Screening Unknown Novel Psychoactive Substances Using GC-MS Based Machine Learning

Abstract

In recent years, there is a large increase in structural diversity of novel psychoactive substances (NPS), exacerbating drug abuse issues as these variants evade classical detection methods such as spectral library matching. Gas chromatography mass spectrometry (GC-MS) is commonly used to identify these NPS. To tackle this issue, machine learning models are developed to address the analytical challenge of identifying unknown NPS, using only GC-MS data. 891 GC-MS spectra are used to train and evaluate multiple supervised machine learning classifiers, namely artificial neural network (ANN), convolutional neural network (CNN) and balanced random forest (BRF). 7 classes, comprising 6 NPS classes (cathinone, cannabinoids, phenethylamine, piperazine, tryptamines and fentanyl) and other unrelated compounds can be effectively classified with a macro-F1 score ~ 0.9, averaged across all cross-validation folds. These results indicate that machine learning models are a promising complement as an effective NPS detection tool.

Application of gas chromatography in the analysis of phytocannabinoids: An update (2020–2023)

“…A GC-MS approach has recently been adopted for studying the effect of nitrogen deficiency on phytocannabinoids biosynthesis in medical and industrial Cannabis by facilitating a metabolic shift towards the production of low-N metabolites. 102 Various GC-FID and GC-MS methods have been employed for the analyses of different samples of marijuana or hashish (cannabis) [92][93][94][95][96][97] and marijuana blunts. 98 Chemical characterisation of marijuana blunt smoke by nontargeted chemical analysis was achieved by GC-MS analyses, where a TOF-MS was used.…”

“…A PhD research project carried out by Shipley 95 reported the validation of a quantitative analytical method to detect nine cannabinoids, CBD (3), CBG (5), CBN (7), Δ 8 -THC (9), THC (10), (6aR, 9R)-Δ 10 -THC, (6aR,9S)-Δ 10 -THC, 9(R)-Δ 6a,10a -THC, and 9(S)-Δ 6a,10a -THC, from marijuana samples for forensic drug analysis using GC-MS, while several phytocannabinoids could be detected in the seized Hashish samples as a part of screening unknown novel psychoactive substances using GC-MS-based machine learning. 96 A GC-MS method together with a thermal separation mode has recently been employed for the characterisation of Hashish samples obtained from the Egyptian illicit trafficking market, leading to the identification of several cannabinoids including CBC (1), CBD (3), CBE (16), CBG (5), CBN (7), Δ 8 -THC (9), THC (10), THCV (13), and exo-THC. 97 Radwan et al 103 have reported the isolation and characterisation of impurities in commercially marketed Δ 8 -THC (9)-containing products like vapes, gummies, and distillates, using GC-FID and GC-MS methods.…”

“…GC‐based analyses of cannabinoids in various consumer products, 61–103 published during 2020–2023, are summarised in Table 1. Phytocannabinoids, particularly CBD ( 3 ), have been claimed to offer some benefits to the human skin, and this beneficial property has recently generated a trend in the use of phytocannabinoids in various cosmetics and cosmeceutical products 70 .…”

“…Various GC‐FID and GC‐MS methods have been employed for the analyses of different samples of marijuana or hashish (cannabis) 92–97 and marijuana blunts 98 . Chemical characterisation of marijuana blunt smoke by nontargeted chemical analysis was achieved by GC‐MS analyses, where a TOF‐MS was used 98 .…”

“…A GC‐FID‐based analysis of 44 seized marijuana samples from the United States was reported, where seven phytocannabinoids were identified in the sampels 94 ; this study correctly predicted the geographical origins of 100% of Washington cannabis, 86% of both Oregon and Vermont samples, and 71% of Ohio samples. A PhD research project carried out by Shipley 95 reported the validation of a quantitative analytical method to detect nine cannabinoids, CBD ( 3 ), CBG ( 5 ), CBN ( 7 ), Δ 8 ‐THC ( 9 ), THC ( 10 ), (6aR, 9R)‐Δ 10 ‐THC, (6aR,9S)‐Δ 10 ‐THC, 9(R)‐Δ 6a,10a ‐THC, and 9(S)‐Δ 6a,10a ‐THC, from marijuana samples for forensic drug analysis using GC‐MS, while several phytocannabinoids could be detected in the seized Hashish samples as a part of screening unknown novel psychoactive substances using GC‐MS‐based machine learning 96 . A GC‐MS method together with a thermal separation mode has recently been employed for the characterisation of Hashish samples obtained from the Egyptian illicit trafficking market, leading to the identification of several cannabinoids including CBC ( 1 ), CBD ( 3 ), CBE ( 16 ), CBG ( 5 ), CBN ( 7 ), Δ 8 ‐THC ( 9 ), THC ( 10 ), THCV ( 13 ), and exo‐THC 97 .…”