitative-structure activity relationship studies. The software is developed using client-server protocol (Java and C++ languages) and is available for world-wide users on the Web site of the authors.

Pharmaceutical fingerprinting in phase space. 2. Pattern recognition.

The current study introduces an approach for pattern recognition of drug manufacturers according to their HPLC trace impurity data. This method considers signals in phase space and accounts for two different types of noise: additive and perturbative. The pharmaceutical fingerprints are estimated as mean trajectories of HPLC trace impurity data and are used as reference models for recognition of new data by the minimal length classifier. The chromatographic trace organic impurity patterns collected from six different manufacturers of L-tryptophan are analyzed as an example. The prediction ability of the new method tested using three different cross-validation procedures remains about 95% even if the number of available data in the training sets decreases by 5 times. The accuracy of prediction in phase space is superior compared to results calculated using a Window Preprocessing method and artificial neural networks. The difference in performance between new and previous methods becomes more significant under particular conditions that are more adequate for practical application of the method. In addition, the current approach enables simple and comprehensive interpretation of the calculated results.

Evaluation of an extended period of use for preserved eye drops in hospital practice.

AIM: To evaluate and compare the microbial contamination arising from 1 and 2 weeks' use of eye drops by hospital inpatients and hence determine the validity of apportioning a 2 week in use expiry date for these preparations. METHODS: Eye drop residues were collected from inpatients of Worthing, Southlands, and Brighton General hospitals after 7 days' use (341 samples) and also after 14 days' use (295 samples). The contents of the containers were examined for the presence of contaminating bacteria and fungi. RESULTS: The incidence of microbial contamination was shown to be not significantly different (p > 0.1 chi 2 test) between the 7 and 14 day samples. In addition, the contaminating micro-organisms were of a broadly similar pattern between the two sample groups and were mostly those normally associated with the skin. Less frequent contaminants were organisms of environmental origin. None of the micro-organisms isolated were considered to be of clinical significance and the mean number of cells found per sample was very low. CONCLUSIONS: The evidence therefore suggests that increasing the period of use for eye drops in hospitals from 7 to 14 days would not present a clinically significant threat to patients' health and yet may lead to annual savings to the NHS of Pounds 0.5 million.

Data modelling with neural networks: advantages and limitations.

The origins and operation of artificial neural networks are briefly described and their early application to data modelling in drug design is reviewed. Four problems in the use of neural networks in data modelling are discussed, namely overfitting, chance effects, overtraining and interpretation, and examples are given of the means by which the first three of these may be avoided. The use of neural networks as a variable selection tool is shown and the advantage of networks as a nonlinear data modelling device is discussed. The display of multivariate data in two dimensions employing a neural network is illustrated using experimental and theoretical data for a set of charge transfer complexes.

Neural network studies. 2. Variable selection.

Quantitative structure-activity relationship (QSAR) studies usually require an estimation of the relevance of a very large set of initial variables. Determination of the most important variables allows theoretically a better generalization by all pattern recognition methods. This study introduces and investigates five pruning algorithms designed to estimate the importance of input variables in feed-forward artificial neural network trained by back propagation algorithm (ANN) applications and to prune nonrelevant ones in a statistically reliable way. The analyzed algorithms performed similar variable estimations for simulated data sets, but differences were detected for real QSAR examples. Improvement of ANN prediction ability was shown after the pruning of redundant input variables. The statistical coefficients computed by ANNs for QSAR examples were better than those of multiple linear regression. Restrictions of the proposed algorithms and the potential use of ANNs are discussed.

Analysis of linear and nonlinear QSAR data using neural networks.

The use of feed forward back propagation neural networks to perform the equivalent of multiple linear regression has been examined using artificial structured data sets and real literature data. Their predictive ability has been assessed using leave-one-out cross-validation and training/test set protocols. While networks have been shown to fit data sets well, they appear to suffer from a number of disadvantages. In particular, they have performed poorly in prediction for the QSAR data examined here, they are susceptible to chance effects, and the relationships developed by the networks are difficult to interpret. This investigation reports results for one particular form of artificial neural network; other architectures and applications, however, may be more suitable.

Computational techniques for the prediction of toxicity.

Computational methods for the prediction of biological activity have been in use since the early 1960s. The application of these approaches in attempts to predict toxicological endpoints was reported within a few years of the establishment of the techniques, although early work was criticized on a number of grounds. This report reviews three approaches to the prediction of toxicity with examples of each approach. A new commercially available system (APEX) for the identification of pharmacophores (toxicophores) based on the three-dimensional structures of database compounds has been applied to a literature dataset of mutagenicity results. This program achieved a success rate of approximately 75% when trained on a set of 105 compounds; comparable results from other prediction systems, although trained on a much larger dataset, are approximately 85-90%.

Analysis of the biological and molecular properties of phencyclidine-like compounds by chemometrics.

The quantitative structure-activity relationships (QSAR's) of 3 series of arylcyclohexylamines have been investigated using computational chemistry and multivariate statistics. Principal component analysis of the aromatic ring data set demonstrated some clustering of activity categories. Biological activity of the cyclohexane ring data set was correlated with molar refractivity. These findings may be useful for predicting the activity of novel neuroprotective agents.

Chiral chromatography and multivariate quantitative structure-property relationships of benzimidazole sulphoxides.

Various benzimidazole sulphoxides were chirally resolved employing an amylase-based chiral stationary phase. The structure-property relationships of these compounds were investigated using calculated physicochemical properties, molecular modelling and multivariate statistical techniques. A data set of 254 molecular descriptors was used to represent the series of compounds. Analysis of the data set using principal components analysis and non-linear mapping suggested that the separation factor of each enantiomeric pair was associated with nine molecular properties and, in particular, molar refractivity of the Z substituent and the partial charge of atom 6. The separation factor of a sulphoxide not used in the analysis was well predicted thus suggesting that these models may be used to generalize.

Multivariate quantitative structure-toxicity relationships in a series of dopamine mimetics.

The techniques of principal components analysis and non-linear mapping are routinely used by computer chemists at SmithKline Beecham Pharmaceuticals in the process of drug development by relating the structure of a compound to its chemical activity. To our knowledge these techniques had not previously been applied to the association between the structure of a compound and its toxicological properties. Using a series of 12 structurally related compounds (11 were active dopamine mimetics and one was inactive), of which five were known to be teratogenic and seven were non-teratogenic, it was possible to demonstrate that molecular modelling techniques could be applied to differentiate toxicological data. The structure/property relationships of these compounds were investigated using calculated physicochemical properties, molecular modelling and multivariate statistical techniques. A data set of 56 molecular descriptors was used to represent this series of compounds. Analysis of the data set using principal components analysis and non-linear mapping suggested that teratogenicity was associated with four molecular properties. Moreover, the electronic nature of the 4-phenyl group appeared to be an important determinant of the teratogenesis.

Structure-activity relationships of pyrethroid insecticides. Part 2. The use of molecular dynamics for conformation searching and average parameter calculation.

Molecular dynamics simulations have been performed on a number of conformationally flexible pyrethroid insecticides. The results indicate that molecular dynamics is a suitable tool for conformational searching of small molecules given suitable simulation parameters. The structures derived from the simulations are compared with the static conformation used in a previous study. Various physicochemical parameters have been calculated for a set of conformations selected from the simulations using multivariate analysis. The averaged values of the parameters over the selected set (and the factors derived from them) are compared with the single conformation values used in the previous study.