aug-MIA-QSPR on the modeling of sweetness values of disaccharide derivatives

Multivariate image analysis applied to quantitative structure-property relationships (MIA-QSPR) was used to model the sweetness values relative to sucrose, log (RS), of a series of 40 disaccharides. Two dimensional molecular shape and atomic sizes were capable of predicting the changes in log (RS), giving results of calibration (R-2 = 0.97) and validation (R-CV(2) = 0.86 and R-test(2) = 0.94) very satisfactory. A novel polyhalogenated derivative with high estimated sweetness was proposed and a glucophoric pattern based on the well-known AH-B-gamma system has been discussed. The technique avoids unnecessary time and cost-demanding procedures, since new compounds (e.g. food additives) can be designed and their property estimated prior to compound preparation. (C) 2012 Elsevier Ltd. All rights reserved.