Systematic pre-annotation explains the "dark matter" in LC-MS metabolomics

Area of Science:

• Analytical Chemistry
• Biochemistry
• Metabolomics

Background:

• Global metabolomics using high-resolution mass spectrometry generates numerous unidentified features, termed "dark matter".
• Distinguishing real compounds from artifacts like in-source fragments is crucial for data interpretation and field advancement.

Purpose of the Study:

• To systematically analyze the nature of unidentified features in large-scale liquid chromatography-mass spectrometry (LC-MS) metabolomics datasets.
• To investigate the contribution of in-source fragments to the metabolomic "dark matter".

Main Methods:

• Analysis of 61 diverse public LC-MS metabolomics datasets.
• Application of Khipu-based pre-annotation to assess ion patterns of abundant features.

Main Results:

• In-source fragments constitute less than 10% of features in LC-MS metabolomics data.
• The majority of abundant features exhibit identifiable ion patterns, suggesting they originate from real compounds.
• The number of unique compounds is significantly smaller than the total number of detected features.

Conclusions:

• The metabolomic "dark matter" in LC-MS is largely composed of real compounds, with abundance correlating to feature identification.
• Most detected features represent known compounds, but the majority of these compounds remain unidentified.
• The findings suggest a vast pool of yet-to-be-identified compounds in metabolomics research.