A large-scale study has revealed that precursor ion type has a critical – and often overlooked – influence on tandem mass spectrometry (MS/MS) fragmentation patterns. By analyzing more than half a million spectra from the NIST 20 library, researchers from the University of British Columbia, Canada, found that some adducts, such as [M + Na]⁺, fragment very differently from protonated ions, while others show closer correspondence. These differences affect not only spectral library matching but also molecular networking, machine learning models, and interpretations of the “dark metabolome.”
We spoke with corresponding author Tao Huan to learn more about the motivation behind the study, its key discoveries, and the implications for metabolomics workflows.
What initially motivated your team to take a closer look at how precursor ion types shape MS/MS spectra?
Around a year ago, I reviewed a manuscript on best practices in metabolomics. It was a well-written piece, and I provided several suggestions to further strengthen its quality. Among its points, the authors briefly noted that precursor ion types can influence MS/MS fragmentation patterns. However, they offered no experimental evidence to support that statement. This gap made me realize that the effect of precursor ion type on MS/MS fragmentation had not been systematically investigated.
Notably, while most compound annotation platforms allow users to specify different precursor ion types for MS1-based candidate selection, typically MS/MS matching is performed only against reference spectra of [M + H]⁺ or [M − H]⁻ ions without considering their experimental precursor ion types. After raising this topic during a group meeting, one of my students, Botao Liu, expressed strong interest in pursuing it. Given our lab’s prior experience in conducting systematic MS/MS evaluations using the well-established NIST tandem MS library, we leveraged this expertise to analyze MS/MS similarity across different precursor ion types using a similar strategy.
One of your key findings is that [M + Na]⁺ fragments very differently from [M + H]⁺ – how does this affect compound annotation in untargeted metabolomics workflows?
The impact of this is wide-ranging. First and foremost, greater attention must be given to precursor ion type annotation – the MS/MS spectrum of [M + Na]⁺ is not simply that of [M + H]⁺ shifted by 22 Da (the mass difference between Na⁺ and H⁺). While this distinction is well understood by researchers with extensive mass spectrometry training, it may not be apparent to metabolomics practitioners from other disciplines. Without careful consideration of precursor ion types, the MS/MS of [M + Na]⁺ could be mistakenly interpreted as a unique spectrum and, consequently, misclassified as part of the so-called “dark metabolome.” Second, for compound annotation via spectral library searching, the substantial fragmentation differences between [M + Na]⁺ and [M + H]⁺ mean that direct library matching is prone to failure. Similarly, spectral similarity analysis and molecular networking can fail. This is because the unique fragmentation pattern of [M + Na]+ ions differs from those of [M+H]+ ions, resulting in a considerably low spectral similarity. The low MS/MS similarity between the pair of structurally related compounds can further hinder the construction of molecular networking-based annotation. Third, discrepancies in fragmentation patterns across precursor ion types can obscure the true structural details in MS/MS data, which creates challenges for machine learning models trained on such MS/MS spectra.
Nonetheless, these challenges also present opportunities. The distinct fragmentation behavior of [M + Na]⁺ can provide complementary structural information, potentially enabling the differentiation of structural isomers that remain indistinguishable in [M + H]⁺-based MS/MS analyses.
Was there a key breakthrough during your research?
I am fortunate to have worked with an outstanding student, Botao Liu, who is both highly skilled and deeply dedicated to this project. He is exceptionally patient and meticulous, with a strong attention to detail. By carefully and manually examining the results – rather than scanning through the data superficially – he uncovered several important findings that would otherwise have been overlooked.
In particular, there were two key breakthrough moments: (i) Unexpectedly low spectral similarity between adduct types: We were surprised to find that both cosine similarity and modified cosine similarity values were very low, when comparing [M + Na]⁺ and [M + H]⁺ spectra. This observation suggests the substantial fragmentation differences between these adducts and reinforces the need for adduct-specific considerations in MS/MS interpretation. (ii) Influence of collision energy: we also discovered that the effect of collision energy on fragmentation patterns is not uniform. While most precursor ion types exhibited consistent trends with [M + H]⁺ (e.g., higher collision energy resulting in higher MS/MS similarity), others – particularly dimers – showed the opposite pattern. It took considerable effort to develop a suitable hypothesis to explain this phenomenon. These insights have significantly advanced our understanding of adduct-dependent fragmentation behavior, opening new avenues for improving compound annotation workflows.
What was the biggest analytical or computational challenge you faced - and how did you overcome it?
As a first-year PhD candidate, Botao had a strong background in chemistry but limited experience with programming. As a result, it took some time to overcome the initial learning curve, particularly in developing the computational skills to manipulate large-scale MS/MS data. We also invested considerable effort in improving the quality and clarity of data visualization. As you can imagine, mapping tens of thousands of chemical entities and their associated analysis results into a single figure required a significant amount of creative thinking and careful planning.
Moreover, while relevant discussions on MS/MS fragmentation mechanisms do exist in literature, much of this work dates back several decades and is scattered across older publications. Many in the new generation of mass spectrometrists, myself included, have not been exposed to this in-depth body of knowledge during their academic training. To bridge these gaps, we reached out to multiple senior mass spectrometrists whose expertise provided valuable historical context and practical insights. We also obtained several classic reference books on mass spectrometry and studied them closely during the manuscript revision process. This combination of modern analytical approaches, historical perspectives, and foundational literature review ultimately enabled us to develop a more rigorous and well-supported explanation for our findings.
What implications do your findings have for the "dark metabolome" concept?
We appreciate the concept of the “dark metabolome.” The chemical diversity of Earth is far more complex than that encompassed by our existing base of scientific knowledge, and exploring this unknown territory is both exciting and essential for advancing our understanding of biology, ecology, and environmental chemistry. However, we must also approach the study of the “dark metabolome” with caution. It is easy to misclassify artifacts or false features as novel metabolites – examples include signals arising from in-source fragmentation, isotope peaks, background contaminants, or the MS/MS spectra of different precursor ion types being mistakenly attributed to entirely new compounds. Such misinterpretations not only inflate the apparent size of the dark metabolome but also risk diverting research efforts toward artifacts, opposed to genuine chemical discoveries.
What are the next steps for your team?
Our next step is to develop a systematic understanding of the interplay between chemical structure and precursor ion type, and how this relationship shapes the uniqueness of MS/MS spectra. Specifically, we will focus on two main research areas: First, we seek to mechanistically understand how and why changes in precursor ion type influence MS/MS fragmentation across a wide range of chemical structures. Second, we aim to leverage the distinct fragmentation patterns of different adducts – particularly [M + Na]⁺ and [M + H]⁺ – to develop a generic strategy for integrating their complementary MS/MS data. We hope this combined approach may enhance structural elucidation and improve the confidence and depth of compound identification.
Tao Huan is an Associate Professor in the Department of Chemistry at the University of British Columbia, Canada, and holds a Canada Research Chair in Metabolomics and Exposomics.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
