Bayesian-regularized Genetic Neural Networks Research Articles

QSAR Study of the Inhibitors of the Acetyl-CoA Carboxylase 1 and 2 using Bayesian Regularized Genetic Neural Networks: A Comparative Study

Linear and non-linear quantitative structure-activity relationship (QSAR) models were presented for modeling and predicting anti-diabetic activities of a set of inhibitors of acetyl-CoA carboxylase 1 and 2 (ACC1 and ACC2). Different algorithms were utilized to choose the best variables among large numbers of descriptors and then these selected descriptors were used for non-linear (artificial neural network) and linear (multiple linear regression) modeling. The variable selection methods were consisted of stepwise-multiple linear regression (stepwise-MLR), successive projections algorithm (SPA), genetic algorithm-multiple linear regression (GA-MLR) and Bayesian regularized genetic neural networks (BRGNNs). The prediction abilities of the models were evaluated by Monte Carlo cross validation (MCCV) in variable selection and modeling steps. The results revealed that the best variables for describing the inhibition mechanism of ACC were among topological charge indices, radial distribution function, geometrical, and autocorrelation descriptors. The statistical parameters of R2 and root mean square error (RMSE) indicated that BRGNNs is superior for modeling the inhibitory activity of ACC modulators over the other methods. The sensitivity analysis together with the frequency of the selected molecular descriptors in this work can establish an understanding to the mechanism of ACC inhibitory activity of small molecules.

ChemInform Abstract: Genetic Algorithm Optimization in Drug Design QSAR: Bayesian‐Regularized Genetic Neural Networks (BRGNN) and Genetic Algorithm‐Optimized Support Vectors Machines (GA‐SVM)

AbstractReview: 105 refs.

Genetic algorithm optimization in drug design QSAR: Bayesian-regularized genetic neural networks (BRGNN) and genetic algorithm-optimized support vectors machines (GA-SVM)

Many articles in "in silico" drug design implemented genetic algorithm (GA) for feature selection, model optimization, conformational search, or docking studies. Some of these articles described GA applications to quantitative structure-activity relationships (QSAR) modeling in combination with regression and/or classification techniques. We reviewed the implementation of GA in drug design QSAR and specifically its performance in the optimization of robust mathematical models such as Bayesian-regularized artificial neural networks (BRANNs) and support vector machines (SVMs) on different drug design problems. Modeled data sets encompassed ADMET and solubility properties, cancer target inhibitors, acetylcholinesterase inhibitors, HIV-1 protease inhibitors, ion-channel and calcium entry blockers, and antiprotozoan compounds as well as protein classes, functional, and conformational stability data. The GA-optimized predictors were often more accurate and robust than previous published models on the same data sets and explained more than 65% of data variances in validation experiments. In addition, feature selection over large pools of molecular descriptors provided insights into the structural and atomic properties ruling ligand-target interactions.

QSAR Modeling of 1‐(3,3‐Diphenylpropyl)‐Piperidinyl Amides as CCR5 Modulators Using Multivariate Adaptive Regression Spline and Bayesian Regularized Genetic Neural Networks

AbstractThis study deals with developing a quantitative structure‐activity relationship (QSAR) model for describing and predicting the inhibition activity of 1‐(3,3‐diphenylpropyl)‐piperidinyl derivatives as CCR5 modulators. Applying the multiple linear regressions (MLR) and its inability in predicting the inhibition behavior showed that the interaction has no linear characteristics. To assess the nonlinear characteristics of the inhibition activity artificial neural networks (ANN) was used for data modeling. In order to select the variables needed for developing ANNs, three variable selection algorithms were used: Stepwise‐MLR, genetic algorithm‐partial least squares (GA‐PLS), and Bayesian regularized genetic neural networks (BRGNNs). R2 and root mean square error (RMSE) values for training (t) and leave‐one‐out (LOO) procedures revealed that BRGNNs is a robust algorithm for the variable selection and regression method simultaneously. Due to the ‘black box’ limitation of neural networks, multivariate adaptive regression spline (MARS) technique was used for modeling. A prominent advantage of MARS with respect to ANN is its ability in interpreting of the results of the model. Q2LOO and Rt2 (0.982 and 0.947) reveal that MARS can describe and predict inhibition activity of these modulators and is as robust as ANN. Because the MARS model can explain the activity of molecules, it is a useful model for designing novel CCR5 inhibitors.

Artificial Neural Networks from MATLAB® in Medicinal Chemistry. Bayesian-Regularized Genetic Neural Networks (BRGNN): Application to the Prediction of the Antagonistic Activity Against Human Platelet Thrombin Receptor (PAR-1)

Artificial neural networks (ANNs) have been widely used for medicinal chemistry modeling. In the last two decades, too many reports used MATLAB environment as an adequate platform for programming ANNs. Some of these reports comprise a variety of applications intended to quantitatively or qualitatively describe structure-activity relationships. A powerful tool is obtained when there are combined Bayesian-regularized neural networks (BRANNs) and genetic algorithm (GA): Bayesian-regularized genetic neural networks (BRGNNs). BRGNNs can model complicated relationships between explanatory variables and dependent variables. Thus, this methodology is regarded as useful tool for QSAR analysis. In order to demonstrate the use of BRGNNs, we developed a reliable method for predicting the antagonistic activity of 5-amino-3-arylisoxazole derivatives against Human Platelet Thrombin Receptor (PAR-1), using classical 3D-QSAR methodologies: Comparative Molecular Field Analysis (CoMFA) and Comparative Molecular Similarity Indices Analysis (CoMSIA). In addition, 3D vectors generated from the molecular structures were correlated with antagonistic activities by multivariate linear regression (MLR) and Bayesian-regularized neural networks (BRGNNs). All models were trained with 34 compounds, after which they were evaluated for predictive ability with additional 6 compounds. CoMFA and CoMSIA were unable to describe this structure-activity relationship, while BRGNN methodology brings the best results according to validation statistics.

Comparative modeling of the conformational stability of chymotrypsin inhibitor 2 protein mutants using amino acid sequence autocorrelation (AASA) and amino acid 3D autocorrelation (AA3DA) vectors and ensembles of Bayesian-regularized genetic neural networks

Predicting protein stability changes upon point mutation is important for understanding protein structure and designing new proteins. Autocorrelation vector formalism was extended to amino acid sequences and 3D conformations for encoding protein structural information with modeling purpose. Protein autocorrelation vectors were weighted by 48 amino acid/residue properties selected from the AAindex database. Ensembles of Bayesian-regularized genetic neural networks (BRGNNs) trained with amino acid sequence autocorrelation (AASA) vectors and amino acid 3D autocorrelation (AA3DA) vectors yielded predictive models of the change of unfolding Gibbs free energy change (ΔΔG) of chymotrypsin Inhibitor 2 protein mutants. The ensemble predictor described about 58 and 72% of the data variances in test sets for AASA and AA3DA models, respectively. Optimum sequence and 3D-based ensembles exhibit high effects on relevant structural (volume, solvent-accessible surface area), physico-chemical (hydrophilicity/hydrophobicity-related) and thermodynamic (hydration parameters) properties.

Protein radial distribution function ( P-RDF) and Bayesian-Regularized Genetic Neural Networks for modeling protein conformational stability: Chymotrypsin inhibitor 2 mutants

Development of novel computational approaches for modeling protein properties is a main goal in applied Proteomics. In this work, we reported the extension of the radial distribution function (RDF) scores formalism to proteins for encoding 3D structural information with modeling purposes. Protein-RDF ( P-RDF) scores measure spherical distributions on protein 3D structure of 48 amino acids/residues properties selected from the AAindex data base. P-RDF scores were tested for building predictive models of the change of thermal unfolding Gibbs free energy change (ΔΔ G) of chymotrypsin inhibitor 2 upon mutations. In this sense, an ensemble of Bayesian-Regularized Genetic Neural Networks (BRGNNs) yielded an optimum nonlinear model for the conformational stability. The ensemble predictor described about 84% and 70% variance of the data in training and test sets, respectively.

Amino acid sequence autocorrelation vectors and bayesian‐regularized genetic neural networks for modeling protein conformational stability: Gene V protein mutants

Development of novel computational approaches for modeling protein properties from their primary structure is the main goal in applied proteomics. In this work, we reported the extension of the autocorrelation vector formalism to amino acid sequences for encoding protein structural information with modeling purposes. Amino acid sequence autocorrelation (AASA) vectors were calculated by measuring the autocorrelations at sequence lags ranging from 1 to 15 on the protein primary structure of 48 amino acid/residue properties selected from the AAindex data base. A total of 720 AASA descriptors were tested for building predictive models of the change of thermal unfolding Gibbs free energy change (delta deltaG) of gene V protein upon mutation. In this sense, ensembles of Bayesian-regularized genetic neural networks (BRGNNs) were used for obtaining an optimum nonlinear model for the conformational stability. The ensemble predictor described about 88% and 66% variance of the data in training and test sets respectively. Furthermore, the optimum AASA vector subset not only helped to successfully model unfolding stability but also well distributed wild-type and gene V protein mutants on a stability self-organized map (SOM), when used for unsupervised training of competitive neurons.

QSAR models for predicting the activity of non-peptide luteinizing hormone-releasing hormone (LHRH) antagonists derived from erythromycin A using quantum chemical properties

Multiple linear regression (MLR) combined with genetic algorithm (GA) and Bayesian-regularized Genetic Neural Networks (BRGNNs) were used to model the binding affinity (pK(I)) of 38 11,12-cyclic carbamate derivatives of 6-O-methylerythromycin A for the Human Luteinizing Hormone-Releasing Hormone (LHRH) receptor using quantum chemical descriptors. A multiparametric MLR equation with good statistical quality was obtained that describes the features relevant for antagonistic activity when the substituent at the position 3 of the erythronolide core was varied. In addition, four-descriptor linear and nonlinear models were established for the whole dataset. Such models showed high statistical quality. However, the BRGNN model was better than the linear model according to the external validation process. In general, our linear and nonlinear models reveal that the binding affinity of the compounds studied for the LHRH receptor is modulated by electron-related terms.

Ensembles of Bayesian‐regularized Genetic Neural Networks for Modeling of Acetylcholinesterase Inhibition by Huprines

Acetylcholinesterase inhibition was modeled for a set of huprines using ensembles of Bayesian-regularized Genetic Neural Networks. In the Bayesian-regularized Genetic Neural Network approach the Bayesian regularization avoids overfitted regressions and the genetic algorithm allows exploring a wide pool of three-dimensional descriptors. The predictive capacity of our selected model was evaluated by averaging multiple validation sets generated as members of neural network ensembles. When 60 members are assembled, the neural network ensemble provides a reliable measure of training and test set R(2)-values of 0.945 and 0.850 respectively. In other respects, the ability of the nonlinear selected genetic algorithm space for differentiate the data were evidenced when total data set was well distributed in a Kohonen self-organizing map. The analysis of the self-organizing map zones allows establishing the main structural features differentiated by our vectorial space.

Amino Acid Sequence Autocorrelation Vectors and Ensembles of Bayesian-Regularized Genetic Neural Networks for Prediction of Conformational Stability of Human Lysozyme Mutants

Development of novel computational approaches for modeling protein properties from their primary structure is a main goal in applied proteomics. In this work, we reported the extension of the autocorrelation vector formalism to amino acid sequences for encoding protein structural information with modeling purposes. Amino Acid Sequence Autocorrelation (AASA) vectors were calculated by measuring the autocorrelations at sequence lags ranging from 1 to 15 on the protein primary structure of 48 amino acid/residue properties selected from the AAindex database. A total of 720 AASA descriptors were tested for building predictive models of the thermal unfolding Gibbs free energy change of human lysozyme mutants. In this sense, ensembles of Bayesian-Regularized Genetic Neural Networks (BRGNNs) were used for obtaining an optimum nonlinear model for the conformational stability. The ensemble predictor described about 88% and 68% variance of the data in training and test sets, respectively. Furthermore, the optimum AASA vector subset was shown not only to successfully model unfolding thermal stability but also to distribute wild-type and mutant lysozymes on a stability Self-organized Map (SOM) when used for unsupervised training of competitive neurons.

Bayesian-regularized genetic neural networks applied to the modeling of non-peptide antagonists for the human luteinizing hormone-releasing hormone receptor

Bayesian-regularized genetic neural networks (BRGNNs) were used to model the binding affinity (IC 50) for 128 non-peptide antagonists for the human luteinizing hormone-releasing hormone (LHRH) receptor using 2D spatial autocorrelation vectors. As a preliminary step, a linear dependence was established by multiple linear regression (MLR) approach, selecting the relevant descriptors by genetic algorithm (GA) feature selection. The linear model showed to fit the training set ( N = 102) with R 2 = 0.746, meanwhile BRGNN exhibited a higher value of R 2 = 0.871. Beyond the improvement of training set fitting, the BRGNN model overcame the linear one by being able to describe 85% of test set ( N = 26) variance in comparison with 73% the MLR model. Our non-linear QSAR model illustrates the importance of an adequate distribution of atomic properties represented in topological frames and reveals the electronegativities, masses and polarizabilities as the most influencing atomic properties in the structures of the heterocycles under analysis for having an appropriate LHRH antagonistic activity. Furthermore, the ability of the non-linear selected variables for differentiating the data was evidenced when total data set was well distributed in a Kohonen self-organizing map (SOM).

2D Autocorrelation modeling of the negative inotropic activity of calcium entry blockers using Bayesian-regularized genetic neural networks

Negative inotropic potency of 60 benzothiazepine-like calcium entry blockers (CEBs), Diltiazem analogs, was successfully modeled using Bayesian-regularized genetic neural networks (BRGNNs) and 2D autocorrelation vectors. This approach yielded reliable and robust models whilst by means of a linear genetic algorithm (GA) search routine no multilinear regression model was found describing more than 50% of the training set. On the contrary, the optimum neural network predictor with five inputs described about 84% and 65% variances of 50 randomly selected training and test sets. Autocorrelation vectors in the nonlinear model contained information regarding 2D spatial distributions on the CEB structure of van der Waals volumes, electronegativities, and polarizabilities. However, a sensitivity analysis of the network inputs pointed out to the electronegativity and polarizability 2D topological distributions at substructural fragments of sizes 3 and 4 as the most relevant features governing the nonlinear modeling of the negative inotropic potency.

Modeling of acetylcholinesterase inhibition by tacrine analogues using Bayesian-regularized Genetic Neural Networks and ensemble averaging

Acetylcholinesterase inhibition was modeled for a set of 136 tacrine analogues using Bayesian-regularized Genetic Neural Networks (BRGNNs). In the BRGNN approach the Bayesian-regularization avoids overtraining/overfitting and the genetic algorithm (GA) allows exploring an ample pool of 3D-descriptors. The predictive capacity of our selected model was evaluated by averaging multiple validation sets generated as members of diverse-training set neural network ensembles (NNEs). The ensemble averaging provides reliable statistics. When 40 members are assembled, the NNE provides a reliable measure of training and test set R values of 0.921 and 0.851 respectively. In other respects, the ability of the nonlinear selected GA space for differentiating the data was evidenced when the total data set was well distributed in a Kohonen Self-Organizing Map (SOM). The location of the inhibitors in the map facilitates the analysis of the connection between compounds and serves as a useful tool for qualitative predictions.