SMART: An Approach for Accurate Formula Assignment in Spatially Resolved Metabolomics

Area of Science:

• Metabolomics
• Biochemistry
• Analytical Chemistry

Background:

• Spatially resolved metabolomics is crucial for understanding tissue metabolism.
• Accurate metabolite annotation in mass spectrometry imaging (MSI) remains a challenge.
• Existing methods struggle to keep pace with the vast data generated by MSI.

Purpose of the Study:

• To introduce SMART, an open-source platform for precise formula assignment in mass spectrometry imaging.
• To bridge the annotation gap between MSI and LC-MS-based approaches.
• To enhance the systematic annotation of metabolites within biological tissues.

Main Methods:

• SMART utilizes a KnownSet database of 2.8 million formulas linked by ChemEdges and BioEdges.
• A multiple linear regression model is employed to extract formula networks.
• Candidate formulas are scored based on linked formulas, database scores, chemical/biological links, and ppm values.

Main Results:

• SMART achieves prediction accuracy rates up to 95.8% when benchmarked against reference datasets.
• The platform successfully annotated 986 formulas in developing mouse embryos using MSI.
• SMART demonstrates robust performance in systematic formula annotation.

Conclusions:

• SMART significantly improves the accuracy and efficiency of metabolite formula assignment in MSI.
• This platform enhances the understanding of metabolic heterogeneity in complex biological samples.
• SMART provides a robust solution for data analysis in spatially resolved metabolomics.