• Bulletin of the Korean Chemical Society
• /
• v.33 no.2
• /
• pp.613-619
• /
• 2012

We applied several machine learning methods for developing QSAR models for prediction of acute toxicity to fathead minnow. The multiple linear regression (MLR) and artificial neural network (ANN) method were applied to predict 96 h $LC_{50}$ (median lethal concentration) of 555 chemical compounds. Molecular descriptors based on 2D chemical structure were calculated by PreADMET program. The recursive partitioning (RP) model was used for grouping of mode of actions as reactive or narcosis, followed by MLR method of chemicals within the same mode of action. The MLR, ANN, and two RP-MLR models possessed correlation coefficients ($R^2$) as 0.553, 0.618, 0.632, and 0.605 on test set, respectively. The consensus model of ANN and two RP-MLR models was used as the best model on training set and showed good predictivity ($R^2$=0.663) on the test set.

• Environmental Mutagens and Carcinogens
• /
• v.21 no.2
• /
• pp.118-121
• /
• 2001

The predictive screening of various molecular descriptors for predicting carcinogenic, mutagenic, teratogenic and alkylation activity of chlorinated disinfection byproducts (DBPs) has been investigated for the application of quantitative structure-activity relationships (QSAR). The toxicity index for 29 compounds were computed by the PASS program and active values were employed in this study. Studies show that different descriptors account for the model equation of each genotoxic endpoint and that thermodynamic descriptors significantly played a major role on prediction of endpoints of chlorinated aliphatic compounds.

• Advances in environmental research
• /
• v.1 no.4
• /
• pp.257-276
• /
• 2012

QSAR models for chlorinated alkenes were developed between $E_{HOMO}$ and their chlorine and carbon content. The aim is to provide valid QSAR model which is statistically validated for $E_{LUMO}$ prediction. Different molecular descriptors, $N_{Cl}$, $N_C$ and $E_{HOMO}$ have been used to take into account relevant information provided by molecular features and physicochemical properties. The best model were selected using Partial Least Square (PLS) and Multiple Linear Regression (MLR) led to models with satisfactory predictive ability for a data set of 15 chlorinated alkene compounds.

• Environmental Analysis Health and Toxicology
• /
• v.16 no.3
• /
• pp.115-120
• /
• 2001

The predictive screening of various molecular descriptors for predicting cyclooxygenase inhibitor, lipooxygenase inhibitor, leucotriene synthesis inhibitor, leucotriene antagonist activities of Stilbene moieties have been investigated for the application of quantitative structure-activity relationships (QSAR). The biological activities for 36 compounds were computed by the PASS program and molecular descriptors are cited from literatures or calculated, to investigate feasibility of screening relevant descriptors and of their applications among biological endpoints. Fairly good correlations varying from 0.7828 to 0.9032 were obtained using 12 descriptors with 29 Stilbene derivatives and 5 diazo-compounds. Our studies reveal that LogKow, electron density(X), electron density (Y),4th-order valence connectivity and water solubility can be usefully employed to predict biological activities of stilbene derivatives with simple regression models.

• Applied Biological Chemistry
• /
• v.41 no.6
• /
• pp.477-482
• /
• 1998

In order to discover new Leukotriene $D_4$ antagonists, Quantitative Structure-Activity Relationships (QSAR) were applied based on the known data. A series of chalcone derivatives were selected for the training set. A candidate was predicted using QSAR and synthesized, and its biological activity was tested.

• Journal of Integrative Natural Science
• /
• v.4 no.3
• /
• pp.216-221
• /
• 2011

c-Jun N-terminal kinase-3 (JNK-3) has been shown to mediate neuronal apoptosis and make the promising therapeutic target for neurodegenerative diseases such as Parkinson's disease, Alzheimer's disease, and other CNS disorders. In order to better understand the structural and chemical features of JNK-3, comparative molecular field analysis (CoMFA) was performed on a series of 3,5-disubstituted quinolines derivatives. The best predictions were obtained CoMFA model ($q^2$=0.707, $r^2$=0.972) and the statistical parameters from the generated 3D-QSAR models were indicated that the data are well fitted and have high predictive ability. The resulting contour map from 3D-QSAR models might be helpful to design novel and more potent JNK3 derivatives.

• Journal of Integrative Natural Science
• /
• v.4 no.3
• /
• pp.210-215
• /
• 2011

Holographic quantitative structure-activity relationships (HQSAR) is a useful tool to correlates structures with their biological activities. HQSAR is a two dimensional (2D) QSAR methodology, which generates QSAR equations through 2D fingerprint and correlates it with biological activity. Here, we report a 2D-QSAR model for a series of fifty-one 3,4-dihydroxychalcones derivatives utilizing HQSAR methodology. We developed HQSAR model with 6 optimum numbers of components (ONC), which resulted in cross-validated correlation coefficient ($q^2$) of 0.855 with 0.283 standard error of estimate (SEE). The non-cross-validated correlation coefficient (r2) with 0.966 indicates the model is predictive enough for analysis. Developed HQSAR model was binned in to a hologram length of 257. Atomic contribution map revealed the importance of dihydroxy substitution on phenyl ring.

• Journal of the Korean Chemical Society
• /
• v.59 no.6
• /
• pp.483-487
• /
• 2015

The different calculated quantum chemical descriptors by DFT method were used for prediction of some sulphanilamide Schiff bases inhibitor activity as a binding constant (log K). Multiple linear regression (MLR) and artificial neural network (ANN) were employed for developing the useful quantitative structure activity relationship (QSAR) model. The obtained results presented superiority of ANN model over the MLR one. The offering QSAR model is very easy to computation and Physico-Chemically interpretable. Sensitivity analysis was used to determine the relative importance of each descriptor in ANN model. The order of importance of each descriptor according to this analysis is: molecular volume, molecular weight and dipole moment, respectively. These descriptors appear good information related to different structure of sulphanilamide Schiff bases can participate in their inhibitor activity.

• YAKHAK HOEJI
• /
• v.59 no.4
• /
• pp.151-157
• /
• 2015

The purpose of this study is to demonstrate how lead compounds are best optimized with the application of in silico QSAR and PBPK modeling at the early drug discovery stage. Several predictive QSAR models such as $IC_{50}$ potency model, intrinsic clearance model and brain penetration model were built and applied to a set of virtually synthesized library of the BACE1 inhibitors. Selected candidate compounds were also applied to the PBPK modeling for comparison between the predicted animal pharmacokinetic parameters and the observed ones in vivo. This novel lead optimization strategy using QSAR and PBPK modelings could be helpful to expedite the drug discovery process.

• Bulletin of the Korean Chemical Society
• /
• v.25 no.12
• /
• pp.1874-1876
• /
• 2004

Quantitative structure activity relationship has been probed for spirosuccinimide-fused tetrahydropyrrolo[1,2-a]pyrazine-1,3-dione derivatives acting as aldose reductase inhibitors. While the spirosuccinimide compounds contain a chiral center, the aldose reductase inhibition assay was performed with racemic mixtures in the published work. As the physicochemical descriptors of the QSAR analysis must be evaluated for a definite molecular structure, we devise a new 'racemic' descriptor as the arithmetic mean of the (R)-enantiomer descriptor and the (S)-enantiomer descriptor. The resultant QSAR model derived from the racemic descriptors outperforms the original QSAR models, closely reproducing the observed activity of optically pure enantiomers as well as racemic mixtures.