QSAR analysis on a large and diverse set of potent phosphoinositide 3-kinase gamma (PI3Kγ) inhibitors using MLR and ANN methods (original) (raw)
Related papers
Journal of Enzyme Inhibition and Medicinal Chemistry, 2008
Quantitative structure-activity relationship (QSAR) studies have been carried out on indolyl aryl sulfones, a class of novel HIV-1 non-nucleoside reverse transcriptase inhibitors, using physicochemical, topological and structural parameters along with appropriate indicator variables. The statistical tools used were linear methods (e.g., stepwise regression analysis, partial least squares (PLS), factor analysis followed by multiple regression (FA-MLR), genetic function approximation combined with multiple linear regression (GFA-MLR) and GFA followed by PLS or G/PLS and nonlinear method (artificial neural network or ANN). In case of physicochemical parameters, GFA-MLR generated the best Equation (n ¼ 97, R 2 ¼ 0.862, Q 2 ¼ 0.821). Using topological parameters, the best Equation (based on leave-one-out Q 2) was obtained with stepwise regression technique (n ¼ 97, R 2 ¼ 0.867, Q 2 ¼ 0.811). When topological and physicochemical parameters were used in combination, statistical quality increased to a great extent (n ¼ 97, R 2 ¼ 0.891, Q 2 ¼ 0.849 from stepwise regression). Furthermore, the whole dataset had been divided into test (25% of whole dataset) and training (remaining 75%) sets. Models were developed based on the training set and predictive potential of such models was checked from the test set. The selection of the training set was based on K-means clustering of the standardized descriptors (topological and physicochemical). In this case also the best results were obtained with stepwise regression (n ¼ 72, R 2 ¼ 0.906, Q 2 ¼ 0.853) but external predictive capacity of this model (R 2 pred ¼ 0:738) was inferior to the model developed from GFA-MLR technique (R 2 ¼ 0.883, Q 2 ¼ 0.823, R 2 pred ¼ 0:760). However, the squared regression coefficient between observed activity and predicted activity values of the test set compounds for the best linear model, i.e., GFA-MLR (r 2 ¼ 0.736) was lower in comparison to the best nonlinear model developed using artificial neural network (r 2 ¼ 0.781). Thus, based on external validation, the ANN models were superior to the linear models. The predictive potential of the best linear Equation (stepwise regression model) was superior to that of the previously published CoMFA (Q 2 ¼ 0.81, SDEP Test ¼ 0.89) on the same data set (Ragno R. et al., J Med Chem 2006, 49, 3172-3184). Furthermore, the physicochemical parameter based models also supported the previous observations based on docking (Ragno R. et al.,
ANN-QSAR model for selection of anticancer leads from structurally heterogeneous series of compounds
European Journal of Medicinal Chemistry, 2007
Developing a model for predicting anticancer activity of any classes of organic compounds based on molecular structure is very important goal for medicinal chemist. Different molecular descriptors can be used to solve this problem. Stochastic molecular descriptors so-called the MARCH-INSIDE approach, shown to be very successful in drug design. Nevertheless, the structural diversity of compounds is so vast that we may need non-linear models such as artificial neural networks (ANN) instead of linear ones. SmartMLP-ANN analysis used to model the anticancer activity of organic compounds has shown high average accuracy of 93.79% (train performance) and predictability of 90.88% (validation performance) for the 8:3-MLP topology with different training and predicting series. This ANN model favourably compares with respect to a previous linear discriminant analysis (LDA) model [H. González-Díaz et al., J. Mol. Model 9 (2003) 395] that showed only 80.49% of accuracy and 79.34% of predictability. The present SmartMLP approach employed shorter training times of only 10 h while previous models give accuracies of 70–89% only after 25–46 h of training. In order to illustrate the practical use of the model in bioorganic medicinal chemistry, we report the in silico prediction, and in vitro evaluation of six new synthetic tegafur analogues having IC50 values in a broad range between 37.1 and 138 μg mL−1 for leukemia (L1210/0) and human T-lymphocyte (Molt4/C8, CEM/0) cells. Theoretical predictions coincide very well with experimental results.
QSAR modeling of PTP1B inhibitor by using Genetic algorithm-Neural network methods
Journal of Physics: Conference Series
Type-2 diabetes mellitus is an epidemic disease that is characterized by the chronic increase of glucose level. The insulin hormone is known to correspond to this disease, while PTP1B involved in the regulation of this hormone. Hence, PTP1B has become the primary target of drug development to treat this disease. In this study, we aim to develop QSAR model to predict PTP1B inhibitor by using a neural network method. Genetic algorithm (GA) method was used to select the set of the molecular descriptor. We improved the performance of the models by performing a hyperparameter tuning procedure. From the results of validation analysis, we found that the model 2 containing 5 descriptors as the best model. This confirms by the value of MCC (0.68) and AUC (0.89) of this model that is higher than the others. Also, the additional y-scrambled analysis confirms that this model does not correspond to coincidental correlation, indicated by a very low value of MCC of the scrambled model.
European Journal of Medicinal Chemistry, 2014
Phosphoinositide 3-kinase gamma (PI3Kg) is member of a family of enzymes involved in cancer pathogenesis. Accordingly, considerable efforts have been carried out to develop new PI3Kg inhibitors. Towards this end we explored the pharmacophoric space of PI3Kg using three diverse sets of inhibitors. Subsequently, we employed genetic algorithm-based QSAR analysis to select optimal combination of pharmacophoric models and physicochemical descriptors that can explain bioactivity variation within training inhibitors. Interestingly, two successful pharmacophores were selected within two statistically consistent QSAR models. The close similarity among the two binding models prompted us to merge them in a hybrid pharmacophore. The resulting model showed superior receiver operator characteristic curve (ROC) and closely resembled binding interactions seen in crystallographic ligandePI3Kg complexes. The resulting model was employed to screen the national cancer institute (NCI) list of compounds to search for new PI3Kg ligands. After testing captured hits in vitro, 19 compounds showed nanomolar IC 50 values against PI3Kg. The chemical structures and purities of most potent hits were validated using NMR and MS experiments.
Phosphoinositide 3-kinase gamma (PI3Kg) is member of a family of enzymes involved in cancer path-ogenesis. Accordingly, considerable efforts have been carried out to develop new PI3Kg inhibitors. Towards this end we explored the pharmacophoric space of PI3Kg using three diverse sets of inhibitors. Subsequently, we employed genetic algorithm-based QSAR analysis to select optimal combination of pharmacophoric models and physicochemical descriptors that can explain bioactivity variation within training inhibitors. Interestingly, two successful pharmacophores were selected within two statistically consistent QSAR models. The close similarity among the two binding models prompted us to merge them in a hybrid pharmacophore. The resulting model showed superior receiver operator characteristic curve (ROC) and closely resembled binding interactions seen in crystallographic ligandePI3Kg complexes. The resulting model was employed to screen the national cancer institute (NCI) list of compounds to search for new PI3Kg ligands. After testing captured hits in vitro, 19 compounds showed nanomolar IC 50 values against PI3Kg. The chemical structures and purities of most potent hits were validated using NMR and MS experiments.
3D-QSAR models to predict anti-cancer activity on a series of protein P38 MAP kinase inhibitors
Journal of Taibah University for Science, 2017
Protein kinases are essential components of various signaling pathways and represent attractive targets for therapeutic interventions. Kinase inhibitors are currently used to treat malignant tumors, as well as autoimmune diseases, due to their involvement in immune cell signaling. In this study, three-dimensional quantitative structure-activity relationship (3D-QSAR) analyses, including Multiple Linear Regression (MLR), Partial Least Squares (PLS), Multiple Non-Linear Regression (MNLR), Artificial Neural Network (ANN) and cross-validation analyses, were performed on a set of P38 MAP kinases as anti-cancer agents. This method, which is based on molecular modeling (molecular mechanics, Hartree-Fock (HF)), was used to determine the structural parameters, electronic properties, and energy associated with the molecules we examined. MLR, PLS, and MNLR analyses were performed on 46 protein P38 MAP kinase analogs to determine the relationships between molecular descriptors and the anti-cancer properties of the P38 MAP kinase analogs. The MLR model was validated by the external validation and standardization approach. The ANN, given the descriptors obtained from the MLR, exhibited a correlation coefficient close to 0.94. The predicted model was confirmed by two methods, leave-one-out (LOO) cross-validation and scrambling (or Y-randomization). We observed a high correlation between predicted and experimental activity, thereby both validating and demonstrating the high quality of the QSAR model that we described.
Arkivoc, 2007
Artificial neural networks (ANNs) can be utilized to generate predictive models of quantitative structure-activity relationships (QSAR) between a set of molecular descriptors and activity. In the present work, QSAR analysis for a set of 95 1-[(2-hydroxyethoxy)-methyl]-6-(phenylthio)thymine (HEPT) derivatives has been investigated by means of a three-layered neural network (NN). It has been shown that NN can be a potential tool in the investigation of QSAR analysis compared with the models given in the literature. The results obtained by using the NN adopted for QSAR models showing not only good statistical significance in fitting, but also high predictive ability. (0.916< r <0.968 and q 2 = 0.8779). The relevant factors controlling the anti-HIV-1 activity of HEPT derivatives have been identified. The results are along the same lines as those of our previous studies on HEPT derivatives and indicate the importance of the hydrophobic parameter in modelling the QSAR for HEPT derivatives
Medicinal Chemistry Research, 2013
The mTOR-mediated PI3K/AKT/mTOR signal transduction pathway plays a key role in a broad spectrum of cancers. In the present article, QSAR and pharmacophore studies were carried out using a series of 61 benzothiazole class of PI3Ka inhibitors to characterize molecular features and structural requirements crucial for biological interaction. QSAR study performed using TSAR 3.3 by multiple regression analysis and partial least square methods identified inertia moment-1-size, kier chiv5 (path) index, and number of H-bond donors as important descriptors responsible for PI3Ka inhibitory activity. Further analysis of pharmacophore model by means of Phase module of Schrodinger revealed that two hydrogen-bond acceptors, one hydrogen-bond donors, and two hydrophobic aromatic rings as crucial molecular features that predict binding affinity for high-affinity ligands to the PI3Ka enzyme. These observations provide important insights to the key structural requirements of these molecules for potent PI3Ka inhibition. Excellent statistical results of developed models strongly suggest that these models are reasonable for the prediction of the activity of new inhibitors and in future drug design. Keywords Anticancer Á N-Acetyl 2-aminobenzothiazole Á Pharmacophore model Á QSAR Á PI3Ka Á Multiple regression analysis Á Partial least square analysis Abbreviations QSAR Quantitative structure-activity relationship PI3Ka Phosphoinositide-3-kinase-alpha mTOR Mammalian target of rapamycin PIKKs Phosphatidylinositol-3-kinase-related kinases MRA Multiple regression analysis PLS Partial least square VIF Variance inflation factor LMO Leave-many-out cv Coefficient of variance RMSD Relative mean square deviation RMSE Root mean square error Electronic supplementary material The online version of this article (
Network Modeling Analysis in Health Informatics and Bioinformatics, 2020
A combination of ligand-receptor interactions and drug-like indexes have been used to develop a quantitative structure-activity relationship model to predict anti-HIV activity (pEC 50) of 73 azine derivatives as non-nucleoside reverse transcriptase inhibitors. Ligand-receptor interactions were derived from the best position (best pose) of studied compounds, as ligands, in the active site of receptors using Autodock 4.2 software and named as molecular docking descriptors. The drug-like indexes were calculated using DRAGON 5.5 software. Two groups of descriptors were mixed, and the stepwise regression method was used for the selection of the most relevant descriptors. Four selected descriptors were subsequently used to construct the quantitative structure-activity relationship model using the Levenberg-Marquardt artificial neural network method. Dataset was randomly divided into the train (53 compounds), validation (10 compounds) and test set (10 compounds). The best model was selected according to the lowest mean square error value of the validation set. The accuracy and predictability of the model were evaluated using test and validation sets and the leave-one-out technique. According to the predicted results, the coefficient of determination of the test set (R 2 = 0.86) and all data (Q 2 LOO = 0.73) were acceptable. The mean square error value for the test set was equal to 0.11. The obtained results emphasized the good prediction ability and generalizability of the developed model in the prediction of pEC 50 values for new compounds.
Journal of King Saud University - Science, 2018
Cancer-causing nature is one of the toxicological endpoints bringing about the most elevated concern. Likewise, the standard bioassays in rodents used to survey the cancer-mitigating capability of chemicals and medications are expensive and require the sacrifice of animals. Thus, we have endeavored the development of a worldwide QSAR model utilizing an information set of 85 compounds, including drugs for their anti-leukemia potential. Considering expansive number of information focuses with different structural elements utilized for model development (ntraining = 68) and model validation (ntest = 17), the model developed in this study has an encouraging statistical quality (leave-one-out Q2 = 0.833, R2pred = 0.716) for pLC50 and (leave-one-out Q2 = 0.744, R2pred = 0.614) for pGI50. Our developed model suggests that the absence of methanal fragments, low dipole moment and presence of some 2D autocorrelated molecular descriptors reduces the carcinogenicity. Branching, size and shape are found to be crucial factors for drug-mitigating carcinogenicity.