Quantitative Structure-property Relationship Analysis Research Articles

Density functionals of chemical bonding.

The behavior of electrons in general many-electronic systems throughout the density functionals of energy is reviewed. The basic physico-chemical concepts of density functional theory are employed to highlight the energy role in chemical structure while its extended influence in electronic localization function helps in chemical bonding understanding. In this context the energy functionals accompanied by electronic localization functions may provide a comprehensive description of the global-local levels electronic structures in general and of chemical bonds in special. Becke-Edgecombe and author’s Markovian electronic localization functions are discussed at atomic, molecular and solid state levels. Then, the analytical survey of the main workable kinetic, exchange, and correlation density functionals within local and gradient density approximations is undertaken. The hierarchy of various energy functionals is formulated by employing both the parabolic and statistical correlation degree of them with the electronegativity and chemical hardness indices by means of quantitative structure-property relationship (QSPR) analysis for basic atomic and molecular systems.

Prediction of Solubility of Nonionic Solutes in Anionic Micelle (SDS) Using a QSPR Model

AbstractA Quantitative structure–property relationships analysis has been applied to a series of chemical compounds which were of interest because of their roles in environmental samples. The incorporation neutral solute (log Ks) in anionic micelle (SDS) has been predicted by using molecular structural descriptors for a heterogeneous set of 62 neutral solutes with a range of more than 4log units. Two linear correlating models, Multiple Linear Regression (MLR) and Partial Least Squares (PLS) regression methods were used. The stepwise MLR of SPSS software was used for the selection of the variables. After variables selection, MLR and PLS methods used leave‐one‐out cross‐validation for building the regression models. Appropriate models with low standard errors and high correlation coefficients were selected. The results showed that both MLR and PLS methods could model the relationship between solubility and their electronic and thermodynamic descriptors perfectly. The predictive quality of the QSPR models were tested for an external prediction set of 11 compounds randomly chosen from 62 compounds. The squared regression coefficients of prediction for the MLR and PLS regression methods were 0.9679 and 0.9728 respectively.

Development of a New Regression Analysis Method Using Independent Component Analysis

In this paper, independent component analysis (ICA) and regression analysis are combined to extract significant components. ICA is a method that extracts mutually independent components from explanatory variables. A relationship between the independent components and an objective variable is constructed by the least-squares method. This method is named ICA-MLR (MLR = multiple linear regression). We verified the superiority of ICA-MLR over partial least squares (PLS) with simulation data and tried to apply this method to a quantitative structure-property relationship analysis of aqueous solubility. We constructed models between aqueous solubility and 173 molecular descriptors. PLS and genetic algorithm PLS models were constructed for a comparison of ICA-MLR. R2, Q2, and Rpred2 values of the PLS model are 0.836, 0.819, and 0.848, respectively. These values of the ICA-MLR model are 0.937, 0.868, and 0.894, respectively. ICA-MLR achieved higher predictive accuracy than PLS. ICA-MLR could extract effective components from explanatory variables and construct the regression model with high predictive accuracy. In addition, the information of regression coefficients bICA-MLR indicates the magnitude of contribution of each descriptor in the analysis of aqueous solubility.

Photolysis of mono-through deca-chlorinated biphenyls by ultraviolet irradiation in n-hexane and quantitative structure-property relationship analysis

The photolysis of 16 polychlorinated biphenyls (PCBs) (including mono- through deca-chlorinated) in n-hexane was investigated under ultraviolet irradiation using a 500-W high-pressure mercury lamp. Photolysis of PCBs follows pseudo-first-order reaction kinetics, with photolysis rate constants ranging between 0.0011 s(-1) for PCB-52 and 0.0574 s(-1) for PCB-118. The degradation rates of PCBs by high-pressure mercury lamp irradiation were remarkably independent with respect to the degree of chlorination. Furthermore, partial least squares (PLS) models were developed to provide insight into which aspect of the molecular structure influenced PCB photolysis rate constants. It was found that the photolysis rates of PCBs increased with an increase in the net charge on the carbon atom (qc), (E(LUMO)-E(HOMO))2, and the Y-axis dipole moment (mu(y)) values, or the decrease in the energy of the second highest occupied molecular orbital (E(HOMO-1)), energy of the lowest unoccupied molecular orbital (E(LUMO)), E(LUMO) + E(HOMO), E(LUMO)--E(HOMO), most positive atomic charge (q+), and the twist angle of the chlorine atom (TA) values.

QSPR analysis of air-to-blood distribution of volatile organic compounds

Quantitative structure property relationship (QSPR) models for the prediction of human blood:air partition coefficient (log K blood) of volatile organic compounds (VOCs) has been developed based on the linear heuristic method (HM) and non-linear radial basis function neural networks (RBFNNs). Molecular descriptors that are calculated from the structures alone were used to represent the characteristics of the compounds. HM was used both to pre-select the whole descriptor sets and to build the linear model. RBFNN was performed to obtain more accurate models. Both the linear and the non-linear models can give very satisfactory prediction results: the correlation coefficient R was 0.964 and 0.979, and the root-mean-square (RMS) error was 0.3303 and 0.2542 for the whole data set, respectively. The prediction result of the non-linear model is better than that obtained by the linear model. In addition, this paper provides an effective method for predicting log K blood from its structures and gives some insight into the structural features related to the solubility of VOCs in human blood.

Quantitative structure–property relationship study of n-octanol–water partition coefficients of some of diverse drugs using multiple linear regression

A quantitative structure–property relationship (QSPR) study was performed to develop models those relate the structures of 150 drug organic compounds to their n-octanol–water partition coefficients (log P o/w). Molecular descriptors derived solely from 3D structures of the molecular drugs. A genetic algorithm was also applied as a variable selection tool in QSPR analysis. The models were constructed using 110 molecules as training set, and predictive ability tested using 40 compounds. Modeling of log P o/w of these compounds as a function of the theoretically derived descriptors was established by multiple linear regression (MLR). Four descriptors for these compounds molecular volume (MV) (geometrical), hydrophilic–lipophilic balance (HLB) (constitutional), hydrogen bond forming ability (HB) (electronic) and polar surface area (PSA) (electrostatic) are taken as inputs for the model. The use of descriptors calculated only from molecular structure eliminates the need for experimental determination of properties for use in the correlation and allows for the estimation of log P o/w for molecules not yet synthesized. Application of the developed model to a testing set of 40 drug organic compounds demonstrates that the model is reliable with good predictive accuracy and simple formulation. The prediction results are in good agreement with the experimental value. The root mean square error of prediction (RMSEP) and square correlation coefficient ( R 2) for MLR model were 0.22 and 0.99 for the prediction set log P o/w.

A QSPR Study on the GC Retention Times of a Series of Fatty, Dicarboxylic and Amino Acids by MLR and ANN

Quantitative structure-property relationship (QSPR) analysis has been carried out to a series of fatty, amino and dicarboxylic acids to model their GC retention times. A genetic partial least square method (GAPLS) was applied as a variable selection tool. Modeling of retention times of these compounds as a function of the theoretically derived descriptors was established by multiple linear regression (MLR) and artificial neural network (ANN). The neural network employed here is a connected back-propagation system with a 3-4-1 architecture. Three topological indices for these compounds, namely, mean information index on atomic composition (AAC), average connectivity index chi-0 (X0A) and total information index of atomic composition (IAC) taken as inputs for the regression models. The results indicate that the GA is a very effective variable selection approach for QSPR analysis. The comparison of the two regression methods used showed that ANN has better prediction ability than MLR. The statistical figure of merits of the two models showed the successful modeling of the retention times with molecular descriptors.

QSPR analysis of the toxicity of aromatic compounds to the algae (Scenedesmus obliquus)

The quantitative structure-property relationship (QSPR) model was developed for the 50% effective inhibition concentration (48h - EC(50)) of 36 selected substituted benzenes for the algae Scenedesmus obliquus by the application of the Characteristic Root Index (CRI) model. To increase the predictive power of the CRI-based model, the following semi-empirical molecular descriptors calculated by the quantum chemical PM3 method were included: the energy of the highest occupied molecular orbital (E(HOMO)), the energy of the lowest unoccupied molecular orbital (E(LUMO)), and the dipole moment (mu). A two-descriptor model with a correlation coefficient of r=0.926 was developed without the outliers from multiple regression analysis [-logEC(50)=0.494 (+/-0.072) CRI-0.798 (+/-0.063) E(LUMO)+1.985 (+/-0.169)]. E(LUMO) was the most important parameter, followed by the CRI. E(LUMO) reflects electronic properties, whereas the CRI reflects hydrophobicity, molecular size, and branching. The statistical robustness of the developed model was validated by the modified jackknife test. The predictive accuracy of the proposed model was compared with the recently published study in which a toxicity model was developed for the same algae. Because of its high statistical significance, the validated model has been used to predict -logEC(50) values of compounds for which there are no experimental measurements.

QSPR analysis for intrinsic viscosity of polymer solutions by means of GA-MLR and RBFNN

A quantitative structure–property relationship (QSPR) treatment of intrinsic viscosity of polymer solutions was performed by means of a genetic algorithm based multivariate linear regression (GA-MLR). A five parameters correlation, with squared correlation coefficient R2=0.8275 gives good predictions for 65 polymer solutions. In preparation of this model, 1664 molecular descriptors for each polymer and 1664 molecular descriptors for each solvent were checked and finally, five molecular descriptors were selected. For considering the nonlinear behavior of these five molecular descriptors, a radial based function neural network (RBFNN) with squared correlation coefficient R2=0.9100 was constructed. Notably, all the parameters involved in these equations can be derived solely from the chemical structure of the polymers repeating unit and the solvents which makes them very useful for prediction of the intrinsic viscosity of unknown or unavailable polymer solutions.

Combination of Genetic Algorithm and Partial Least Squares for Cloud Point Prediction of Nonionic Surfactants from Molecular Structures

Quantitative structure-property relationship (QSPR) analysis has been directed to a series of pure nonionic surfactants containing linear alkyl, cyclic alkyl, and alkey phenyl ethoxylates. Modeling of cloud point of these compounds as a function of the theoretically derived descriptors was established by multiple linear regression (MLR) and partial least squares (PLS) regression. In this study, a genetic algorithm (GA) was applied as a variable selection method in QSPR analysis. The results indicate that the GA is a very effective variable selection approach for QSPR analysis. The comparison of the two regression methods used showed that PLS has better prediction ability than MLR.

Cyclic voltammetric, computational, and quantitative structure–electrochemistry relationship studies of the reduction of several 9,10-anthraquinone derivatives

The electrochemical behavior of 33 newly synthesized 9,10-anthraquinone (Aq) derivatives were studied in acetonitrile solution by cyclic voltammetry, under the same experimental conditions. In the absence of a proton donor source, the anthraquinones reduced in two successive one-electron steps. The first step is reversible while the second step is reversible or quasi-reversible. Some electrochemical properties such as peak potentials and half-wave reduction potentials were determined from the recorded cyclic voltammograms. Semi-empirical PM3 and DFT methods at the level of B3LYP have been used to compute the redox potentials for 15 typical anthraquinone derivatives in acetonitrile solution. A linear relationship was observed between the theoretically predicted values and experimentally determined half-wave reduction potentials of the 9,10-anthraquinone derivatives. The classification of Aq derivatives was carried out by the application of principal component analysis on the current data of the cyclic voltamograms and plotting the resulted scores. The quantitative effects of structural characteristics of the molecules on their electrochemical properties were investigated by quantitative structure–property relationship analysis. Linear multi-parametric equations between electrochemical properties as dependent variable and molecular descriptors as independent variables were obtained by genetic algorithm-multiple linear regression analysis. The resulting QSPR equations indicated that the topology of the molecules is an important factor in the first reduction wave while the electronic parameters significantly affect the second reduction wave.

Quantitative structure-property relationship research of main group compounds

New approaches were applied to improve the molecular connectivity indicesmχgu. The vertex valence is redefined and it was reasonable for hydrogen atom. The distances between vertices were used to propose novel connectivity topological indexes. The vertices and the distances in a molecular graph were taken into account in this definition. The linear regression was used to develop the structural property models. The results indicate that the novel connectivity topological indexes are useful model parameters for Quantitative Structure-Property Relationship (QSPR) analysis.

Density of multicomponent silica-based potential bioglasses: Quantitative structure-property relationships (QSPR) analysis

The results of a quantitative structure-property relationships (QSPR) analysis of multicomponent silica-based potential bioglasses (containing Na 2O, CaO, P 2O 5 and/or ZnO) are here presented. A quantitative model explaining the variation of the density data measured for series of glasses with different compositions has been obtained by means of a structural descriptor derived from molecular dynamics (MD) simulations. A descriptor able to rationalize the variation in density caused by the overall packing degree of the structural units in the glasses examined has been defined. It is worth noting that the descriptor used allows the fitting of glasses with different composition (presence–absence of P 2O 5, ZnO, Na 2O and CaO) in the same correlation. The validity of the QSPR approach, which has recently been introduced for the rationalization and prediction of the technology-related properties of a series of complex multicomponent glasses, is confirmed by this work on a larger series of glasses of various compositions.

The RP-HPLC measurement and QSPR analysis of log Po/w values of several Pt(II) complexes

The n-octanol/water partition coefficient, log P o/w, for a set of 24 Pt(II)-complexes was estimated by means of reversed-phase high performance liquid chromatography (RP-HPLC) technique using a C18 (ODS, octadecyl silane) column as a stationary phase and water/methanol mixtures as mobile phases. Based on the known log P o/w of several Pt(II)-complexes, we set a method to correlate the partition coefficient of this kind of complexes with the corresponding retention parameters. The best result was obtained from extrapolation to 0% of the organic modifier (MeOH) of the aqueous eluant. A quantitative structure–property relationship (QSPR) was constructed using molecular descriptors derived from density functional theory (DFT) calculations, which was found to correlate and predict these values with good accuracy. The use of DFT calculations is required because group-additive methods fail due to lack of values for appropriate fragments for many Pt(II)-complexes.

The Classification of Solvents by Combining Classical QSPR Methodology with Principal Component Analysis

The results of a quantitative structure-property relationship (QSPR) analysis of 127 different solvent scales and 774 solvents using the CODESSA PRO program are presented. QSPR models for each scale were constructed using only theoretical descriptors. The high quality of the models is reflected by the squared multiple correlation coefficients that range from 0.726 to 0.999; only 18 models have R2< 0.800. This enables direct theoretical calculation of predicted values for any scale and/or for any organic solvent, including those previously unmeasured. The molecular descriptors involved in the models are classified and discussed according to (i) the origin of their calculation (i.e., constitutional, geometric, charge-related, etc.) and (ii) the commonly accepted classification of physical interactions between the solute and solvent molecules in liquid (condensed) media. A reduced matrix 774 (solvents) x 100 (solvent scales) was selected for the principal component analysis (PCA) by taking into account only the solvent scales with more than 20 experimental data points. The first 5 principal components account for 75% of the total variance. The robustness of the PCA model obtained was validated by the comparison models development for restricted submatrices of data and with the results obtained for the full data set. The total variance accounted for by the first three PCs, for the submatrices with the same number of solvent scales but different numbers of solvents, varies from 68.2% to 59.0%. This demonstrates that the total variance described by the first 3 components is essentially stable as the number of solvents involved varies from 100 to 774. Subsequently, a matrix with 703 diverse solvents and 100 solvent scales was selected for the general classification of the solvents and scales according to the scores and loadings obtained from the PCA treatment. Classification of the theoretical molecular descriptors, derived from the chemical structure alone, according to their relevance to specific types of intermolecular interaction (cavity formation, electrostatic polarization, dispersion, and hydrogen bonding) in liquid media enables a more easily comprehensible physical interpretation of the QSPR of molecular properties in liquids and solutions. The reported QSPR models for solvent scales with theoretical molecular descriptors and the results of the PCA analysis are potentially of great practical importance, as they extend the applicability of correlations with empirical solvent scales to many previously unmeasured systems.

A General Treatment of Solubility. 3. Principal Component Analysis (PCA) of the Solubilities of Diverse Solutes in Diverse Solvents

A phenomenological study of solubility has been conducted using a combination of quantitative structure-property relationship (QSPR) and principal component analysis (PCA). A solubility database of 4540 experimental data points was used that utilized available experimental data into a matrix of 154 solvents times 397 solutes. Methodology in which QSPR and PCA are combined was developed to predict the missing values and to fill the data matrix. PCA on the resulting filled matrix, where solutes are observations and solvents are variables, shows 92.55% of coverage with three principal components. The corresponding transposed matrix, in which solvents are observations and solutes are variables, showed 62.96% of coverage with four principal components.

Highly diverse, massive organic data as explored by a composite QSPR strategy: An advanced study of boiling point

An improved strategy of quantitative structure–property relationship (QSPR) studies of diverse and inhomogeneous organic datasets has been proposed. A molecular connectivity term was successively corrected for different structural features encoded in fragmental descriptors. The so-called solvation index 1χs (a weighted Randić index) was used as a “leading” variable and standardized molecular fragments were employed as “corrective” class-specific variables. Performance of the new approach was illustrated by modelling a dataset of experimental normal boiling points of 833 organic compounds belonging to 20 structural classes. Firstly, separate QSPR models were derived for each class and for eight groups of structurally similar classes. Finally, a general model formed by combining all the classes together was derived ( ). The strategy outlined can find application in QSPR analyses of massive, highly diverse databases of organic compounds.

Combination of Antimicrobial and Endotoxin-Neutralizing Activities of Novel Oleoylamines

A combination of antimicrobial and endotoxin-neutralizing activities is desired in order to prevent progression from infection to sepsis due to the release of lipopolysaccharide from dying gram-negative bacteria. Lipopolyamines have emerged as a new type of endotoxin-neutralizing compound, but their antimicrobial activity has not been investigated. We synthesized a series of 10 oleoylamines differing in the polyamino head group, particularly in the number and separation between nitrogen atoms and the position of the oleoyl moiety. Compounds showed activity against both gram-negative and gram-positive bacteria in the micromolar range. Compounds were able to provide penetration of ethidium bromide into bacteria, indicating effects on the bacterial membrane. Oleoylamines neutralized endotoxin in Limulus amoebocyte lysate assays and by neutralization of tumor necrosis factor alpha release in human blood. Comparison of biological activities of compounds identified structural properties responsible for antimicrobial activity, and quantitative structure-property relationship analysis provided a quantitative model for prediction of activity of oleoylamines.

Qualitative and Quantitative Structure−Property Relationships Analysis of Multicomponent Potential Bioglasses

The results of a qualitative and quantitative structure-property relationships analysis of multicomponent potential bioglasses of composition (2 - y) SiO2 x 1 Na2O x 1.1 CaO x y P2O5 x x ZnO (x = 0, 0.16, 0.35, 0.78 and y = 0.10, 0.20, 0.36) are presented. Quantitative models are obtained by means of structural descriptors derived by molecular dynamics simulations and experimental data measured for density, thermal analysis, 29Si and 31P magic angle spinning NMR, and chemical durability in water. Analysis of the crystal species obtained upon glass crystallization helped in the rationalization of the structural role of the different components. Finally, glass surface characterization with scanning electron microscopy, transmission electron microscopy, and X-ray diffraction after soaking in acellular simulated body fluid demonstrated the in vitro bioactivity of the newly obtained 1.80 SiO2 x 1 Na2O x 1.1 Ca x 0.20 P2O5 x 0.16 ZnO (HP5Z5) glass, corresponding to x = 0.16 and y = 0.20.

Genetic algorithms and self-organizing maps: a powerful combination for modeling complex QSAR and QSPR problems

Modeling non-linear descriptor-target activity/property relationships with many dependent descriptors has been a long-standing challenge in the design of biologically active molecules. In an effort to address this problem, we couple the supervised self-organizing map with the genetic algorithm. Although self-organizing maps are non-linear and topology-preserving techniques that hold great potential for modeling and decoding relationships, the large number of descriptors in typical quantitative structure-activity relationship or quantitative structure-property relationship analysis may lead to spurious correlation(s) and/or difficulty in the interpretation of resulting models. To reduce the number of descriptors to a manageable size, we chose the genetic algorithm for descriptor selection because of its flexibility and efficiency in solving complex problems. Feasibility studies were conducted using six different datasets, of moderate-to-large size and moderate-to-great diversity; each with a different biological endpoint. Since favorable training set statistics do not necessarily indicate a highly predictive model, the quality of all models was confirmed by withholding a portion of each dataset for external validation. We also address the variability introduced onto modeling through dataset partitioning and through the stochastic nature of the combined genetic algorithm supervised self-organizing map method using the z-score and other tests. Experiments show that the combined method provides comparable accuracy to the supervised self-organizing map alone, but using significantly fewer descriptors in the models generated. We observed consistently better results than partial least squares models. We conclude that the combination of genetic algorithms with the supervised self-organizing map shows great potential as a quantitative structure-activity/property relationship modeling tool.