Prediction of the GC-MS retention time for terpenoids detected in sage (Salvia officinalis L.) essential oil using QSRR approach

Abstract

This work aimed to obtain a validated model for prediction of retention time of terpenoids isolated from sage herbal dust using supercritical fluid extraction. In total 32 experimentally obtained retention time of terpenes, which were separated and detected by GC-MS were further used to build a prediction model. The quantitative structure-retention relationship was employed to predict the retention time of essential oil compounds obtained in GC-MS analysis, using six molecular descriptors selected by a genetic algorithm. The selected descriptors were used as inputs of an artificial neural network, to build a retention time predictive quantitative structure-retention relationship model. The coefficient of determination for training cycle was 0.837, indicating that this model could be used for prediction of retention time values for essential oil compounds in sage herbal dust extracts obtained by supercritical fluid extraction due to low prediction error and moderately high r2. Results suggested that a 2D autocorrelation descriptor AATS0v was the most influential parameter with an approximately relative importance of 25.1 %.