QSRR Model Research Articles

Development of QSRR model of a set of Polycyclic Aromatic Hydrocarbons (PAHs) using simple regression analysis in silico

A structure/retention indices relationship was searched for 59 PAHs while promoting the simple linear regression by genetic algorithm MOBIDYGS software, the structural parameters being calculated with the software Spartan and DRAGON. Among about a hundred of one-regression models gotten, we selected the one that present best values of the prediction parameter (Q2) and of the determination coefficient (R2). The robustness of obtaining model were illustrated using different techniques: leave-many-out, external-validation, randomization test, applicability domain.

Extract identification and evaluation of the cytotoxic activity of Polygala fallax Hemsl in Heilongjiang ethnic medicine against tumors.

Heilongjiang Province is a frontier province with distinctive characteristics, fertile land and rich products. This study provides a new method for qualitatively studying flavonoids in traditional Chinese medicine and a new auxiliary means for identifying flavonoid isomers. The flavonoids in Polygala fallax Hemsl were identified by ultra-performance liquid chromatography-photo-diode array (PDA)-quadrupole-electro- static field orbitrap mass spectrometry tandem by UV Spectrum, primary and secondary high-resolution mass spectrometry (MS1/MS2) cleavage of fragments combined with databases, mass spectrometry cleavage patterns and literature. The established QSRR model was used to verify the flavonoids identified from the Polygala fallax Hemsl. The structure of multiple Polygala fallax Hemsl has been identified using various spectral methods. The tumor cytotoxic activity of the isolated compounds was evaluated. This paper is of great significance for further elucidating the pharmacodynamic substance basis and further developing and utilizing Polygala fallax Hemsl.

Application of liquid chromatography in defining the interaction of newly synthesized chalcones and related compounds with human serum albumin

Defining the interaction of newly synthesized compounds with plasma proteins is an important step in the drug development process. Chromatographic techniques can be successfully used in predicting the biopharmaceutical and pharmacokinetic properties of newly synthesized compounds. The aim of this study is to investigate and isolate the most important molecular properties that affect the interaction of 20 newly synthesized chalcones and commercial compounds (lopinavir, ritonavir, darunavir and ivermectin) with human serum albumin (HSA). The retention behaviour of the selected compounds was tested on a CHIRALPAK?HSA column. A mixture of phosphate buffer (pH 7.0) and isopropanol (80:20 volume ratio) was used as the mobile phase, and the support vector method was used to form the quantitative structure retention relationship (QSRR) model. Based on the obtained values of retention parameters, it was observed that halogenated derivatives show the strongest, and methylated chalcone derivatives the weakest interaction with HSA. By correlating the retention and physicochemical properties of the tested compounds, it was shown that the structural (SDSCH) and electronic properties (MAXQ, EEM_F1) groups have the greatest influence on the retention behaviour and the interaction of the tested compounds with HSA. The obtained QSRR model can be applied in the prediction of the retention characteristics of new, structurally related chalcone derivatives on HSA stationary phase.

Optimization of chromatographic separation of aripiprazole and impurities: Quantitative structure-retention relationship approach

A new optimization strategy based on the mixed quantitative structure? retention relationship (QSRR) model is proposed for improving the RPHPLC separation of aripiprazole and its impurities (IMP A-E). Firstly, experimental parameters (EPs), namely mobile phase composition and flow rate, were varied according to Box?Behnken design and thereafter, an artificial neural network (ANN) as a QSRR model was built correlating EPs and selected molecular descriptors (ovality, torsion energy and non-1,4-van der Waals energy) with the log-transformed retention times of the analytes. Values of the root mean square error (RMSE) were used for an estimation of the quality of the ANNs (0.0227, 0.0191 and 0.0230 for the training, verification and test set, respectively). The separations of critical peak pairs on chromatogram (IMP AB and IMP D-C) were optimized using ANNs for which the EPs served as inputs and the log-transformed separation criteria s as the outputs. They were validated by application of leave-one-out cross-validation (RMSE values 0.065 and 0.056, respectively). The obtained ANNs were used for plotting response surfaces upon which the analyses chromatographic conditions resulting in optimal analytes retention behaviour and the optimal values of the separation criteria s were defined. The optimal conditions were 54 % of methanol at the beginning and 79 % of methanol at the end of gradient elution programme with a mobile phase flow rate of 460 ?L min-1.

A QSRR model for predicting gas chromatography retention indices of essential oils using an improved chemical reaction optimization algorithm

A QSRR model for predicting gas chromatography retention indices of essential oils using an improved chemical reaction optimization algorithm

Achillea clypeolata Sibth. & Sm. essential oil composition and QSRR model for predicting retention indices

The aim of this study was the prediction model of retention indices of compounds from the aboveground parts of Achillea clypeolata Sibth. & Sm. essential oil, obtained by hydrodistillation and analysed by GC?MS. The quantitative structure?retention relationship analysis was applied in order to anticipate the retention time of the obtained compounds. The selection of the seven molecular descriptors was done by a genetic algorithm. The chosen descriptors were uncorrelated and were used to construct an artificial neural network. A total of 40 experimentally obtained retention indices was used to build this prediction model. The coefficient of determination for the training, testing and validation cycles were: 0.950, 0.825 and 1.000, respectively, indicating that this model could be used for prediction of retention indices for A. clypeolata, essential oil compounds.

QSRR Model for predicting retention indices of Satureja kitaibelii Wierzb. ex Heuff. essential oil composition

A prediction model of retention indices of compounds from the aboveground parts of Satureja kitaibelii essential oil, obtained by hydrodistillation and analysed by Gas Chromatography coupled with Mass Spectrometry (GC-MS), was the aim of this study. The quantitative structure–retention relationship was employed to predict the retention time using five molecular descriptors selected by a genetic algorithm. The selected descriptors were used as inputs of an artificial neural network. Total of 53 experimentally obtained retention indices (log RI) were used to build a prediction model. The selected descriptors were used as inputs of an artificial neural network model, to build a prediction time predictive quantitative structure-retention relationship model. The coefficient of determination for the training cycle was 0.962, indicating that this model could be used for prediction of retention indices for S. kitaibelii essential oil compounds.

QSRR Model for Predicting Retention Indices of Geraniol Chemotype of Thymus serpyllum Essential Oil

A total of 26 experimentally obtained retention indices on a logarithmic scale (log RI) from Thymus serpyllum essential oil were used to build a robust predictive model. The selected descriptors were used as inputs of an artificial neural network model to build a predictive quantitative structure-retention time relationship model. The coefficient of determination for the training cycle was 0.977, indicating that this model could be used for the prediction of retention indices for T. serpyllum essential oil compounds. These 26 compounds comprise about 99.8 % of the total oil, but among them only 6 compounds had the average relative concentration over 1 percent: geraniol (63.4 %), nerol (or cis-geraniol) (18.9 %), geranyl acetate (4.7 %), trans- caryophyllene (4.6 %), β-bisabolene (2.0 %) and geranial (1.2 %). According to these results, it can be concluded that T. serpyllum from village Sesalac (Serbia) belongs to geraniol chemotype, in total 82.3 % in both, trans and cis forms (63.4 % and 18.9 %, respectively).

Modeling and Prediction of Gas Chromatography Relative Retention Times of Volatile Organic Compounds

The theme of this paper is to foresee relative retention time of 122 volatile organic compounds. QSRR analysis was accomplished on a serial of 122 VOCs. Multiple Linear Regression (MLR) and support vector machine (SVM) methods were used to build linear and nonlinear (QSRR) models, respectively, which correlate the (RTT) values of these chemical substance to their structural descriptors. At first, the data set was separated using Kennard and Stone algorithm into a training set (92 chemicals) and a test set (30 chemicals) for statistical external validation. The five-dimensional models were developed using as independent variables the theoretical descriptors derived from the DRAGON software during the application of the procedure GA (genetic algorithm) - VSS (Variable Subset Selection). The robustness and the predictive performance of the MLR model have been demonstrated by inside and outer statistical validation. Non-linear technique leads to the best QSRR model with good internal and external predictive abilities. It is based on support vector machines using the RBF function for the optimal parameters values. The values of and were 17, 0.2, 0.2, respectively.

Modeling and Prediction of Gas Chromatography Relative Retention Times of Volatile Organic Compounds

The theme of this paper is to foresee relative retention time of 122 volatile organic compounds. QSRR analysis was accomplished on a serial of 122 VOCs. Multiple Linear Regression (MLR) and support vector machine (SVM) methods were used to build linear and nonlinear (QSRR) models, respectively, which correlate the (RTT) values of these chemical substance to their structural descriptors. At first, the data set was separated using Kennard and Stone algorithm into a training set (92 chemicals) and a test set (30 chemicals) for statistical external validation. The five-dimensional models were developed using as independent variables the theoretical descriptors derived from the DRAGON software during the application of the procedure GA (genetic algorithm) - VSS (Variable Subset Selection). The robustness and the predictive performance of the MLR model have been demonstrated by inside and outer statistical validation. Non-linear technique leads to the best QSRR model with good internal and external predictive abilities. It is based on support vector machines using the RBF function for the optimal parameters values. The values of and were 17, 0.2, 0.2, respectively.

Chemical structure and antifungal activity of mint essential oil components

The objective of this research was to determine chemical composition and to evaluate the antifungal activity of essential oil of Mentha piperita. By the application of GC/MS analysis of essential mint oil, 27 components were identified. The major components were menthol (39.9 %), menton (23.51 %), menthyl acetate (7.29 %), 1.8-cineol (5.96 %), isomenton (5.24 %), isomenthol (3.17 %), trans-caryophyllene (2.88 %), limonene (2.14 %), pulegon (1.38 %), beta-pinene (1.14 %) and piperiton (1.03 %). The quantitative structure? retention relationship (QSRR) was employed to predict the retention time (RT) of Mentha piperita essential oil compounds obtained in GC/MS analysis, using twelve molecular descriptors selected by genetic algorithm. The selected descriptors were used, as inputs of an artificial neural network, to build an RT predictive QSRR model. The coefficient of determination was 0.983, during training cycle, indicating that this model could be used for prediction of RT values for essential oil compounds in Mentha piperita essential oil extracts. Essential oil of Mentha piperita showed antifungal activity on all tested isolates in the minimal inhibitory concentration range of 0.2?1.7 ?l/ml and a minimal fungicidal concentration (MFC) range of 1.7?454.5 ?l/ml. The most powerful antifungal activity of mint was observed in C. cladosporioides of MFC value 1.7 ?l/ml. P. aurantiogriseum showed the lowest sensitivity of MFC value 454.5 ?l/ml.

Quantitative structure -retention relationship modeling of selected antipsychotics and their impurities in green liquid chromatography using cyclodextrin mobile phases.

Applying green chromatography methods is currently one of the challenges in liquid chromatography. Among different strategies, using cyclodextrin (CD) mobile phase modifiers was applied in this paper. CDs can form inclusion complexes with a wide variety of hydrophobic organic compounds and, consequently, affect their retention behavior. CD-containing mobile phases possess complicated complexation and adsorption equilibria so retention is dependent not only on chromatographic parameters and properties of the compound but also on properties of compound-CD complex. Docking study was used to calculate association constants of the selected antipsychotics (risperidone, olanzapine, and their impurities) and β-CD complexes and predict which part of the molecule structure will most likely incorporate into the β-CD cavity. Quantitative structure-retention relationship model (QSRR) for selected model substances was built employing artificial neural network (ANN) technique. Reliable QSRR model was obtained using molecular descriptors, complex association constants, and chromatographic factors. The multilayer perceptron network with 11-8-1 topology, trained with back propagation algorithm, showed the best performance. Root mean square error for training, validation, and test was 0.2954, 0.3633, and 0.4864, respectively. The correlation coefficient (R2) between experimentally obtained retention factor values [k(exp)] and values computed or predicted by ANN [k(ANN)] was 0.9962 for training, 0.9927 for validation, and 0.9829 for test, indicating good predictive ability of the model. The optimized network was used for development of green chromatography method for separation of risperidone and its related impurities, as well as olanzapine and its related impurities in a relatively short run time and with low consumption of organic modifier. The developed methods were validated in accordance with ICH Q2 (R1) quideline and all parameters fulfilled the defined criteria. The greenness of the proposed methods has also been demonstrated. Graphical Abstract Complex association constants as inputs of QSRR model in β-cyclodextrin modified HPLC system and development of green chromatography methods.

QSRR Evaluation of the New Anticancer Sulfonamides in View of the cis-trans Isomerism

Background: Fifteen pairs of geometric isomers of new sulfonamide derivatives, exhibiting antitumor activities were analyzed with the use of quantitative structure-retention relationships method. The set of studied compounds bears thiohydrazone moiety, which leads to in-solution formation of an equilibrium between cis and trans isomer. The semi-empirical level of in silico molecular modeling was performed for calculations of molecular descriptors. Methods: During the study, OPLS and MLR were compared, as two types of chemometric approach. QSRR model was proposed applying partial least squares regression method based on chosen descriptors. Results: The relationship between the isomers geometry and chromatographic retention properties derived from in silico model was able to describe that observed pairs of peaks are most likely cis and trans isomers. Conclusion: The applied chemometric techniques also revealed the most influential features of molecules responsible for chromatographic behavior. In silico applied approach seems to be a good supportive tool helpful in identifying geometric isomers with the use of chromatographic retention data. Keywords: Anticancer sulfonamides, cis-trans isomers, descriptors, MLR, OPLS, QSRR.

Kernel-Based, Partial Least Squares Quantitative Structure-Retention Relationship Model for UPLC Retention Time Prediction: A Useful Tool for Metabolite Identification.

We propose a new QSRR model based on a Kernel-based partial least-squares method for predicting UPLC retention times in reversed phase mode. The model was built using a combination of classical (physicochemical and topological) and nonclassical (fingerprints) molecular descriptors of 1383 compounds, encompassing different chemical classes and structures and their accurately measured retention time values. Following a random splitting of the data set into a training and a test set, we tested the ability of the model to predict the retention time of all the compounds. The best predicted/experimental R2 value was higher than 0.86, while the best Q2 value we observed was close to 0.84. A comparison of our model with traditional and simpler MLR and PLS regression models shows that KPLS better performs in term of correlation (R2), prediction (Q2), and support to MetID peak assignment. The KPLS model succeeded in two real-life MetID tasks by correctly predicting elution order of Phase I metabolites, including isomeric monohydroxylated compounds. We also show in this paper that the model's predictive power can be extended to different gradient profiles, by simple mathematical extrapolation using a known equation, thus offering very broad flexibility. Moreover, the current study includes a deep investigation of different types of chemical descriptors used to build the structure-retention relationship.

Quantitative structure–retention relationships applied to development of liquid chromatography gradient-elution method for the separation of sartans

QSRR are mathematically derived relationships between the chromatographic parameters determined for a representative series of analytes in given separation systems and the molecular descriptors accounting for the structural differences among the investigated analytes. Artificial neural network is a technique of data analysis, which sets out to emulate the human brain's way of working. The aim of the present work was to optimize separation of six angiotensin receptor antagonists, so-called sartans: losartan, valsartan, irbesartan, telmisartan, candesartan cilexetil and eprosartan in a gradient-elution HPLC method. For this purpose, ANN as a mathematical tool was used for establishing a QSRR model based on molecular descriptors of sartans and varied instrumental conditions. The optimized model can be further used for prediction of an external congener of sartans and analysis of the influence of the analyte structure, represented through molecular descriptors, on retention behaviour. Molecular descriptors included in modelling were electrostatic, geometrical and quantum-chemical descriptors: connolly solvent excluded volume non-1,4 van der Waals energy, octanol/water distribution coefficient, polarizability, number of proton-donor sites and number of proton-acceptor sites. Varied instrumental conditions were gradient time, buffer pH and buffer molarity. High prediction ability of the optimized network enabled complete separation of the analytes within the run time of 15.5min under following conditions: gradient time of 12.5min, buffer pH of 3.95 and buffer molarity of 25mM. Applied methodology showed the potential to predict retention behaviour of an external analyte with the properties within the training space. Connolly solvent excluded volume, polarizability and number of proton-acceptor sites appeared to be most influential paramateres on retention behaviour of the sartans.

Theoretical Models and QSRR in Retention Modeling of Eight Aminopyridines.

In this article, retention modeling of eight aminopyridines (synthesized and characterized at the Faculty of Pharmacy) in reversed-phase high performance liquid chromatography (RP-HPLC) was performed. No data related to their retention in the RP-HPLC system were found. Knowing that, it was recognized as very important to describe their retention behavior. The influences of pH of the mobile phase and the organic modifier content on the retention factors were investigated. Two theoretical models for the dependence of retention factor of organic modifier content were tested. Then, the most reliable and accurate prediction of log k was created, testing multiple linear regression model-quantitative structure-retention relationships (MLR-QSRR) and support vector regression machine-quantitative structure-retention relationships (SVM-QSRR). Initially, 400 descriptors were calculated, but four of them (POM, log D, M-SZX/RZX and m-RPCG) were included in the models. SVM-QSRR performed significantly better than the MLR model. Apart from aminopyridines, four structurally similar substances (indapamide, gliclazide, sulfamethoxazole and furosemide) were followed in the same chromatographic system. They were used as external validation set for the QSRR model (it performed well within its applicability domain, which was defined using a bounding box approach). After having described retention of eight aminopyridines with both theoretical and QSRR models, further investigations in this field can be conducted.

Correlation study of retention data and antimalarial activity of 1,2,4,5-mixed tetraoxanes with their molecular structure descriptors and LSER parameters

The chromatographic behavior of mixed 1,2,4,5-tetraoxanes, cholic and deoxycholic acid derivatives with distinct biological activity, was examined by high-performance thin-layer chromatography in order to correlate their structure and retention. Chromatographic systems were consisted of RP-18 or CN-silica as stationary phase, and binary mixtures of water with methanol, dioxane or acetone as mobile phase. Based on the respective retentions, the lipophilicity of the investigated compounds was determined. Multiple linear regression and partial least squares have been used to select variables that best describe the behavior of the investigated compounds in chromatographic systems and to quantify influences of most important parameters. The validation and cross-validation of the QSRR model suggest its applicability for prediction and understanding of retention of congeners. The models indicate the importance of nonpolar properties of the solutes and their ability for hydrophobic interactions, as well as the importance of proton donating abilities, hydrophilic and π interactions pointing out on that way the possible separation mechanism in the studied chromatographic systems. Observed correlations between structure and biological activity of mixed 1,2,4,5-tetraoxanes, indicate that the antimalarial activity against W2 and D6 Plasmodium falciparum strains, is governed by hydrophobic feature (measured with lipophilicity parameter), hydrophilic feature (measured with HLB, %HS, HB and HBA descriptors), and electronic feature (HOMO).

Semi-Empirical Topological Method for Prediction of the Relative Retention Time of Polychlorinated Biphenyl Congeners on 18 Different HR GC Columns.

High resolution gas chromatographic relative retention time (HRGC-RRT) models were developed to predict relative retention times of the 209 individual polychlorinated biphenyls (PCBs) congeners. To estimate and predict the HRGC-RRT values of all PCBs on 18 different stationary phases, a multiple linear regression equation of the form RRT = a o + a 1 (no. o-Cl) + a 2 (no. m-Cl) + a 3 (no. p-Cl) + a 4 (V M or S M) was used. Molecular descriptors in the models included the number of ortho-, meta-, and para-chlorine substituents (no. o-Cl, m-Cl and p-Cl, respectively), the semi-empirically calculated molecular volume (V M), and the molecular surface area (S M). By means of the final variable selection method, four optimal semi-empirical descriptors were selected to develop a QSRR model for the prediction of RRT in PCBs with a correlation coefficient between 0.9272 and 0.9928 and a leave-one-out cross-validation correlation coefficient between 0.9230 and 0.9924 on each stationary phase. The root mean squares errors over different 18 stationary phases are within the range of 0.0108–0.0335. The accuracy of all the developed models were investigated using cross-validation leave-one-out (LOO), Y-randomization, external validation through an odd–even number and division of the entire data set into training and test sets.Electronic supplementary materialThe online version of this article (doi:10.1365/s10337-010-1696-5) contains supplementary material, which is available to authorized users.

Structural characterization and Kovats retention index prediction for oxygen-containing organic compounds

A method of molecular structural characterization (MSC) called hydrogen-association classified molecular electronegativity-distance vector (H-MEDV) was used to describe the molecular structure of 106 oxygen-containing organic compounds. The samples were divided into a training set and a test set. For the training set, a QSRR model was built up by using multiple linear regression (MLR) and this model was evaluated by performing the cross-validation with the leave-one-out (LOO) procedure. The correlation coefficient ( R) and the cross-verification correlation coefficient ( R CV) of the model were 0.964 and 0.915, respectively. Furthermore, the validity of the model was tested by the test set and satisfied results with correlation coefficient ( R est) of 0.948. The results show that the model constructed in this work could provide estimation stability and favorable predictive ability.

Quantitative structure–retention relationships of polycyclic aromatic hydrocarbons gas-chromatographic retention indices

Polycyclic aromatic compounds (PAHs) are of concern in environmental chemistry and toxicology. In the present work, a QSRR study was performed for 209 previously reported PAHs using quantum mechanics and other sources descriptors estimated by different approaches. The B3LYP/6-31G* level of theory was used for geometrical optimization and quantum mechanics related variables. A good linear relationship between gas-chromatographic retention index and electronic or topologic descriptors was found by stepwise linear regression analysis. The molecular polarizability ( α) and the second order molecular connectivity Kier and Hall index ( 2 χ) showed evidence of significant correlation with retention index by means of important squared coefficient of determination, ( R 2), values ( R 2 = 0.950 and 0.962, respectively). A one variable QSRR model is presented for each descriptor and both models demonstrates a significant predictive capacity established using the leave-many-out LMO (excluding 25% of rows) cross validation method's q 2 cross-validation coefficients q 2 CV-LMO25%, (obtained q 2 CV-LMO25% 0.947 and 0.960, respectively). Furthermore, the physicochemical interpretation of selected descriptors allowed detailed explanation of the source of the observed statistical correlation. The model analysis suggests that only one descriptor is sufficient to establish a consistent retention index–structure relationship. Moderate or non-significant improve was observed for quantitative results or statistical validation parameters when introducing more terms in predictive equation. The one parameter QSRR proposed model offers a consistent scheme to predict chromatographic properties of PAHs compounds.