The molecular weight and electrotopological E-state indices were used to estimate by Artificial Neural Networks aqueous solubility for a diverse set of 1291 organic compounds. The neural network with 33-4-1 neurons provided highly predictive results with r(2) = 0.91 and RMS = 0.62. The used parameters included several combinations of E-state indices with similar properties. The calculated results were similar to those published for these data by Huuskonen (2000). However, in the current study only E-state indices were used without need of additional indices (the molecular connectivity, shape, flexibility and indicator indices) also considered in the previous study. In addition, the present neural network contained three times less hidden neurons. Smaller neural networks and use of one homogeneous set of parameters provides a more robust model for prediction of aqueous solubility of chemical compounds. Limitations of the developed method for prediction of large compounds are discussed. The developed approach is available online at http://www.lnh.unil.ch/~itetko/logp.

A new method, ALOGPS v 2.0 (http://www.lnh.unil.ch/~itetko/logp/), for the assessment of n-octanol/water partition coefficient, log P, was developed on the basis of neural network ensemble analysis of 12 908 organic compounds available from PHYSPROP database of Syracuse Research Corporation. The atom and bond-type E-state indices as well as the number of hydrogen and non-hydrogen atoms were used to represent the molecular structures. A preliminary selection of indices was performed by multiple linear regression analysis, and 75 input parameters were chosen. Some of the parameters combined several atom-type or bond-type indices with similar physicochemical properties. The neural network ensemble training was performed by efficient partition algorithm developed by the authors. The ensemble contained 50 neural networks, and each neural network had 10 neurons in one hidden layer. The prediction ability of the developed approach was estimated using both leave-one-out (LOO) technique and training/test protocol. In case of interseries predictions, i.e., when molecules in the test and in the training subsets were selected by chance from the same set of compounds, both approaches provided similar results. ALOGPS performance was significantly better than the results obtained by other tested methods. For a subset of 12 777 molecules the LOO results, namely correlation coefficient r(2)= 0.95, root mean squared error, RMSE = 0.39, and an absolute mean error, MAE = 0.29, were calculated. For two cross-series predictions, i.e., when molecules in the training and in the test sets belong to different series of compounds, all analyzed methods performed less efficiently. The decrease in the performance could be explained by a different diversity of molecules in the training and in the test sets. However, even for such difficult cases the ALOGPS method provided better prediction ability than the other tested methods. We have shown that the diversity of the training sets rather than the design of the methods is the main factor determining their prediction ability for new data. A comparative performance of the methods as well as a dependence on the number of non-hydrogen atoms in a molecule is also presented.

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.

The origin and laminar arrangement of the homolateral and callosal projections to the anterior (AAF), primary (AI), posterior (PAF) and secondary (AII) auditory cortical areas were studied in the cat by means of electrophysiological recording and WGA-HRP tracing techniques. The transcallosal projections to AAF, AI, PAF and AII were principally homotypic since the major source of input was their corresponding area in the contralateral cortex. Heterotypic transcallosal projections to AAF and AI were seen, originating from the contralateral AI and AAF, respectively. PAF received heterotypic commissural projections from the opposite ventroposterior auditory cortical field (VPAF). Heterotypic callosal inputs to AII were rare, originating from AAF and AI. The neurons of origin of the transcallosal connections were located mainly in layers II and III (70-92%), and less frequently in deep layers (V and VI, 8-30%). Single unit recordings provided evidence that both homotypic and heterotypic transcallosal projections connect corresponding frequency regions of the two hemispheres. The regional distribution of the anterogradely labeled terminals indicated that the homotypic and heterotypic auditory transcallosal projections are reciprocal. The present data suggest that the transcallosal auditory interconnections are segregated in 3 major parallel components (AAF-AI, PAF-VPAF and AII), maintaining a segregation between parallel functional channels already established for the thalamocortical auditory interconnections. For the intrahemispheric connections, the analysis of the retrograde tracing data revealed that AAF and AI receive projections from the homolateral cortical areas PAF, VPAF and AII, whose neurons of origin were located mainly in their deep (V and VI) cortical layers. The reciprocal interconnections between the homolateral AAF and AI did not show a preferential laminar arrangement since the neurons of origin were distributed almost evenly in both superficial (II and III) and deep (V and VI) cortical layers. On the contrary, PAF received inputs from the homolateral cortical fields AAF, AI, AII and VPAF, originating predominantly from their superficial (II and III) layers. The homolateral projections reaching AII originated mainly from the superficial layers of AAF and AI, but from the deep layers of VPAF and PAF. The laminar distribution of anterogradely labeled terminal fields, when they were dense enough for a confident identification, was systematically related to the laminar arrangement of neurons of origin of the reciprocal projection: a projection originating from deep layers was associated with a reciprocal projection terminating mainly in layer IV, whereas a projection originating from superficial layers was associated with a reciprocal projection terminating predominantly outside layer IV.(ABSTRACT TRUNCATED AT 400 WORDS)