Alternatively for mipdatabase quick searches go here

Custom Search

Abstract: Computational approach for evaluating the feasibility of template-monomer complexes has great potential in assisting the selection of appropriate functional monomers for template molecule of interest. A quantitative structure-property relationship (QSPR) study of template-monomer complexes was investigated for the prediction of imprinting factor of molecularly imprinted polymers (MIPs). The data set was based on uniformly-sized MIP particles taken from the literature and was used in our previous study for computing the imprinting factor using molecular descriptors derived from charge density-based electronic properties of molecules. In this study, we examined the feasibility of using quantum chemical descriptors and artificial neural networks for prediction of the imprinting factor. The proposed methodology reliably predicted the imprinting factor of MIPs with correlation coefficient from 0.7083 to 0.8378 albeit to a lesser degree than charge-based descriptors, which yielded correlation coefficient as high as 0.9680. The importance of mobile phase descriptors on the predictive performance of the QSPR model has surprisingly shown that the use of mobile phase descriptors alone was able to predict the imprinting factor with good performance
Author keywords: Imprinting factor, molecular imprinting, molecularly imprinted polymer, Neural network, MIP, QSPR