Predicting Collision-Induced-Dissociation Tandem Mass Spectra (CID-MS/MS) Using Ab Initio Molecular Dynamics

Area of Science:

• Computational Chemistry
• Metabolomics
• Mass Spectrometry

Background:

• Compound identification in metabolomics relies on matching experimental mass spectra to libraries.
• Most compounds lack commercial availability, necessitating MS/MS spectra prediction.
• Existing prediction models, except for lipids, have limited success.

Purpose of the Study:

• To present the Collision-Induced Dissociation via Molecular Dynamics (CIDMD) framework for predicting MS/MS spectra.
• To model the physical process of molecular collisions in CID tandem mass spectrometry using first-principles molecular dynamics.
• To assess the accuracy of CIDMD predictions against experimental MS/MS spectra for small metabolites.

Main Methods:

• Simulating molecular ion collisions with argon atoms using density functional theory and molecular dynamics.
• Modeling bond breakages over time (≥1,000 fs) from multiple collisional directions.
• Generating in silico mass spectra by accumulating fragmentations from repeated simulations.

Main Results:

• CIDMD predicted 261 spectra for 12 small metabolites (<205 Da).
• Average entropy similarity scores reached 624 ± 189, improving to 828 ± 77 with optimal parameters.
• Prediction accuracy varied with protomers, collider velocities, and molecular mass; higher velocities improved results for larger molecules.

Conclusions:

• CIDMD is a viable tool for predicting collision-induced dissociation mass spectra of small metabolites.
• The framework's accuracy is influenced by protomer choice and simulation parameters.
• Understanding fragmentation pathways provides mechanistic insights into metabolite identification.