Evaluating the Robustness and Ruggedness of a Statistical Method for Comparison of Mass Spectral Data for Seized Drug Identification

Identification of seized drugs often involves visual comparison of mass spectra of the submitted sample and a suitable reference. However, the continued emergence of novel psychoactive substances (NPS) has resulted in a multitude of new, structurally similar, compounds, including positional isomers. The electron-ionization mass spectra of such compounds are so similar that identification based on visual comparison is not reliable.

To provide a more objective means for mass spectral comparison, our laboratory previously developed a statistical-based method. The method uses the unequal variance t-test to compare ion intensities at corresponding m/z values across the scan range. In cases where there is no significant difference in intensity at any m/z value, the two spectra are considered to be statistically indistinguishable. In contrast, if there is a significant difference in intensity at any m/z value, the two spectra are statistically distinguishable. Successful application of the method has been demonstrated for association and discrimination of amphetamine-type stimulants and for salvinorins extracted from Salvia divinorum.

The proposed work seeks to fully evaluate the robustness and ruggedness of the statistical comparison method for implementation in operational forensic laboratories. A set of 21 NPS, including six sets of positional isomers, will be used in this work to provide a rigorous test of method capabilities. Spectra of isomers will be statistically compared, assessing the effect of concentration, mass spectral tune parameters, and different GC-MS instruments on spectral quality and, ultimately, the rate of false positive and false negative identifications. Through partnerships with operational forensic science laboratories, the method will be more thoroughly evaluated, taking into account the effect of different analysts, methods, and instruments on the rate of false positive and false negative identifications.

The results from this work will be disseminated broadly in a number of ways. First, results of the robustness and ruggedness studies will be presented at regional and national forensic science conferences and published in peer-reviewed forensic science journals. A workshop will then be presented to train analysts in the theory and application of the method. Finally, toward the end of the project, a webinar-based training will also be presented to reach a wider audience. All collected data, conference presentations, peer-reviewed publications, and webinar materials will be submitted to NIJ, along with the final research summary. The proposed research will provide a validated tool for the objective comparison of mass spectral data to increase confidence in seized drug identifications. Note: This project contains a research and/or development component, as defined in applicable law, and complies with Part 200 Uniform Requirements - 2 CFR 200.210(a)(14). CA/NCF