Random Forest Processing of Direct Analysis in Real-Time Mass Spectrometric Data Enables Species Identification of Psychoactive Plants from Their Headspace Chemical Signatures.
ACS omega September 24, 2019 Meghan Grace Appley, Samira Beyramysoltan, Rabi Ann Musah 18 citations
Several plant-based 'legal highs' designated by the United Nations as plants of concern can be reliably identified by their headspace chemical profiles. Using direct analysis in real-time high-resolution mass spectrometry (DART-HRMS), the headspace volatiles of 15 species—including Cannabis sativa, Mitragyna speciosa, and Salvia divinorum—were analyzed. Intraspecies similarities and interspecies differences were observed. Supervised random forest classification achieved 99% accuracy in distinguishing species. A conformal predictor based on this classifier was valid at an 8% significance level with an error rate of 0. The technique demonstrates proof-of-concept for a database to detect and identify plant-based legal highs through headspace analysis.