Non-linear QSAR modeling by using multilayer perceptron feedforward neural networks trained by back-propagation.

The use of multilayer perceptrons (MLP) feedforward neural networks trained by back-propagation (BP) for non-linear QSAR model building is presented and explained in detail through a case study. This method was compared with others often used in this field, such as multiple linear regression (MLR), partial least squares (PLS) and quadratic PLS (QPLS). The case study deals with a series of 18 alpha adrenoreceptors agonists belonging to three different classes (alpha-1, alpha-2 and alpha-1,2) according to their different pharmacological effects. Each of them is described by 15 chemical features (the X block). Six pharmacological responses were also measured for each one to build the matrix of biological responses (the Y block). The results obtained indicated a slightly better performance of MLP against the other procedures, when using the correlation coefficient of the observed versus predicted response plots as an indicator of the goodness of the fit.

Computational program for evaluating and optimizing response-surface curves based on uniform shell designs.

A program for optimizing uniform shell designs, called DOEHLOPT, is proposed. The scope and features of DOEHLOPT are discussed and compared with those of a general purpose program (GOSSET). DOEHLOPT was applied to a worked example for the sake of illustration.

Inorganic indicators of the origin of edible salts marketed in Spain from a chemometric approach.

Twenty edible salts marketed in Spain were studied for content of inorganic constituents in order to identify one or more indicators of the origins of salts. In Spain the principal source of salt is seawater, either from the Mediterranean coast or from the Atlantic coast. In addition, some manufactured salts in the dietary category were also considered. Five "Atlantic," 12 "Mediterranean," and 3 dietary salts were analyzed for potassium, magnesium, calcium, strontium, manganese, zinc, and sulfate contents and for alkalinity. Salt samples were first dried for moisture determination. Metals were determined by inductively coupled plasma atomic emission spectrometry, sulfate was determined by turbidimetry, and alkalinity was determined by titration with hydrochloric acid. All the results were expressed as the means for triplicate experiments, as mg of analyte per kg of salt on a dry basis. From these results and by application of chemometric strategies (principal component analysis, variable-variable plots, and cluster analysis) it was found that the potassium content distinguishes sea salts from dietary salts very well, and the contents of strontium and manganese discriminate salts of Atlantic origin versus salts of Mediterranean origin. Dietary salts showed levels of potassium higher than the sea salts, and salts of Atlantic origin had greater strontium and manganese content values than salts of Mediterranean origin.

Spectrophotometric determination of total procyanidins in wine vinegars.

A widely-used method for the spectrophotometric determination of procyanidins in wines has been adapted to wine vinegar samples. Reagent concentrations have been established and the analytical method tested for possible matrix effects. The recovery of catechin was approximately 98% and the limit of reliable measurement was 0.48 mg l(-1). The within-day and between-day precisions were evaluated and according to the two-tailed F-test the precisions were statistically equivalent. Application to wine vinegars obtained by traditional and quick acetification methods showed differences in concentration between the two groups.

Classification of tea samples by their chemical composition using discriminant analysis.

Multivariate analysis in combination with pattern recognition procedures has been applied to samples of green and black tea. Discriminant analysis has been used for classification purposes. Aqueous extract, polyphenols, amino acids, caffeine, theobromine and theophylline were used as chemical descriptors.