Answers are grounded in this paper and cite their sources.
Empirical data resource enhances spectral comparison and machine learning for small molecules, implying improved metabolite identification.
METLIN 960 K represents the largest collection of experimentally acquired small-molecule MS/MS spectra currently available. We introduce a reengineered publicaly accessible METLIN platform integrating high-resolution tandem mass spectrometry (MS/MS) data for over 960,000 empirically validated molecular standards. This scale was enabled by a high-throughput experimental framework integrating acoustic droplet ejection with high-throughput LC-MS/MS acquisition, allowing systematic empirical generation of MS/MS spectra from authentic standards. In addition to scale, METLIN 960 K provides a uniquely standardized MS/MS data set, with spectra acquired under controlled and consistent conditions across ionization modes and collision energies, enabling reproducible spectral comparison and machine-learning applications. Each compound is characterized by MS/MS spectra acquired in both positive and negative ionization modes across four collision energies (0, 10, 20, and 40 eV), enabling comprehensive fragmentation coverage and improved structural annotation. Designed as a reference library for XCMS-METLIN and compatible with machine-learning workflows, METLIN 960 K supports high-fidelity spectral matching, neutral loss analysis, and filtering of misannotations, including annotation of in-source fragments and biologically synchronized ranking of candidate metabolites. The platform also provides empirically derived MRM transitions on all standards (via METLIN-MRM), supporting quantitative method development across a chemically diverse range of metabolites, natural products, lipids, peptides, pharmaceuticals, and toxicants. A redesigned interface enables efficient querying by exact mass, formula, or structure with direct access to curated spectra and metadata. Two additional resources enhance identification: (1) METLIN Core, a high-frequency-use subset for rapid searching, and (2) > 1.02 million additional structures without MS/MS data for hypothesis generation. Derived exclusively from authentic standards, METLIN 960 K (https://metlin.scripps.edu) provides the largest publicly available empirical MS/MS database, delivering high-confidence annotation for both untargeted and targeted mass spectrometry workflows.
No takes yet. Share an insight, caveat, or question.
Uritboonthai et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: