From many tools to one: an all-in-one R pipeline for untargeted NPS screening in LC-HRMS/MS data

The novel psychoactive substances (NPS) market is evolving at an unprecedented pace, continuously introducing new compounds that analytical workflows must be able to detect. Currently, several platforms are available to support researchers in NPS identification; however, each platform is designed to perform a single task, requiring analysts to use multiple tools in series to obtain a comprehensive result. This fragmented approach increases operator workload and extends the time required to produce a reliable outcome. The present work describes an R-based script specifically developed for untargeted NPS screening, which integrates into a single automated pipeline several analytical tasks that currently require multiple separate platforms. The script applies exclusion filters to distinguish compounds of analytical interest from background noise, and incorporates a signal rescue mechanism capable of recovering weak signals — including trace-level compounds — that would otherwise be discarded by standard quality criteria. Of particular relevance is the retrospective metabolic search functionality, which reconstructs putative parent masses from unknown compounds. This approach enables compound identification even when the substance has a very short half-life or when the sample was collected hours after consumption, at a point when the parent drug has been fully metabolized. The script is fully configurable by the operator through a dedicated parameter section, requiring no programming expertise, and can be adapted to diverse analytical and laboratory requirements. All processes have been developed with optimization of both analytical performance and processing time as primary objectives. The tool operates locally, making it particularly suitable for laboratories that, due to data privacy constraints, cannot use cloud-based platforms for sample processing.