Quantitative structure-property relationship study of liquid vapor pressures for polychlorinated diphenyl ethers

A quantitative structure-property relationship (QSPR) study was performed to correlate the logarithmic values of liquid vapor pressure of polychlorinated diphenyl ethers (PCDEs) with their molecular structures. The entire set of 106 PCDEs and diphenyl ether was divided into a training set of 72 samples and a test set of 35 samples using the DUPLEX algorithm. An extended set of molecular descriptors was calculated to represent the molecular structures by DRAGON software. Multiple linear regression (MLR) was used to select descriptors and develop models. A three-descriptor equation was obtained for the training set, with a squared correlation coefficient (R2) of 0.997 and a standard error(s) of 0.069. The robustness and predictive performance of the proposed model were assessed by different approaches, including leave-many-out cross-validation, Y-randomization test, and external validation through test set. Satisfactory results of R2 = 0.995 and a mean absolute error of 0.066 for the test set confirmed the model being very useful to predict the vapor pressure of PCDEs. Furthermore, the applicability domain of the models was analyzed based on the Williams plot.