Multilaboratory Untargeted Mass Spectrometry Metabolomics Collaboration to Identify Bottlenecks and Comprehensively...

Area of Science:

• Metabolomics
• Analytical Chemistry
• Bioinformatics

Background:

• Metabolomics enables the study of small molecules in biological systems.
• Accurate annotation of mass spectrometry data is crucial for identifying metabolites.
• Untargeted metabolomics aims to detect and identify a wide range of metabolites.

Purpose of the Study:

• To identify challenges in annotating untargeted mass spectrometry metabolomics datasets.
• To propose strategies for overcoming annotation difficulties.
• To assess the reproducibility and accuracy of metabolite annotation across different expert teams.

Main Methods:

• Analysis of ashwagandha (Withania somnifera) extract using liquid chromatography-mass spectrometry (LC-MS) on Orbitrap and Q-ToF platforms.
• Data-dependent acquisition (DDA) in positive ion mode.
• Annotation of 12 datasets by ten expert metabolomics teams.
• Cross-checking annotations to establish a consensus list of 142 putative analytes.

Main Results:

• Significant variability in analyte reporting among expert teams (24-57% of consensus list).
• Correct assignment of ion species (clusters, fragments) in MS spectra emerged as a major bottleneck.
• In-source redundancy was frequently mistaken for independent analytes, leading to annotation errors and overestimation of sample complexity.
• Only 13 out of 142 consensus analytes were confirmed by comparison with standards.

Conclusions:

• Current annotation strategies in untargeted metabolomics face significant challenges, particularly in spectral interpretation and redundancy handling.
• There is a critical need for improved computational tools and standardized approaches for feature identification and grouping.
• Enhanced methods for querying spectral and taxonomic databases are required to improve the accuracy of putative analyte structure assignment.