Approaches to identify novel psychoactive substances

As consumption of cannabinoids explodes with legalization and decriminalization, so too has the emergence of novel psychoactive substances (NPS). Since 2010, there has been an unprecedented increase in the number, type, and availability of new synthetic drugs of abuse. These novel compounds can be grouped as cannabinoids, cathinones, phenethylamines, opioids, tryptamines, benzodiazepines, arylalkylamines, and others. From a public health perspective, it is critical to identify which compounds are responsible for impairment, overdose, and death to educate officials, first responders, medical professionals, and the public. However, when novel compounds appear on the street, there is a lack of analytical data, pure standards, and studies on adverse effects, making their detection a major toxicological challenge. Furthermore, biodisposition data are lacking for novel drugs of abuse and their potencies vary considerably, making NPS concentrations difficult to interpret in any biological matrix.

Rapid proliferation of NPS is at least in part driven by the desire to circumvent forensic detection. Untargeted screening of psychoactive substances in biological matrices is one of the most difficult yet exciting goals of forensic toxicology. Admittedly, truly untargeted screening is still out of reach. Recent interest in untargeted high resolution mass spectrometry (HRMS) is promising but still relies on targeted data processing, reference standard availability, and dereplication through spectral library matching. Computational algorithms are needed to mine untargeted mass spectral data and identify structural relationships and patterns without a spectral library. In pursuit of this, we have a National Institute of Justice funded project (15PNIJ-21-GG-04171-COAP) leveraging open access molecular networking as a crowd-sourced spectral annotation strategy. Molecular networking organizes mass spectral data to produce maps of component structural similarity without any prior knowledge of the sample’s composition. We generate molecular networks where spectra of compounds with structural similarity are clustered together. We then seed the molecular networks with known compounds through traditional spectral library matching. This hybrid targeted/untargeted spectral annotation approach enables efficient processing of mass spectral data and rapid identification of NPS that are similar to known parent compounds. Molecular networking infrastructure is crowd-sourced so once new compounds are identified and validated, that information is immediately propagated through the entire molecular networking database. To facilitate adoption of molecular networking within the forensic toxicology community, we are curating and making public a drug of abuse and metabolite spectral library within an open access platform. Furthermore, our datasets will also be made public for other research groups to build upon or annotate.