Python Pipeline, Mass Spectrometry, and Quantitative Structure‐Retention Relationship Models: A Powerful Trio for Metabolite Characterization

ABSTRACT Data processing using high‐resolution mass spectrometry (HR‐MS) for drug metabolite characterization remains challenging and time‐consuming. Commercial software tools are costly and lack cross‐platform generality. Furthermore, potential isomeric metabolites are difficult to characterize solely based on mass fragment ions, necessitating complementary approaches for unambiguous identification. We developed a hybrid workflow that combines a Python‐based MS data processing tool, molecular networking, and quantitative structure‐retention relationship (QSRR) models to enable comprehensive drug metabolite characterization. This package integrated several data mining algorithms for drug metabolite characterization, including the mass defect filter, nitrogen rule filter, dot‐product algorithm, and neutral loss filter. The relationships between the fragment ions of the parent drugs and their metabolites were elucidated and confirmed using a visual molecular network. The characterized isomeric metabolites were further discriminated based on their predicted retention times using a QSRR model. The metabolites of the model drugs, midazolam, diclofenac, and testosterone, were characterized using this hybrid workflow. In total, 7, 12, and 11 metabolites were identified, respectively. All of these matched the reported metabolites, confirming the feasibility of the hybrid strategy. A powerful trio integrating a Python‐based MS data‐processing tool, molecular networks, and QSRR models was developed for rapid and visual drug metabolite characterization.