Illuminating the Dark: Molecular Networking as a Novel Psychoactive Substance Identification Strategy

Illuminating the Dark: Molecular Networking as a Novel Psychoactive Substance Identification Strategy

Drug use is a considerable threat to the public good. Not only does drug use contribute to the societal burden of disease and death, it also generates crime, impedes community and family cohesion, and degrades the quality of our health, educational, and social systems. Recently, drug use patterns have changed and there has been a surge in novel psychoactive substance (NPS) use. As of December 2020, over 1,000 unique novel psychoactive substances have been reported to the United Nations Office on Drugs and Crime (UNODC) Early Warning Advisory. 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 algorithims are needed to mine untargeted mass spectral data and identify structural relationships and patterns without a spectral library. In pursuit of this, we propose 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. Our work will begin by developing and validating an untargeted sample preparation and screening method. This method will be validated for several forensically relevant compounds, focusing on the National Safety Council’s Tier I and II drugs and metabolites. Next, we will generate a molecular network where spectra of compounds with structural similarity are clustered together. We will then seed the molecular network with known compounds through traditional spectral library matching. This hybrid targeted/untargeted spectral annotation approach will enable 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 will curate and make public a drug of abuse and metabolite spectral library within an open access platform. Furthermore, our proposed datasets will also be made public for other research groups to build upon or annotate.