Publications

  • [Show abstract] [Hide abstract]
    ABSTRACT: Quantitative Structure-Activity Relationships (QSARs) and Quantitative Structure-Property Relationships (QSPRs) are mathematical models used to describe and predict a particular activity/property of compounds. On the other hand, the Artificial Neural Network (ANN) is a tool that emulates the human brain to solve very complex problems. The exponential need for new compounds in the drug industry requires alternatives for experimental methods to decrease development time and costs. This is where chemical computational methods have a great relevance, especially QSAR/QSPR-ANN. This chapter shows the importance of QSAR/QSPR-ANN and provides examples of its use.
    Methods in molecular biology (Clifton, N.J.) 01/2015; 1260:319-33. · 1.29 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Abstract Quantitative Structure-Activity Relationship (QSAR) models for binding affinity constants (log Ki) of 78 flavonoid ligands towards the benzodiazepine site of GABA (A) receptor complex were calculated using the machine learning methods: artificial neural network (ANN) and support vector machine (SVM) techniques. The models obtained were compared with those obtained using multiple linear regression (MLR) analysis. The descriptor selection and model building were performed with 10-fold cross-validation using the training data set. The SVM and MLR coefficient of determination values are 0.944 and 0.879, respectively, for the training set and are higher than those of ANN models. Though the SVM model shows improvement of training set fitting, the ANN model was superior to SVM and MLR in predicting the test set. Randomization test is employed to check the suitability of the models.
    Journal of Enzyme Inhibition and Medicinal Chemistry 10/2013; · 1.50 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Vaptans are compounds that act as non-peptide vasopressin receptor antagonists. These compounds have diverse chemical structures. In this study, we used a combined approach of protein folding, molecular dynamics simulations, docking and Quantitative structure-activity relationship (QSAR) to elucidate the detailed interaction of the vasopressin receptor V1a (V1aR) with some of its blockers (134). QSAR studies were performed using MLR analysis and were gathered into one group to perform an Artificial Neural Network (ANN) analysis. For each molecule, 1481 molecular descriptors were calculated. Additionally, 15 quantum chemical descriptors were calculated. The final equation was developed by choosing the optimal combination of descriptors after removing the outliers. Molecular modeling enabled us to obtain a reliable tridimensional model of V1aR. The docking results indicated that the great majority of ligands reach the binding site under π-π, π-cation and hydrophobic interactions. The QSAR studies demonstrated that the heteroatoms N and O are important for ligand recognition, which could explain the structural diversity of ligands that reach V1aR This article is protected by copyright. All rights reserved.
    Chemical Biology &amp Drug Design 09/2013; · 2.47 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Quantitative Structure-Activity Relationship (QSAR) modeling was performed for imidazo[1,5-a]pyrido[3,2-e]pyrazines, which constitute a class of phosphodiesterase 10A inhibitors. Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) were used as feature selection techniques to find the most reliable molecular descriptors from a large pool. Modeling of the relationship between the selected descriptors and the pIC50 activity data was achieved by linear (Multiple Linear Regression; MLR) and nonlinear [Locally Weighted Regression (LWR) based on both Euclidean (E) and Mahalanobis (M) distances] methods. In addition, a stepwise MLR model was built using only a limited number of quantum chemical descriptors, selected because of their correlation with the pIC50 . The model was not found interesting. It was concluded that the LWR model, based on the Euclidean distance, applied on the descriptors selected by PSO has the best prediction ability. However, some other models behaved similarly. The root mean squared errors of prediction (RMSEP) for the test sets obtained by PSO/MLR, GA/MLR, the PSO/LWRE, PSO/LWRM, GA/LWRE and GA/LWRM models were 0.333, 0.394, 0.313, 0.333, 0.421 and 0.424, respectively. The PSO-selected descriptors resulted in the best prediction models, both the linear and nonlinear. This article is protected by copyright. All rights reserved.
    Chemical Biology &amp Drug Design 08/2013; · 2.47 Impact Factor
  • Omar Deeb, Mohammad Goodarzi
    [Show abstract] [Hide abstract]
    ABSTRACT: Undesirable toxicity is still a major block in the drug discovery process. Obviously, capable techniques that identify poor effects at a very early stage of product development and provide reasonable toxicity estimates for the huge number of untested compounds are needed. In silico techniques are very useful for this purpose, because of their advantage in reducing time and cost. These case studies give the description of in silico validation techniques and applied modeling methods for the prediction of toxicity of chemical compounds. In silico toxicity prediction techniques can be classified into two categories: Molecular Modeling and methods that derive predictions from experimental data. Molecular modeling is a computational approach to mimic the behavior of molecules, from small molecules (e.g. in conformational analysis) to biomolecules. But the same approaches can also be applied for toxicological purposes, if the mechanism is receptor mediated. Quantitative Structure-Toxicity Relationships (QSTRs) models are typical examples for the prediction of toxicity which relates variations in the molecular structures to toxicity. There are many applied modeling techniques in QSTR such as Partial Least Squares, Artificial Neural Networks, and Principal Component Regression (PCR). The applicability of these techniques in predictive toxicology will be discussed with different examples of sets of chemical compounds.
    Current drug safety. 10/2012;
  • [Show abstract] [Hide abstract]
    ABSTRACT: Quantitative structure-activity relationship study was performed to understand the inhibitory activity of a set of 192 vascular endothelial growth factor receptor-2 (VEGFR-2) compounds. QSAR models were developed using multiple linear regression (MLR) and partial least squares (PLS) as linear methods. While principal component - artificial neural networks (PC-ANN) modeling method with application of eigenvalue ranking factor selection procedure was used as nonlinear method. The results obtained offer good regression models having good prediction ability. The results obtained by MLR and PLS are close and better than those obtained by principal component-artificial neural network. The best model was obtained with a correlation coefficient of 0.87. The strength and the predictive performance of the proposed models was verified using both internal (cross-validation and Y-scrambling) and external statistical validations.
    Current pharmaceutical design 09/2012; · 4.41 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Linear and nonlinear quantitative structure activity relationship models for predicting the inhibitory activities of sulfonamides toward different carbonic anhydrase isozymes were developed based on multilinear regression, principal component-artificial neural network and correlation ranking-principal component analysis, to identify a set of structurally based numerical descriptors. Multilinear regression was used to build linear quantitative structure activity relationship models using 53 compounds with their quantum chemical descriptors. For each type of isozyme, separate quantitative structure activity relationship models were obtained. It was found that the hydration energy plays a significant role in the binding of ligands to the CAI isozyme, whereas the presence of five-membered ring was detected as a major factor for the binding to the CAII isozyme. It was also found that the softness exhibited significant effect on the binding to CAIV isozyme. Principal component-artificial neural network and correlation ranking-principal component analysis analyses provide models with better prediction capability for the three types of the carbonic anhydrase isozyme inhibitory activity than those obtained by multilinear regression analysis. The best models, with improved prediction capability, were obtained for the hCAII isozyme activity. Models predictivity was evaluated by cross-validation, using an external test set and chance correlation test.
    Chemical Biology &amp Drug Design 12/2011; 79(4):514-22. · 2.47 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: A QSAR study of antiamoebic agents isolated from natural products was performed by multi linear regression (MLR), artificial neuron network (ANN), and regression through origin (RTO). After several procedures to reduce the number of descriptors, 11 descriptors were selected from the descriptor pool by a complete MLR methodology. The best proportion between training:predicted:validation sets is 100:43:16 molecules. The Mor23m descriptor is a 3D-MoRSE descriptor and it is the main descriptor in the models studied. This result suggests that the three-dimensional structure and atomic properties like masses are very important in the models. The best quantitative structure–activity relationship model was proved to be independent of chance correlation.
    Medicinal Chemistry Research 09/2011; 21(9). · 1.61 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Several non-peptide heterocyclic compounds reported as potent thrombin inhibitors in vitro were chosen to carry out a QSAR study upon them using MLR and ANN analysis. In order to identify the best QSAR models, the input for ANN consisted of those subsets of descriptors used in the MLR models. The best QSAR models contained the SIC₀ descriptor as the main topological descriptor. To identify the physical and chemical properties involved in the ligand-thrombin complexes, an automated ligand-flexible docking procedure was used. The docking results suggest that the thrombin inhibition by these heterocyclic compounds is driven by π-π, hydrogen bonds and salt bridge interactions. The best Gibbs free energy of ligand binding was found at the thrombin sites S1 and D. We have shown that it is possible to build MLR models with geometries taken from two different sources (semi-empirical and MD geometries) and obtain similar results.
    Journal of Enzyme Inhibition and Medicinal Chemistry 06/2011; 27(2):174-86. · 1.50 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Melting point is a basic physical property that specifies the transition temperature between solid and liquid phases. Melting point has numerous applications in biochemical and environmental sciences due to its relationship with solubility. Sufficient aqueous solubility is essential for a compound to be transferred to the site of action within an organism. In spite of the huge number of available melting point data, few useful guidelines exist for understanding the relationship between the compound melting point and its chemical structure. Therefore, methods for estimating the melting point of organic compounds would considerably help medicinal chemists in designing new drugs within a specified range of melting point and solubility. A highly effective tool depending on quantitative structure–property relationship (QSPR) can be utilized to predict melting point for drug-like compounds with no literature values. QSPR models were developed using genetic algorithms, multiple linear regression and neural networks analyses. Predictive non-linear QSPR models were developed for the relevant descriptors. The results obtained offers excellent regression models that possesses good prediction ability.
    Molecular Physics 01/2011; 109(4):507-516. · 1.67 Impact Factor
  • Source
    Omar Deeb, Mojahed Drabh
    [Show abstract] [Hide abstract]
    ABSTRACT: Quantitative structure-activity relationship study was performed to understand analgesic activity for a set of 95 heterogeneous analgesic compounds. This study was performed by using the principal component-artificial neural network modeling method, with application of eigenvalue ranking factor selection procedure. The results obtained by principal component-artificial neural network give advanced regression models with good prediction ability using a relatively low number of principal components. A 0.834 correlation coefficient was obtained using principal component-artificial neural network with six extracted principal components.
    Chemical Biology &amp Drug Design 09/2010; 76(3):255-62. · 2.47 Impact Factor
  • Source
    Omar Deeb, Mohammad Goodarzi
    [Show abstract] [Hide abstract]
    ABSTRACT: The support vector machine (SVM) and partial least square (PLS) methods were used to develop quantitative structure activity relationship (QSAR) models to predict the inhibitory activity of non-peptide HIV-1 protease inhibitors. Genetic algorithm (GA) was employed to select variables that lead to the best-fitted models. A comparison between the obtained results using SVM with those of PLS revealed that the SVM model is much better than that of PLS. The root mean square errors of the training set and the test set for SVM model were calculated to be 0.2027, 0.2751, and the coefficients of determination (R(2)) are 0.9800, 0.9355 respectively. Furthermore, the obtained statistical parameter of leave-one-out cross-validation test (Q(2)) on SVM model was 0.9672, which proves the reliability of this model. The results suggest that TE2, Ui, GATS5e, Mor13e, ATS7m, Ss, Mor27e, and RDF035e are the main independent factors contributing to the inhibitory activities of the studied compounds.
    Chemical Biology &amp Drug Design 05/2010; 75(5):506-14. · 2.47 Impact Factor
  • Omar Deeb, Mohammad Goodarzi
    [Show abstract] [Hide abstract]
    ABSTRACT: Pesticide contamination of surface water and groundwater due to agricultural activities has been of concern for a long time. Water solubility indicates the tendency of a pesticide to be removed from soil by runoff or irrigation and to reach surface water and indicates the tendency to precipitate at the soil surface. The experimental procedures determining the solubility in water of pesticides are always time-consuming and expensive, and it is difficult to accurately distinguish species with similar physicochemical properties. A highly effective tool depending on a quantitative structure-property relationship (QSPR) can be utilised to predict solubility in water for those pesticide compounds with no literature values. QSPR models were developed using multiple linear regression, partial least squares and neural networks analyses. Following the removal of a small number of outliers, linear and non-linear QSPR models to predict the solubility of pesticide compounds in water were developed for the relevant descriptors. Consistent with experimental studies, the results obtained offer excellent regression models having good prediction ability.
    Molecular Physics 01/2010; 108(2):181-192. · 1.67 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The terms bioaccumulation and bioconcentration refer to the uptake and build-up of chemicals that can occur in living organisms. Experimental measurement of bioconcentration is time-consuming and expensive, and is not feasible for a large number of chemicals of potential regulatory concern. A highly effective tool depending on a quantitative structure-property relationship (QSPR) can be utilized to describe the tendency of chemical concentration organisms represented by, the important ecotoxicological parameter, the logarithm of Bio Concentration Factor (log BCF) with molecular descriptors for a large set of non-ionic organic compounds. QSPR models were developed using multiple linear regression, partial least squares and neural networks analyses. Linear and non-linear QSPR models to predict log BCF of the compounds developed for the relevant descriptors. The results obtained offer good regression models having good prediction ability. The descriptors used in these models depend on the volume, connectivity, molar refractivity, surface tension and the presence of atoms accepting H-bonds.
    Environmental Health Insights 01/2010; 4:33-47.
  • Omar Deeb
    [Show abstract] [Hide abstract]
    ABSTRACT: The performance of the two novel QSAR algorithms, stepwise regression (SR) and correlation ranking (CR) principal component-artificial neural network modeling methods named (SR-PC-ANN and CR-PC-ANN, respectively), combined with two factor selection approaches is compared. These algorithms are applied to predict the toxic activity of 278 substituted benzenes belonging to diverse types of compounds as well as human serum albumin (HSA) binding affinity of a set of 95 compounds. A total number of 1233 and 698 theoretical descriptors are calculated for each molecule for the toxicity and binding affinity data sets, respectively.In PCA analysis, the principal components were extracted by applying two different approaches named the individual PCA approach, PCA(I), and the combined PCA approach, PCA(C). In the PCA(I) approach, descriptors were gathered into groups and then the PCA analysis was performed on each group. In the combined PCA approach, PCA(C), PCA analysis was applied on the pool of all calculated descriptors combined in one group. One is able to see that the models obtained by the CR-PC-ANN(I) approach are superior to those obtained from the other approaches. Both the external and cross-validation methods were used to validate the performances of the resulting models. Randomization test was employed to check the suitability of the models and to investigate the possibility of obtaining chance models.
    Chemometrics and Intelligent Laboratory Systems. 01/2010;
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Five-nanosecond molecular dynamics (MD) simulations were performed on human serum albumin (HSA) to study the conformational features of its primary ligand binding sites (I and II). Additionally, 11 HSA snapshots were extracted every 0.5 ns to explore the binding affinity (K(d)) of 94 known HSA binding drugs using a blind docking procedure. MD simulations indicate that there is considerable flexibility for the protein, including the known sites I and II. Movements at HSA sites I and II were evidenced by structural analyses and docking simulations. The latter enabled the study and analysis of the HSA-ligand interactions of warfarin and ketoprofen (ligands binding to sites I and II, respectively) in greater detail. Our results indicate that the free energy values by docking (K(d) observed) depend upon the conformations of both HSA and the ligand. The 94 HSA-ligand binding K(d) values, obtained by the docking procedure, were subjected to a quantitative structure-activity relationship (QSAR) study by multiple regression analysis. The best correlation between the observed and QSAR theoretical (K(d) predicted) data was displayed at 2.5 ns. This study provides evidence that HSA binding sites I and II interact specifically with a variety of compounds through conformational adjustments of the protein structure in conjunction with ligand conformational adaptation to these sites. These results serve to explain the high ligand-promiscuity of HSA.
    Biopolymers 09/2009; 93(2):161-70. · 2.88 Impact Factor
  • Omar Deeb, Brian W Clare
    [Show abstract] [Hide abstract]
    ABSTRACT: QSAR analysis of a set of 96 heterocyclics with antifungal activity was performed. The results reveals that a pyridine ring is more favorable than benzene as the 6-membered ring, for high activity, but thiazole is unfavorable as the 5-membered ring relative to imidazole or oxazole. Methylene is the spacer leading to the highest activity. The descriptors used are indicator variables, which account for identity of substituent, lipophilicity and volume of substituent, and total polarizability. Unlike previously reported results for this data set, our fits do not exceed the limitations set by the nature of the data itself.
    Journal of Computer-Aided Molecular Design 07/2008; 22(12):885-95. · 3.17 Impact Factor
  • Omar Deeb, Brian W Clare
    [Show abstract] [Hide abstract]
    ABSTRACT: The flip regression procedure that we used earlier for handling the xanthones system has been applied to phenylaminoquinazoline analogues. It is known that the substituents at the 6- and 7- positions of the polycyclic system have been identified as the most important structural features. The steric as well as the electrostatic interactions proved to be the most important for the inhibitory effect. In this contribution it is shown that the orientation of nodes in their occupied pi orbitals, and also the energies of these orbitals explains a further large portion of the variance in their inhibitory activity.
    Journal of Enzyme Inhibition and Medicinal Chemistry 07/2008; 23(6):763-75. · 1.50 Impact Factor
  • Omar Deeb, Brian W Clare
    [Show abstract] [Hide abstract]
    ABSTRACT: The aim of this study is to provide an initial indication regarding the scope and limitations of some state-of-the-art methods in computational chemistry, including semiempirical (AM1) and density functional theory (B3LYP), in the flip regression procedure applied to the inhibition of phenylisopropylamines. The results show that the models established based on the density functional theory-B3LYP are better than that based on semiempirical method (AM1). It is demonstrated that electron-rich ring systems and highest occupied molecular orbital levels tended to increase activity. In this contribution, it is shown that the orientation of nodes in their occupied pi orbitals, and also the energies of these orbitals explain a further large portion of the variance in their inhibitory activity.
    Chemical Biology &amp Drug Design 05/2008; 71(4):352-62. · 2.47 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Quantitative structure-activity relationship studies were performed to describe and predict the mutagenic activity of a set of 48 nitrated polycyclic aromatic hydrocarbons. From a larger pool of molecular descriptors (topological indices) we arrived at much a smaller set consisting of three correlating parameters. Such a variable selection is made using ncss software in that successive regressions were attempted using maximum-R(2) method. The results are critically discussed using a variety of statistical parameters. Our results have shown that connectivity and shape type indices together with the distance-based Wiener index (W) play a dominating role in modelling of mutagenicity (logTA100). The predictive ability of the models is discussed on the basis of cross-validated parameters.
    Chemical Biology &amp Drug Design 04/2008; 71(3):230-43. · 2.47 Impact Factor

18 Following View all

76 Followers View all