Near-Infrared Spectroscopy and Machine Learning-Based Technique to Predict Quality-Related Parameters in Cannabis and Basil

Jakob Shimshoni, Food Safety and Quality, Agricultural Research Organization, Rishon LeZiyyon, Israel (jakobshimshoni@gmail.com)

Sweet basil and medicinal cannabis are among the most popular edible and medicinal annual herbs, respectively. The quantification of their pharmacologically active constituents is usually achieved by gas and liquid chromatography. These methods are expensive and time consuming, requiring highly trained personnel. Moreover, pre-and post-harvest mold identification and quantification is commonly achieved by inoculating cannabis swabs in culture plates, providing a lab result after 5-10 days. Consequently, the aim of the present study was to develop a simple, accurate, fast, and cost-effective analytical technique for the quantification of major essential oil components such as cannabinoids, terpenes, polyphenols and benzenoids using fourier transform near infra-red spectroscopy (FT-NIRS) coupled with multivariate classification and regression models. Furthermore, a modified FT-NIRS was developed for the first time for the identification and quantification of major fresh cannabis inflorescence molds such as Botrytis and Alternaria, revealing a great potential to replace traditional techniques for mold characterization.   

---

Short Biography of Presenting Author

Jakob Shimshoni is research scientist in the Department of Food Quality and Safety at the Agricultural Research Organization, Volcani Center, Israel. I earned my PhD from the Hebrew University of Jerusalem in 2009, and was a postdoctoral researcher at the NIH and University of Washington (2009-2011). My research group develops novel fluorescence and near-infrared methods coupled with multivariate classification and regression models to identify and quantify pharmacologically active secondary metabolites as well as pathogenic mold species in medicinal and edible herbs.