Genetic Algorithm-multiple Linear Regression Research Articles

Prediction of PCE of fullerene (C60) derivatives as polymer solar cell acceptors by genetic algorithm–multiple linear regression

Quantitative structure property relationship study of Fullerene derivatives was studied to predict the power conversion efficiency of compounds as polymer solar cell acceptors. The data set was split into the training and test set by employing hierarchal cluster technique. The most relevant descriptors were selected using the genetic algorithm (GA) method. The predictive ability of the constructed model was evaluated using Y-randomization test, cross-validation and test set compounds. The GA–MLR model was built based on six molecular descriptors, and it revealed appropriate statistical results. The results suggested that some quantum-chemical descriptors play significant effects on increasing the PCE values.

Application of HS-SPME, SDME and Cold-Press Coupled to GC/MS to Analysis the Essential Oils of Citrus sinensis CV. Thomson Navel and QSRR Study for Prediction of Retention Indices by Stepwise and Genetic Algorithm-Multiple Linear Regression Approaches

Volatile fractions in the essential oils (EOs) obtained from the peels of Citrus sinensis CV. Thomson Navel by head-space solid phase micro-extraction (HS-SPME), single drop micro-extraction (SDME) and cold-press (CP) techniques were analyzed by means of gas chromatography-mass spectrometry (GC-MS). Fifteen, nine and fourteen compounds were identified by HS-SPME, SDME and CP methods, respectively representing 99.75 %, 99.45 % and 98.5 % of the chemical profiles in the peels of C. sinensis. The main components were found to be limonene (72.69 %, 71.96 %, 90.40 %), myrcene (9.65 %, 9.80 %, 1.10 %), sabinene (2.63 %, 9.35 %, 0.90 %) and α-pinene (2.00 %, 7.05 %, 1.60 %) by HS-SPME, SDME and CP, respectively. In addition, a quantitative structure-retention relationship (QSRR) study has been developed for prediction of retention indices (RIs) of the occurred natural compounds. The suitable set of molecular descriptors was calculated and the important descriptors were selected by stepwise (SW) and genetic algorithm (GA) variable selection features combined with multiple linear regression (MLR). A comparison between the attained results indicated the superiority of the genetic algorithm over the stepwise in the feature-selection step. The predictive quality of the QSRR models was tested for an external set of five compounds, randomly chosen out of 27 compounds. The genetic algorithm-multiple linear regression model (GA-MLR) with two selected descriptors gave rise in the best predicted results. The accuracy of the proposed GA-MLR model is illustrated using cross-validation, validation through an external test set and Y-randomization evaluation techniques.

QSAR study of ACK1 inhibitors by genetic algorithm–multiple linear regression (GA–MLR)

In this work, a quantitative structure–activity relationship (QSAR) model was used to predict the ACK1 inhibitory activities. A data set of 37 compounds with known ACK1 inhibitory activities was used. The data set was divided into two subsets of training and test sets, based on hierarchical clustering technique. Genetic algorithm was applied to select the respective variables to build the model in the next step. Multiple linear regressions (MLR) were employed to give the QSAR model. The squared cross-validated correlation coefficient for leave-one-out (QLOO2) of 0.712 and squared correlation coefficient (Rtrain2) of 0.806 were obtained for the training set compounds by GA–MLR model. The given model performed a good stability and predictability when it was verified by internal and external validation. The predicted results from this study can lead to design of better and potent ACK1 inhibitors.

QSAR study of mGlu5 inhibitors by genetic algorithm-multiple linear regressions

In this study, the quantitative structure–activity relationship (QSAR) model for some pyrazole/imidazole amide derivatives as mGlu5 inhibitors was developed. The data set was split into the training and test subsets, randomly. The most relevant variables were selected using the genetic algorithm (GA) variable selection method. Multiple linear regression (MLR) method was used as a linear model to predict the activity of mGlu5 inhibitors based on compounds in training set. The external set of nine compounds selected out of 47 compounds, and used to evaluate the predictive ability of QSAR model. The built model could give high statistical quantities (Rtrain2 = 0.837, Q2 = 0.759, Rtest2 = 0.919) in which proved that the GA-MLR model was a useful tool to predict the inhibitory activity of pyrazole/imidazole amide derivatives. The results suggested that the atomic masses, atomic van der Waals volumes, atomic electronegativities, and the number of imines (aromatic) are the most important independent factors that contribute to the mGlu5 inhibition activity of pyrazole/imidazole amides derivatives.

QSAR study of Nav1.7 antagonists by multiple linear regression method based on genetic algorithm (GA–MLR)

In this work, a quantitative structure–activity relationship study was developed to predict the NaV1.7 antagonist activities. A data set consisted of 36 compounds with known NaV1.7 antagonist activities was split into two subsets of training set and test set using hierarchical clustering technique. To select the most respective descriptors among the pool of descriptors, genetic algorithm was applied. The model based on selected descriptors through genetic algorithm (GA) was built by employing multiple linear regression (MLR) method. The squared correlation coefficient (\( R_{\text{train}}^{2} \)) of 0.813, squared cross-validated correlation coefficient for leave-one-out (\( Q_{\text{LOO}}^{2} \)) of 0.699 and root mean square error of 0.214 were calculated for the training set compounds by GA–MLR model. The proposed model performed good predictive ability when it was verified by internal and external validation tests. The results of predictive model can lead to design better compounds with high NaV1.7 antagonist activities.

Quantitative structure–property relationship study on the intercalation of anticancer drugs with ct-DNA

Studies on the intercalation of anticancer compounds with DNA can provide useful suggestions and guidance for the design of new and more efficient anticancer drugs. A quantitative structure–property relationship (QSPR) study of a series of anticancer and candidate anticancer drugs with calf thymus DNA (ct-DNA) was performed. Constitutional, Topological, and WHIM descriptors, as well as GETAWAY, 3D-MoRSE, and Aromaticity Indices descriptors generated from Dragon, were selected to describe the molecules. The resampling by half-means method was used to detect the outlier molecules. Self-organizing map was used to split the original dataset into training and test set. Genetic algorithm–multiple linear regression technique was used to establish QSPR model for training set. Finally, the best four-molecular descriptor model was developed on a training set of molecules and the external validation was performed on test set of molecules. The stability and predictability of QSPR model were determined with the leave-one-out cross-validated variance and the external-validated variance. This QSPR approach can contribute to a better understanding of structural factors of the compounds responsible for drug–DNA intercalations, and be useful in predicting the binding affinity of other compounds with DNA.

Application of self organizing maps and GA-MLR for the estimation of stability constant of 18-crown-6 ether derivatives with sodium cation

A quantitative structure–property relationship (QSPR) study of a binding constant of 18-crown-6 ether derivatives with sodium cation was performed. Constitutional, Topological and WHIM, as well as GETAWAY, 3D-MoRSE, and Aromaticity Indices descriptors generated from Dragon, were selected to describe the molecules. Self organizing map (SOM) was used to split the original data set into calibration and test set. Genetic algorithm-multiple linear regression technique was used to establish QSPR model for calibration set. Finally, the best six-dimensional model, developed on a calibration set of molecules and the external validation was performed on test set of molecules. The stability and predictability of QSPR model determined with the leave one out cross validated variance and the external-validated variance. This QSPR approach can contribute to a better understanding of structural factors of the crown ethers with metal cations and be useful in predicting the binding constant of other organic compounds with metal ions.

Quantitative structure activity relationship study of p38α MAP kinase inhibitors

The quantitative structure activity relationship (QSAR) of the novel pyrazole derivatives as inhibitors of p38α mitogen activated protein (MAP) kinase was studied. The suitable set of the molecular descriptors was calculated and the important descriptors using the variable selections of the stepwise (SW) and the genetic algorithm (GA) were selected. The predictive quality of the QSAR models was tested for an external set of nine compounds, randomly chosen out of 44 compounds. A comparison between the attained results indicated the superiority of the genetic algorithm over the stepwise method in the feature selection. The genetic algorithm-multiple linear regression (GA-MLR) model with six selected descriptors was obtained. The accuracy of the proposed model is illustrated using the following evaluation techniques: cross-validation, validation through an external test set, applicability domain, and Y-randomization. The analyses may be used to design more potent pyrazole derivatives and predict their activity prior to synthesis.

Prediction of the complexation stabilities of La3+ ion with ionophores applied in lanthanoid sensors

The complexation stability of lanthanum ion with different ionophores was efficiently predicted by the quantitative structure–property relationship (QSPR) model. Since the selectivity of ionophores can be determined by the stability constants of lanthanum–ionophore complexes, QSPR studies on complex stability constants (logK) were performed. The Genetic algorithm-multiple linear regression method was used to develop models based on calculated molecular descriptors. The best model was validated using Leave-One-Out and Leave-Group-Out cross validation, external test set and Y-randomization. The MLR model, with high statistical significance (Rtrain2 = 0.910, QLOO2 = 0.868, QLGO2 = 0.860), could be used to predict the stability constants of lanthanum–ionophores complexes. This paper provided a novel method that only used chemical structures to develop efficient and stable model for testing and estimating ionophores in lanthanoide sensors based on complex stability constants, for the first time.

Prediction of tyrosinase inhibition for drug design using the genetic algorithm–multiple linear regressions

Modeling inhibition data for a series of most potent tyrosinase inhibitors for hyper-pigmentation treatment was done with quantitative structure–activity relationship (QSAR). In this report, multiple linear regression (MLR) methodology coupled with feature selection method, genetic algorithm (GA), was applied to derive QSAR models. A model with seven selected descriptors was obtained by the database that consisted of 49 compounds. The power of the model for prediction was verified with the leave-one-out and leave-group-out cross-validation test, which have values of 0.766 and 0.795, respectively. Moreover, the statistical parameters provided by the GA–MLR model (\( R_{\text{train}}^{ 2} \) = 0.85, \( R_{\text{test}}^{ 2} \) = 0.84, and \( F_{\text{train}}^{ 2} \) = 24.498) lead to a better understanding of the structural requirements of designing novel potent tyrosinase inhibitors. The study showed that the number of six-membered rings, the branching of a molecule, the van der Waals volume of atoms, the lopping centric index, the global charge transfer in a molecule and atomic masses are the main criteria in the inhibitory activity of compounds.

Quantitative Structure-Electrochemistry Relationship Study for Prediction of Half-Wave Reduction Potentials of Some Chlorinated Organic Compounds by Genetic Algorithm-Multiple Linear Regression

Quantitative Structure-Electrochemistry Relationship Study for Prediction of Half-Wave Reduction Potentials of Some Chlorinated Organic Compounds by Genetic Algorithm-Multiple Linear Regression

Is it Possible to Predict the ADI of Pesticides using the QSAR Approach?

Objectives: QSAR methodology was applied to explain two different sets of acceptable daily intake (ADI) data of 74 pesticides proposed by both the USEPA and WHO in terms of setting guidelines for food and drinking water. Methods: A subset of calculated descriptors was selected from Dragon® software. QSARs were then developed utilizing a statistical technique, genetic algorithm-multiple linear regression (GA-MLR). The differences in each specific model in the prediction of the ADI of the pesticides were discussed. Results: The stepwise multiple linear regression analysis resulted in a statistically significant QSAR model with five descriptors. Resultant QSAR models were robust, showing good utility across multiple classes of pesticide compounds. The applicability domain was also defined. The proposed models were robust and satisfactory. Conclusions: The QSAR model could be a feasible and effective tool for predicting ADI and for the comparison of logADIEPA to logADIWHO. The statistical results agree with the fact that USEPA focuses on more subtle endpoints than does WHO.

Quantitative Structure–Property Relationship Study to Predict Speed of Sound in Diverse Organic Solvents from Solvent Structural Information

The interaction of solvents with ultrasonic waves is of drastic importance and has been the subject of many studies in recent years. In this study, the effect of solvent structural parameters on the speed of sound in chemical solvents was investigated through a quantitative structure–property relationship (QSPR). Genetic algorithm–multiple linear regression (GA-MLR) analysis was employed to select the most relevant subset of descriptors and, then, to develop the model. The validity of the obtained 10-parameter model was assessed by most widely used validation techniques. The predictive power of the model was evaluated by use of an external data set. The high level of accuracy of results approved the model. According to the model, those solvents that have stronger solvent–solvent interactions can create a more appropriate medium for passing and propagating sound waves and will result in higher speed of sounds.

QSAR modeling of antimalarial activity of urea derivatives using genetic algorithm–multiple linear regressions

A quantitative structure–activity relationship (QSAR) was performed to analyze antimalarial activities of 68 urea derivatives using multiple linear regressions (MLR). QSAR analyses were performed on the available 68 IC50 oral data based on theoretical molecular descriptors. A suitable set of molecular descriptors were calculated to represent the molecular structures of compounds, such as constitutional, topological, geometrical, electrostatic and quantum-chemical descriptors. The important descriptors were selected with the aid of the genetic algorithm (GA) method. The obtained model was validated using leave-one-out (LOO) cross-validation; external test set and Y-randomization test. The root mean square errors (RMSE) of the training set, and the test set for GA–MLR model were calculated to be 0.314 and 0.486, the square of correlation coefficients (R2) were obtained 0.801 and 0.803, respectively. Results showed that the predictive ability of the model was satisfactory, and it can be used for designing similar group of antimalarial compounds.

Quantitative Near-Infrared Spectrum-Active Blood Efficacy Relationship of Compound Danshen Tablet with Genetic Algorithm–Multiple Linear Regression

ABSTRACT The near-infrared spectroscopy (NIRs) of compound Danshen tablet was scanned and the active blood efficacy of each sample was investigated in mouse. The quantitative spectrum-efficacy relationship (QSER) between NIR and the active blood efficacy was constructed by genetic algorithm-multiple linear regression (GA-MLR). The statistic parameters of the proposed QSER model were R2 = 0.891, , and RMSETrain = 0.085 for the training set while r2 = 0.817 and RMSEEXT = 0.116 for the external test set. The results indicate that the proposed QSER model could be applied to predict the active blood efficacy of compound Danshen tablet from the related NIR spectra.

Exploring Novel QSRRs for Simulation of Gas Chromatographic Retention Indices of Diverse Sets of Terpenoids in Pistacia Lentiscus L. Essential Oil Using Stepwise and Genetic Algorithm Multiple Linear Regressions

The present work describes convenient and interpretable models by step-wise multiple linear regression (SW-MLR)and genetic algorithm- multiple linear regression (GA-MLR). These quantitative structure-retention relationship (QSRR) based strategies have been successfully used to predict the retention indices (RIs) of a series of natural compounds found in the essential oil of Pistacia lentiscus L. The dataset was divided into training set (51 compounds) and test set (25 compounds), randomly. The prediction capabilities of both approaches have been appraised by their exertion to test set compounds. The fitness statistics of the models were assessed to be satisfactory which resulted in accurate predictions. The best optimized models for SW-MLR and GA-MLR consisted 5 (X0, Qtot, Dv, HATS3e and MATS4e) and 6 (MATS2e, DELS, ATS4e, PW5, PCD and W) molecular descriptors, respectively. Our investigation revealed the superiority of SW-MLR (R2=0.96, and for training set; REP=3.7 for test set) standpoint against the best GA-MLR model (R2=0.955, for training set; REP=7.5 for test set) for estimation of RIs of similar or unknown compounds.

Quantitative Structure-Activity Relationship Study of Amino Acid Derivatives as Histone Deacetylase Inhibitors using the Genetic Algorithm – Multiple Linear Regression

The quantitative structure-activity relationship (QSAR) of the amino acid derivatives for prediction of histone deacetylase inhibitors activity was studied. Semi-empirical quantum chemical calculations at the AM1 level were used to find the 3D geometry of the studied molecules. The suitable set of the molecular descriptors was calculated, and the important descriptors were selected by the genetic algorithm (GA). The predictive quality of the QSAR model was tested for an external set of eight compounds, randomly chosen out of 38 compounds. The genetic algorithm-multiple linear regression (GA-MLR) model with six selected descriptors was obtained. The accuracy of the proposed GA-MLR model was illustrated using to following evaluation techniques: cross-validation, validation through an external test set, and Y-randomization. This model, demonstrating high statistical qualities .The predictive ability of the model was found to be satisfactory and could be used for predicting the histone deacetylase inhibitors activity of the molecules with low prediction error.

Quantitative Structure–Activity Relationships and Docking Studies of Calcitonin Gene‐Related Peptide Antagonists

Defining the role of calcitonin gene-related peptide in migraine pathogenesis could lead to the application of calcitonin gene-related peptide antagonists as novel migraine therapeutics. In this work, quantitative structure-activity relationship modeling of biological activities of a large range of calcitonin gene-related peptide antagonists was performed using a panel of physicochemical descriptors. The computational studies evaluated different variable selection techniques and demonstrated shuffling stepwise multiple linear regression to be superior over genetic algorithm-multiple linear regression. The linear quantitative structure-activity relationship model revealed better statistical parameters of cross-validation in comparison with the non-linear support vector regression technique. Implementing only five peptide descriptors into this linear quantitative structure-activity relationship model resulted in an extremely robust and highly predictive model with calibration, leave-one-out and leave-20-out validation R(2) of 0.9194, 0.9103, and 0.9214, respectively. We performed docking of the most potent calcitonin gene-related peptide antagonists with the calcitonin gene-related peptide receptor and demonstrated that peptide antagonists act by blocking access to the peptide-binding cleft. We also demonstrated the direct contact of residues 28-37 of the calcitonin gene-related peptide antagonists with the receptor. These results are in agreement with the conclusions drawn from the quantitative structure-activity relationship model, indicating that both electrostatic and steric factors should be taken into account when designing novel calcitonin gene-related peptide antagonists.

Prediction of gas chromatographic retention indices of some amino acids and carboxylic acids from their structural descriptors

In this work, quantitative structure-retention relationship (QSRR) approaches were applied for modeling and prediction of the retention index of 282 amino acids (AAs) and carboxylic acids (CAs). Descriptors that were used to encode structural features of molecules in a data set were calculated by using the Dragon software. The genetic algorithm (GA) and stepwise multiple linear regression (MLR) methods were used to select the most relevant descriptors. Then support vector machine (SVM), artificial neural network (ANN) and multiple linear regression were utilized to construct nonlinear and linear quantitative structure-retention relationship models. The obtained results using these techniques revealed that nonlinear models were much better than other linear ones. The GA-ANN model has the average absolute relative errors (AARE) of 0.054, 0.059 and 0.100 for training, internal and external test set. Applying the tenfold cross-validation procedure on GA-AAN model obtained the statistics of Q(2)=0.943, which revealed the reliability of this model.

Prediction of Solubility of Fullerene C60 in Various Organic Solvents by Genetic Algorithm-Multiple Linear Regression

Quantitative structure property relationship (QSPR) study is presented for modeling and predicting of solubility of fullerene (C60) in various solvents. A data set consisting of 36 benzene derivatives is used in this study. Various kinds of molecular descriptors were calculated to represent the molecular structures of compounds and the best-fitting descriptors were selected by using stepwise multiple linear regressions (SW-MLR) and a genetic algorithm (GA-MLR) the selection of variables. The models were validated using leave-one-out (LOO), leave-multiple-out (LMO) cross-validation, external test set and Y-randomization test. The outliers were also examined to understand better in which cases large errors were to be expected and to improve the predictive models. Comparison of the results obtained indicated the superiority of the genetic algorithm over the stepwise. Also electronegativity, dispersion interaction in solution and volume of molecule were the main independent factors contributing to the solubility of fullerenes in the studied solvents.