Adduct-Induced Variability in Tandem Mass Spectrometry.

Abstract

Tandem mass spectrometry (MS/MS) provides essential structural information and plays a central role in compound annotation in metabolomics. While different precursor ion types are expected to influence the generation of MS/MS spectra, systematic investigations into precursor ion type-dependent MS/MS variability have been limited. To address this gap, we analyzed over half a million MS/MS spectra of 24,686 unique compounds from the NIST 20 spectral library, covering a broad range of precursor ion types and collision energies (CEs). Using [M + H]+ and [M - H]- spectra as references, we found that alkali cation adducted species such as [M + Na]+ and [M + K]+ exhibited distinct fragmentation behavior and low spectral similarity, likely due to the distinct nature of the alkali charge carriers, which do not promote protonated fragmentation pathways but instead stabilize the precursor ion through coordination. In contrast, [M + NH4]+, [2M + H]+, [M + H - H2O]+, [M + Cl]-, [2M - H]-, and [M - H - H2O]- showed moderate to high similarity to their references, as they often undergo neutral losses that generate [M + H]+ or [M - H]-, or are themselves derived from these ions. Our study also observed that fragmentation is structure-driven at lower CE and energy-driven at higher CE. This pattern allows for a higher spectral similarity among different precursor ion types at high CE. However, [2M + H]+ or [2M - H]- showed reduced similarity at higher CE, likely because the same amount of energy is distributed across more bonds in these larger precursor ions, resulting in less energy per bond. Finally, we demonstrated that ignoring precursor ion types can compromise compound annotation, including spectral library searches, molecular networking, and machine learning model development. Overall, this study underscores the critical influence of precursor ion types on MS/MS spectra and highlights the need for precursor-ion-type-aware strategies in metabolite annotation, which has been largely overlooked in the metabolomics field.