A Chemometric Model for the Rapid Determination of Phenolic Diterpenes in Domestic Rosemary
Thamann, Kathryn D.
MetadataShow full item record
Rosemary is a common household spice of interest due to its antioxidant properties, including specific phenolic diterpenes. These diterpenes are of interest to Kalsec because of their ability to act as a food preservative, thereby eliminating artificial preservatives in their products. This study seeks to improve upon the existing chromatographic method of determining phenolic diterpene content in domestic rosemary by conducting a chemometric analysis of spectral data. A chemometric calibration model was therefore constructed using partial least squares (PLS) regression comparing near infrared (NIR) data to ultra-performance liquid chromatography (UPLC) as a reference. With the inclusion of spectral pretreatments of standard normal variate (SNV) and a Savitzky-Golay second derivative, the model could accurately predict phenolic diterpene levels in a test set of rosemary within a 10% error and below an average prediction error of 1.200 diterpene levels. An internal full cross-validation additionally verified the model’s ability to predict new domestic rosemary samples. Future studies should include additional data points in the model to improve the predictions, as well as the inclusion of a moisture calibration to account for the effects of water in the samples.