Use of quantitative-structure property relationship (QSPR) and artificial neural network (ANN) based approaches for estimating the octanol-water partition coefficients of the 209 chlorinatedtrans-azobenzene congeners

Abstract

Polychlorinated azobenzenes (PCABs) can be found as contaminant by products in 3,4-dichloroaniline and its derivatives and in the herbicides Diuron, Linuron, Methazole, Neburon, Propanil and SWEP. Trans congeners of PCABs are physically and chemically more stable and so are environmentally relevant, when compared to unstable cis congeners. In this study, to fulfill gaps on environmentally relevant partitioning properties of PCABs, the values of n-octanol/water partition coefficients (log K(OW)) have been determined for 209 congeners of chloro-trans-azobenzene (Ct-AB) by means of quantitative structure-property relationship (QSPR) approach and artificial neural networks (ANN) predictive ability. The QSPR methods used based on geometry optimalization and quantum-chemical structural descriptors, which were computed on the level of density functional theory (DFT) using B3LYP functional and 6-311++G basis set in Gaussian 03 and of the semi-empirical quantum chemistry method (PM6) of the molecular orbital package (MOPAC). Polychlorinated dibenzo-p-dioxins (PCDDs), -furans (PCDFs) and -biphenyls (PCBs), to which PCABs are related, were reference compounds in this study. An experimentally obtained data on physical and chemical properties of PCDD/Fs and PCBs were reference data for ANN predictions of log K(OW) values of Ct-ABs in this study. Both calculation methods gave similar results in term of absolute log K(OW) values, while the models generated by PM6 are considered highly efficient in time spent, when compared to these by DFT. The estimated log K(OW) values of 209 Ct-ABs varied between 5.22-5.57 and 5.45-5.60 for Mono-, 5.56-6.00 and 5.59-6.07 for Di-, 5.89-6.56 and 5.91-6.46 for Tri-, 6.10-7.05 and 6.13-6.80 for Tetra-, 6.43-7.39 and 6.48-7.14 for Penta-, 6.61-7.78 and 6.98-7.42 for Hexa-, 7.41-7.94 and 7.34-7.86 for Hepta-, 7.99-8.17 and 7.72-8.20 for Octa-, 8.35-8.42 and 8.10-8.62 for NonaCt-ABs, and 8.52-8.60 and 8.81-8.83 for DecaCt-AB. These log K(OW) values shows that Ct-ABs are compounds of relatively low environmental mobility (log K(OW) > 4.5) and of significant bioaccumulation potential.

Dual Self-Adaptive Intelligent Optimization of Feature and Hyperparameter Determination in Constructing a DNN Based QSPR Property Prediction Model

“…To forecast the solubility of organic molecules in water in terms of their K ow values, Jiang et al leveraged the radial basis function (RBF) neural network and the molecular bond connectivity index . Wilczyńska-Piliszek et al utilized the QSPR approach and artificial neural networks (ANNs) combined with geometry optimization and quantum-chemical structural descriptors to estimate the values of log K ow for 209 congeners of chloro- trans -azobenzene (Ct-AB) . In other related works, QSPR techniques have been combined with the genetic algorithm (GA) to select the most relevant molecular descriptors and eventually to predict log K ow of sulfa drugs. − …”

“…14 Wilczynśka-Piliszek et al utilized the QSPR approach and artificial neural networks (ANNs) combined with geometry optimization and quantum-chemical structural descriptors to estimate the values of log K ow for 209 congeners of chloro-trans-azobenzene (Ct-AB). 15 In other related works, QSPR techniques have been combined with the genetic algorithm (GA) to select the most relevant molecular descriptors and eventually to predict log K ow of sulfa drugs. 16−18 It is interpreted from the discussed literature survey that the development of QSPR models needs a step in which the appropriate number and types of molecular descriptors are selected.…”

Artificial Neural Network in Drug Delivery and Pharmaceutical Research

“…Clark and co-workers used a data set containing 1085 compounds for developing a neural network model for octanolpartition coefficient prediction from the results of semi empirical AM1 calculations [81] and Eros et al developed neural network (fitting and prediction errors were s = 0.48 and s = 0.72 respectively)from database of 625 molecules, 98% of which are registered API showing high structural diversity [82]. Similarly, more recent studies based on QSPR ANN model predicted octanol-water partition coefficients for 209 chlorinated trans-azobenzene derivatives, contaminants in herbicides [83]. QSPR ANN model was also used by Noorizadeh et al to calculate the polar surface area of 32 drug molecules [84].…”