MultiMS2: A curated multi-modal, multi-energy spectral library for metabolomics

Area of Science:

• Analytical Chemistry
• Metabolomics
• Computational Biology

Background:

• Spectral libraries are crucial for metabolite annotation in mass spectrometry-based metabolomics.
• Existing libraries primarily use collision-induced dissociation (CID), which is often insufficient for structural elucidation.
• Electron-activated dissociation (EAD) offers complementary fragmentation but is underrepresented in public datasets, limiting analytical capabilities and machine learning development.

Purpose of the Study:

• To create a comprehensive metabolomics spectral library integrating multiple dissociation methods and acquisition parameters.
• To expand the representation of electron-activated dissociation (EAD) data within public spectral libraries.
• To facilitate advancements in metabolite annotation, benchmarking, and machine learning models for metabolomics.

Main Methods:

• Acquisition of tandem mass spectrometry (MS/MS) spectra using both CID and EAD.
• Utilized three distinct energy levels for each dissociation method.
• Acquired spectra across both positive and negative ionization modes.
• Curated a dataset of 43,728 MS/MS spectra from 2,899 unique compounds.

Main Results:

• Developed MultiMS2, a spectral library containing 43,728 MS/MS spectra for 2,899 metabolites.
• Significantly increased the availability of EAD spectra under controlled conditions.
• Ensured matched acquisition parameters across different energies and dissociation types, enabling direct comparison.

Conclusions:

• MultiMS2 serves as a valuable resource for metabolite annotation, benchmarking, and machine learning in metabolomics.
• The library supports advanced analytical approaches, including energy-aware and dissociation-aware analyses.
• Facilitates methodological innovation and enhances the generalizability of computational metabolomics tools.