The Dark Metabolome: A Figment of Our Fragmentation?

Mass spectrometry, specifically liquid chromatography-tandem mass spectrometry (LC-MS/MS), has revealed hundreds of thousands to millions of metabolites that still need to be characterized. This unknown and extensive collection has been dubbed the “dark metabolome” and it represents more than 98 percent of observed LC-MS/MS spectra. On the face of it, we appear to be far from fully describing the human metabolome. But what if this discrepancy can be explained not by biology but by technology? 

The latter may be closer to the truth according to research from Martin Giera – Center for Proteomics and Metabolomics, Leiden University Medical Center, Netherlands – and Aries Aisporna, Winnie Uritboonthai, and Gary Siuzdak – all from the Scripps Center of Metabolomics and Mass Spectrometry, USA (1).

Several groups have observed a phenomenon called in-source fragmentation (ISF), which relates to the fragmentation of analytes during the initial ionization process within the ESI source and hence before the collision cell. “ISF basically generates a forest from a tree, in other words a single analyte can be presented as a molecular ion and one or many fragments,” write the authors in their Nature Metabolism Correspondence paper. “The employed mass analyzer will blindly isolate and fragment (again) whatever is being sent into the collision cell.” Given this fact, it appeared plausible to the researchers that ISF might partially be responsible for the dark metabolome.

To test their hypothesis, the team mined the METLIN MS/MS database – consisting of more than 930,000 molecular standards – at 0 eV, an energy designed to simulate the fragmentation observed in ISF. They found that ISF could account for over 70 percent of the peaks observed in typical LC-MS/MS metabolomic datasets. 

Put another way: “This finding disrupts the prevailing assumption that the majority of peaks in mass spectra correspond to unique metabolites,” write the authors. “Instead, it suggests that the spectra may be significantly populated by fragment ions generated during the ionization process. Such fragments, if not recognized as such, could be misclassified as distinct molecular entities, thus artificially expanding the metabolome’s perceived complexity.”

Siuzdak and Giera believe the findings could have important implications for anyone doing MS-based analysis of small molecules.

“Mass spectrometry practitioners need to be more rigorous in verifying the identity of the molecules they detect,” they say. “For example, when multiple peaks share identical retention times and peak shapes, the smaller (lower m/z) ions should be scrutinized as potential in-source fragments. This approach will help distinguish real molecules from fragments, ultimately reducing the perceived complexity of LC/MS data.” 

The authors believe that by implementing these identification steps, researchers can avoid misinterpretations and focus on the true molecular species present in their samples. 

And it’s not just LC-MS. “It is also worth noting that, since matrix-assisted laser desorption/ionization (MALDI) is generally considered a higher energy ionization source, as compared to electrospray, MALDI based imaging will also be subjected to significant ISF of small molecules.” 

The findings could be particularly important for researchers in the metabolic networking field, which involves mapping out the complex interactions and pathways of metabolites within biological systems. “Previously, the assumption was that all detected peaks represented intact molecules (as most networking approaches do). This leads to erroneous network interpretations due to the inclusion of data from these in-source fragments artifacts,” say Siuzdak and Giera. “Our study highlights the necessity for more selective data interpretation, ensuring that only real molecules are considered. This shift will enhance the accuracy of metabolic networks, avoiding distortions from non-existent molecules and providing a clearer, more reliable map of metabolic interactions.”

Feedback from colleagues, particularly biologists using LC-MS, have given Siuzdak and Giera confidence in their conclusions: “They have long been puzzled by the excessive number of peaks observed in LC-MS experiments and are pleased to now have a logical explanation for this phenomenon. The identification of in-source fragmentation as a major contributing factor to the overabundance of peaks provides a coherent explanation that aligns with their observations and experiences.”

In their paper, the authors note one positive outcome of their findings: the metabolome appears far more defined than earlier assumptions indicated. How defined? “This is a complex question, and one I approach with caution,” says Siuzdak. “The metabolome is a dynamic and variable entity, especially when factors such as diet and the biological diversity of plants and lipids are considered. A personal anecdote illustrates this well: during a period of frequent blood analyses, a dramatic change in my blood lipidome was noted after consuming a meal rich in lipids (guacamole). This highlighted how dietary intake can significantly alter the metabolome.”

Over time, simplifying the chemical landscape by excluding non-existent molecules could help everyone move forward faster. “Both biologists and chemists can benefit from this simplification, as it provides a clearer, more accurate picture of not only metabolic states but also anywhere LC-MS is used for small molecule analysis.”