Mass Spectrometry: A Molecular Detective of the Invisible World
Updated: Aug 21
Dr. Sophie L. Bennett¹, Dr. Arjun V. Mehta², Prof. Tobias K. Reinhardt³
¹ Department of Analytical and Molecular Sciences, Northbridge Institute of Technology, Cambridge, UK
² Center for Biomedical Mass Spectrometry, Pacific Institute of Molecular Medicine, Singapore
³ Laboratory for Advanced Instrumental Analysis, Rheinland University of Science, Cologne, Germany
[Disclaimer: This is a sample academic article. All author names, affiliations, and institutional details are fictional and have been created solely for illustrative and educational purposes.]
Abstract
Mass spectrometry has become one of the most powerful analytical techniques for identifying and characterizing molecules by converting chemical species into gas-phase ions and measuring their mass-to-charge ratios ((m/z)). A modern mass spectrometer can be viewed as an integrated molecular measurement system in which a sample is ionized, ions are separated according to their physical behavior within electric or magnetic fields, and the resulting signals are detected and reconstructed as a mass spectrum. The development of soft-ionization methods such as electrospray ionization and matrix-assisted laser desorption/ionization transformed the field by allowing large and fragile biomolecules—including peptides, proteins, lipids, metabolites, and oligonucleotides—to be transferred into the gas phase without extensive fragmentation. Mass analyzers including quadrupoles, time-of-flight instruments, ion traps, Fourier-transform ion-cyclotron-resonance systems, and Orbitrap analyzers provide different combinations of mass resolution, accuracy, speed, sensitivity, and dynamic range, while tandem mass spectrometry allows selected ions to be fragmented so that molecular structure can be inferred from characteristic product-ion patterns. These capabilities have enabled applications ranging from proteomics, metabolomics, pharmaceutical development, clinical diagnostics, and biomarker discovery to environmental monitoring, isotope analysis, food authentication, toxicology, and forensic identification. Increasingly compact instruments are also allowing measurements to be moved from centralized analytical laboratories toward field, clinical, industrial, and potentially point-of-care environments. At the same time, machine-learning and artificial-intelligence approaches are increasingly being applied to spectral interpretation, compound annotation, proteomic identification, quantitative analysis, and the extraction of patterns from high-dimensional mass-spectrometric datasets. Particular attention is given to the physics of ion formation and separation, the operation of major mass analyzers, tandem-MS fragmentation, high-resolution molecular identification, quantitative analysis, imaging mass spectrometry, portable instrumentation, and computational interpretation. Mass spectrometry is therefore presented not simply as a molecular weighing technique but as a multidimensional analytical platform in which ionization → separation → detection → fragmentation → molecular inference provides access to chemical information that would otherwise remain invisible.
1. Introduction
Mass spectrometry is an analytical methodology in which atoms or molecules are converted into gas-phase ions and subsequently distinguished according to their mass-to-charge ratio, (m/z), allowing chemical composition to be inferred with sensitivities and molecular specificity that are difficult to obtain using most other analytical techniques. Importantly, a mass spectrometer does not directly “weigh” neutral molecules; the measured quantity reflects the behavior of charged species in electric or magnetic fields, and molecular mass must therefore be reconstructed from the observed (m/z), charge state, isotope pattern, and—in many experiments—fragmentation behavior. The development of time-of-flight analysis established that ions with different (m/z) values could be separated through differences in flight time after controlled acceleration [10.1063/1.1715212], while electrospray ionization later enabled intact multiply charged ions to be generated from proteins and other large biomolecules in solution [10.1126/science.2675315]. Matrix-assisted laser desorption/ionization similarly enabled large biomolecules to be ionized with comparatively limited fragmentation and became particularly effective when coupled to time-of-flight instrumentation [10.1021/ac00171a028]. High-resolution analyzers such as Fourier-transform ion-cyclotron-resonance and Orbitrap instruments subsequently provided greatly improved resolving power and accurate-mass measurements, enabling compounds with closely spaced nominal masses to be distinguished [10.1002/(SICI)1098-2787(1998)17:1<1::AID-MAS1>3.0.CO;2-K; 10.1021/ac991131p]. Modern mass spectrometry should therefore be understood not as a single instrument but as a modular measurement architecture in which sample introduction → ionization → mass analysis → detection → fragmentation → computational identification can be configured differently according to the chemical question being addressed.
2. Results and Discussion
The extraordinary versatility of mass spectrometry is produced by the ability to combine different ionization mechanisms, separation methods, mass analyzers, fragmentation strategies, and computational workflows within a single analytical platform. A volatile environmental contaminant can be separated by gas chromatography and ionized by electron impact, a peptide mixture can be separated by liquid chromatography and transferred through electrospray ionization, a tissue section can be sampled directly by imaging mass spectrometry, and an unknown surface contaminant can be analyzed under ambient conditions without conventional sample preparation [10.1126/science.1104404; 10.1021/ac970888i]. The central analytical challenge is therefore not simply obtaining a spectrum but determining how confidently spectral information can be converted into molecular identity, abundance, structure, and spatial or biological meaning.
2.1. The Fundamental Measurement: Mass-to-Charge Ratio Rather Than Mass
The fundamental coordinate of a conventional mass spectrum is (m/z), where (m) represents ion mass and (z) represents the number of elementary charges carried by that ion. For singly charged species, the measured (m/z) closely corresponds numerically to molecular or ionic mass, but multiply charged ions generated by electrospray can place very large biomolecules within comparatively modest (m/z) ranges [10.1126/science.2675315]. A protein carrying 20 positive charges, for example, produces an ion whose (m/z) is roughly one twentieth of its neutral molecular mass after accounting for the added protons, and a distribution of charge states can therefore be used to reconstruct the intact molecular mass [10.1002/mas.1280090103]. Isotope distributions provide additional information because naturally occurring isotopes generate predictable peak spacings and abundance patterns, while high-resolution instrumentation can resolve small differences between elemental compositions with the same nominal integer mass [10.1021/ac991131p]. The interpretation of mass spectra is consequently an inverse problem in which ion mass, charge, isotope composition, adduct formation, and fragmentation must be disentangled before molecular identity can be assigned.
2.2. Ionization: Making Neutral Molecules Visible to the Instrument
A neutral molecule cannot be manipulated efficiently by the electric and magnetic fields used in most mass analyzers, so ionization represents the critical interface between the chemical sample and the mass spectrometer. Electrospray ionization generates highly charged droplets from solution under a strong electric field, followed by solvent evaporation and progressive evolution toward gas-phase ions; because large biomolecules can acquire multiple charges, proteins and oligonucleotides can be measured without requiring them to be converted into volatile neutral species [10.1126/science.2675315]. Matrix-assisted laser desorption/ionization uses a laser-absorbing matrix to assist the desorption and ionization of embedded analytes and was shown early to produce protein ions with molecular masses exceeding 10,000 Da [10.1021/ac00171a028]. These “soft” ionization approaches changed biological mass spectrometry fundamentally because molecules that would have decomposed under more energetic ionization conditions could be transferred into the gas phase while retaining substantial structural information [10.1126/science.2675315; 10.1021/ac00171a028]. Ionization efficiency nevertheless varies strongly among molecules, meaning that observed signal intensity is not automatically proportional to concentration unless matrix effects, charge competition, extraction, and calibration are properly controlled.
2.3. Time-of-Flight Mass Spectrometry: Separating Ions by Flight Time
Time-of-flight mass spectrometry separates ions according to the relationship between their velocity and (m/z) after acceleration through an electric potential. When ions receive approximately the same kinetic energy, lighter ions or ions with higher charge travel more rapidly than heavier ions with the same charge, giving the approximate relation (t\propto\sqrt{m/z}) between flight time and mass-to-charge ratio [10.1063/1.1715212]. Wiley and McLaren showed that spatial and energy focusing could substantially improve TOF resolution, establishing principles that remain central to modern instruments [10.1063/1.1715212]. TOF analyzers possess no intrinsic upper mass limit imposed by a scanning stability window in the same sense as many trapping analyzers, making them particularly compatible with MALDI and the analysis of high-mass biomolecules [10.1021/ac00171a028]. Modern TOF architectures use reflectrons, orthogonal acceleration, rapid digitization, and high-repetition-rate sources to improve resolving power and mass accuracy, but the underlying physical measurement remains strikingly simple: molecular information is transformed into a precisely measured difference in arrival time.
2.4. Quadrupoles and Ion Traps: Filtering and Storing Ions with Oscillating Fields
Quadrupole mass filters and quadrupole ion traps exploit radio-frequency electric fields to create ion trajectories whose stability depends on (m/z), allowing selected ion populations either to pass through an analyzer or to remain confined within a trapping volume. Quadrupole ion traps can sequentially isolate, excite, fragment, and eject ions, which has made them particularly valuable for tandem and multistage mass spectrometry [10.1002/mas.20250]. Because ion populations can be stored temporarily rather than measured immediately after formation, controlled experiments can be performed on selected precursor ions before their products are detected [10.1002/(SICI)1097-0231(19981030)12:20<1543::AID-RCM343>3.0.CO;2-T]. Triple-quadrupole instruments similarly combine ion selection, collision-induced fragmentation, and product-ion filtering and are widely used when high selectivity and quantitative targeted analysis are required. The central advantage of these architectures is therefore not maximal resolving power but controllable ion manipulation, enabling a molecular ion to be selected from a complex mixture and interrogated through its fragmentation behavior.
2.5. Fourier-Transform and Orbitrap Analysis: Reading Ion Motion as Frequency
Fourier-transform mass spectrometers determine (m/z) indirectly by measuring characteristic frequencies of trapped ion motion rather than sequentially ejecting each mass from the analyzer. In Fourier-transform ion-cyclotron-resonance mass spectrometry, ions orbit within a strong magnetic field at cyclotron frequencies related to their (m/z), and the image currents generated by coherent ion motion can be transformed mathematically into a mass spectrum with extremely high resolving power [10.1002/(SICI)1098-2787(1998)17:1<1::AID-MAS1>3.0.CO;2-K]. The Orbitrap uses an electrostatic field rather than a magnetic field, with trapped ions oscillating axially at frequencies dependent primarily on (m/z); Makarov's original analyzer demonstrated resolving powers approaching 150,000 under experimental conditions [10.1021/ac991131p]. Orbitrap instruments subsequently became widely used for high-resolution accurate-mass analysis in proteomics, metabolomics, toxicology, and environmental screening because complex spectra can be acquired with high mass accuracy and substantial dynamic range [10.1002/mas.20186]. In these instruments, molecular mass is effectively converted into an oscillation frequency, and a computational Fourier transform converts the measured time-domain ion signal into the familiar frequency- and mass-domain spectrum.
2.6. Tandem Mass Spectrometry: Breaking Molecules to Learn Their Structure
A single accurate molecular mass can substantially constrain possible elemental compositions but usually cannot determine a unique molecular structure, because isomers can possess identical elemental formulas and exact masses. Tandem mass spectrometry addresses this limitation by selecting a precursor ion, inducing fragmentation, and measuring the resulting product ions, creating a structural fingerprint that can be compared with theoretical expectations, reference spectra, or sequence databases [10.1126/science.1124619]. In proteomics, peptide fragmentation patterns allow amino-acid sequence information to be extracted and matched to protein databases, transforming complex mixtures containing thousands of peptides into identifiable protein sets [10.1038/nature01511]. Repeating stages of selection and fragmentation in trapping instruments can provide additional structural information, while alternative fragmentation mechanisms reveal complementary aspects of molecular architecture. MS/MS therefore converts mass spectrometry from a method concerned primarily with molecular weight into a technique capable of probing chemical connectivity.
2.7. Proteomics: Identifying Thousands of Proteins through Peptide Mass Spectra
The combination of enzymatic digestion, liquid chromatography, electrospray ionization, tandem mass spectrometry, and computational database searching established mass spectrometry as one of the central analytical technologies of modern proteomics [10.1038/nature01511]. Proteins are commonly digested into peptides, which are separated chromatographically, ionized, measured, and fragmented; the observed fragment spectra are then matched against peptide sequences predicted from genomic or protein databases [10.1126/science.1124619]. Quantitative workflows can additionally use isotope labeling, internal standards, or label-free signal comparison to estimate relative or absolute protein abundance across biological conditions [10.1038/nature01511]. The power of this strategy arises from multiplexing: instead of requiring a separate molecular probe for every protein, a common physicochemical measurement can identify many thousands of peptide species in parallel. The difficulty is correspondingly computational, because false assignments, incomplete sequence coverage, post-translational modifications, dynamic range, missing values, and multiple-testing effects must all be controlled before biological conclusions can be considered reliable [10.1074/mcp.R600012-MCP200].
2.8. Imaging Mass Spectrometry: Turning Molecular Spectra into Spatial Maps
Mass spectrometry can also be used as a form of molecular microscopy when spectra are collected repeatedly across defined spatial coordinates of a tissue or material surface. Caprioli, Farmer, and Gile demonstrated MALDI-TOF imaging in which peptides and proteins were mapped directly across biological tissue sections, creating ion-intensity images for selected (m/z) values while retaining the broader molecular spectrum at each location [10.1021/ac970888i]. Unlike fluorescence microscopy, this approach does not necessarily require an antibody or fluorescent label to be prepared for every molecular species, allowing many chemically distinct molecules to be mapped within the same experiment [10.1002/pmic.201600133]. Modern imaging mass spectrometry has been extended to lipids, metabolites, pharmaceuticals, peptides, proteins, and other compound classes, with spatial resolution, molecular coverage, sensitivity, and identification confidence depending strongly on sample preparation and instrumentation [10.1002/pmic.201600133]. The resulting dataset is intrinsically multidimensional because each pixel contains an entire mass spectrum rather than a single intensity value.
2.9. Environmental Analysis: Finding Pollutants That Were Not Explicitly Searched For
Environmental samples such as river water, wastewater, soil extracts, atmospheric particles, and biological tissues can contain thousands of anthropogenic and naturally occurring chemicals, many of which are unknown before analysis begins. High-resolution mass spectrometry has enabled a transition from strictly targeted analysis toward suspect and non-target screening, in which accurate masses, isotope patterns, retention behavior, and MS/MS fragments are used to search for chemical features without requiring every compound to have been specified before the experiment [10.1021/acs.est.7b02184]. Such methods can reveal transformation products, emerging contaminants, industrial chemicals, pharmaceuticals, pesticides, and other compounds that would remain invisible in a narrowly targeted assay [10.1007/s00216-015-8681-7]. However, detection of an accurate-mass feature is not equivalent to definitive molecular identification, and confidence must increase progressively through formula assignment, isotope evidence, fragmentation, database comparison, reference standards, and orthogonal information [10.1007/s00216-015-8681-7]. Environmental HRMS therefore expands the chemical search space dramatically while simultaneously creating a major challenge in distinguishing meaningful pollutants from thousands of unresolved spectral candidates.
2.10. Forensic Science and Toxicology: Chemical Evidence from Complex Samples
Mass spectrometry is extensively used in forensic and clinical toxicology because drugs, metabolites, poisons, and other xenobiotics must often be identified in complex matrices such as blood, urine, hair, tissue, or unknown powders at low concentrations. High-resolution LC-MS/MS can combine targeted quantitative analysis with broader screening for unexpected compounds, metabolites, and emerging psychoactive substances, providing a more flexible analytical strategy than assays restricted to predefined analyte lists [10.1007/s00216-020-03064-y]. Ambient mass-spectrometric approaches have also demonstrated direct chemical imaging of latent fingerprints, showing that endogenous compounds and externally deposited chemicals can be spatially analyzed without conventional extraction of the entire fingerprint [10.1126/science.1157199]. Paper-spray ionization has enabled rapid direct analysis of complex biological samples and has been demonstrated for therapeutic drugs and illicit compounds in dried or untreated biofluids [10.1021/ac902854g]. Nevertheless, forensic interpretation requires especially rigorous validation because chemical identification must be separated from assumptions about dose, timing, impairment, exposure route, or legal significance.
2.11. Ambient Mass Spectrometry: Bringing the Instrument to the Sample
Traditional mass-spectrometric workflows frequently require samples to be extracted, chromatographically separated, and introduced into a high-vacuum instrument, but ambient ionization methods were developed to reduce this separation between the sample and the detector. Desorption electrospray ionization demonstrated that charged solvent droplets could be directed onto surfaces in the open atmosphere, desorbing analytes that were then transferred directly into a mass spectrometer [10.1126/science.1104404]. This approach enabled rapid analysis of materials in native or minimally prepared states and helped establish the broader field of ambient mass spectrometry [10.1126/science.1119426]. Paper spray subsequently demonstrated that a triangular piece of paper containing a sample could function simultaneously as a substrate, extraction medium, and electrospray emitter after solvent and high voltage were applied [10.1021/ac902854g]. These developments changed the analytical question from “How can the sample be brought into the mass spectrometry laboratory?” toward “How can ionization and mass analysis be moved closer to the sample?”
2.12. Miniaturization: From Laboratory Instrument to Field Analyzer
Conventional high-performance mass spectrometers are large because efficient ion transmission and mass analysis generally require vacuum systems, precision electrodes, high-voltage electronics, pumps, detectors, and stable mechanical structures, but substantial effort has been directed toward miniaturizing these components [10.1146/annurev-anchem-060908-155229]. Rectilinear ion traps, discontinuous atmospheric-pressure interfaces, compact vacuum systems, and miniature electronics have enabled portable instruments with useful tandem-MS capability to be constructed [10.1021/ac800014v; 10.1021/ac801275x]. A handheld rectilinear-ion-trap instrument was demonstrated for monitoring toxic compounds in air, illustrating the potential of portable MS for environmental monitoring, public safety, and industrial hygiene [10.1021/ac070906o]. Wearable backpack systems have subsequently enabled direct in-situ sampling and analysis without transferring the specimen to a centralized laboratory [10.1021/ac403765x]. Portable mass spectrometry nevertheless involves compromises among resolving power, mass range, sensitivity, vacuum quality, battery consumption, ruggedness, and cost, so miniaturization should be understood as optimization for a particular analytical task rather than simply shrinking a laboratory instrument.
2.13. Artificial Intelligence and Computational Molecular Identification
As mass spectrometers have become faster and higher resolution, the major bottleneck has increasingly shifted from generating spectra toward interpreting the enormous number of features and fragmentation patterns contained within those spectra. Computational methods such as SIRIUS combine accurate precursor masses, isotope distributions, fragmentation trees, and molecular-structure databases to infer formulas and candidate metabolite structures from tandem mass spectra [10.1038/s41592-019-0344-8]. Deep-learning methods have also been used to predict peptide fragmentation intensities and chromatographic retention behavior, allowing synthetic spectral libraries to be generated computationally and improving peptide identification in proteomic datasets [10.1038/s41592-019-0426-7]. CANOPUS demonstrated that deep neural networks can classify unknown metabolites into thousands of chemical classes directly from high-resolution fragmentation spectra, including cases for which a specific reference spectrum is unavailable [10.1038/s41587-020-0740-8]. More recent graph-transformer models have been developed to predict tandem mass spectra of small molecules from molecular structure, illustrating how machine learning is increasingly connecting the forward problem—predicting a spectrum from a molecule—with the inverse problem of inferring a molecule from its spectrum [10.1038/s42256-024-00816-8]. Artificial intelligence therefore does not replace the mass spectrometer; it expands the fraction of the instrument's chemical information that can be interpreted.
2.14. Why Accurate Mass Alone Does Not Prove Molecular Identity
Modern high-resolution instruments can measure (m/z) with extremely small errors, but an accurate mass is still not equivalent to an unambiguous molecular structure. Different structural isomers can possess exactly the same elemental composition and therefore essentially identical monoisotopic masses, while adducts, in-source fragments, isotopologues, charge states, and background contaminants can generate additional features that complicate interpretation. Environmental non-target screening has demonstrated the necessity of graded identification confidence because many candidate structures can remain plausible even after accurate-mass and fragmentation information have been collected [10.1007/s00216-015-8681-7]. Computational structure-search tools can rank candidates but cannot automatically transform incomplete spectral evidence into definitive identification [10.1038/s41592-019-0344-8]. The strongest molecular assignments therefore combine multiple independent forms of evidence—including precursor mass, isotope pattern, chromatographic behavior, diagnostic fragments, authentic reference standards, and where appropriate orthogonal spectroscopy—rather than treating a single high-resolution peak as proof of identity.
2.15. Quantification: A Large Peak Is Not Automatically More Molecules
Mass-spectral peak intensity depends not only on the quantity of analyte present but also on extraction efficiency, chromatographic recovery, ionization probability, charge competition, matrix suppression or enhancement, fragmentation efficiency, detector response, and instrumental settings. Electrospray is particularly susceptible to matrix effects because coeluting molecules compete during droplet formation, desolvation, and ion production, so two compounds present at equal concentrations may generate very different signals [10.1002/mas.1280090103]. Reliable quantitative mass spectrometry therefore commonly uses calibration curves, isotopically labeled internal standards, standard addition, matrix-matched controls, or carefully validated normalization strategies rather than assuming a universal relationship between signal intensity and concentration. The analytical distinction is critical: mass spectrometry can be extraordinarily sensitive, but sensitivity and quantitative accuracy are separate properties. A molecular ion can be detected confidently while its abundance remains uncertain unless the measurement process has been calibrated specifically for that analyte and matrix.
3. Conclusion and Outlook
Mass spectrometry has evolved from a technique for separating relatively simple ions into a multidimensional molecular-analysis platform capable of identifying proteins, metabolites, pharmaceuticals, pollutants, toxins, lipids, and other chemical species across samples ranging from purified solutions to intact tissues and environmental surfaces. Electrospray and MALDI expanded the accessible molecular-mass range and enabled fragile biomolecules to be ionized [10.1126/science.2675315; 10.1021/ac00171a028], TOF, FT-ICR, quadrupole, ion-trap, and Orbitrap technologies provided complementary approaches to ion separation and measurement [10.1063/1.1715212; 10.1002/(SICI)1098-2787(1998)17:1<1::AID-MAS1>3.0.CO;2-K; 10.1021/ac991131p], and tandem mass spectrometry transformed ion fragmentation into a practical source of structural information [10.1126/science.1124619]. These advances made large-scale proteomics possible [10.1038/nature01511], enabled molecular imaging directly within tissues [10.1021/ac970888i], and expanded environmental and forensic analysis from predetermined target lists toward broad high-resolution screening [10.1021/acs.est.7b02184; 10.1007/s00216-020-03064-y]. Ambient ionization and miniature instrumentation are progressively moving analysis beyond centralized laboratories [10.1126/science.1104404; 10.1146/annurev-anchem-060908-155229], while machine learning is increasingly being used to predict spectra, classify unknown compounds, and interpret datasets whose complexity exceeds practical manual analysis [10.1038/s41592-019-0426-7; 10.1038/s41587-020-0740-8]. The future of the field is therefore likely to involve a convergence of high-resolution instrumentation → direct sampling → multidimensional separation → autonomous spectral interpretation → portable and real-time molecular decision-making, but the fundamental requirement will remain unchanged: a spectrum is evidence about molecular composition, and the strength of any chemical conclusion must remain proportional to the quality and independence of the evidence supporting it.
References
Wiley, W. C., & McLaren, I. H. (1955). Time-of-flight mass spectrometer with improved resolution. Review of Scientific Instruments, 26, 1150–1157. [10.1063/1.1715212]
Karas, M., & Hillenkamp, F. (1988). Laser desorption ionization of proteins with molecular masses exceeding 10,000 daltons. Analytical Chemistry, 60, 2299–2301. [10.1021/ac00171a028]
Fenn, J. B., Mann, M., Meng, C. K., Wong, S. F., & Whitehouse, C. M. (1989). Electrospray ionization for mass spectrometry of large biomolecules. Science, 246, 64–71. [10.1126/science.2675315]
Fenn, J. B., Mann, M., Meng, C. K., Wong, S. F., & Whitehouse, C. M. (1990). Electrospray ionization—principles and practice. Mass Spectrometry Reviews, 9, 37–70. [10.1002/mas.1280090103]
Marshall, A. G., Hendrickson, C. L., & Jackson, G. S. (1998). Fourier transform ion cyclotron resonance mass spectrometry: A primer. Mass Spectrometry Reviews, 17, 1–35. [10.1002/(SICI)1098-2787(1998)17:1<1::AID-MAS1>3.0.CO;2-K]
Makarov, A. (2000). Electrostatic axially harmonic orbital trapping: A high-performance technique of mass analysis. Analytical Chemistry, 72, 1156–1162. [10.1021/ac991131p]
Aebersold, R., & Mann, M. (2003). Mass spectrometry-based proteomics. Nature, 422, 198–207. [10.1038/nature01511]
Takáts, Z., Wiseman, J. M., Gologan, B., & Cooks, R. G. (2004). Mass spectrometry sampling under ambient conditions with desorption electrospray ionization. Science, 306, 471–473. [10.1126/science.1104404]
Domon, B., & Aebersold, R. (2006). Mass spectrometry and protein analysis. Science, 312, 212–217. [10.1126/science.1124619]
Cooks, R. G., Ouyang, Z., Takáts, Z., & Wiseman, J. M. (2006). Ambient mass spectrometry. Science, 311, 1566–1570. [10.1126/science.1119426]
Caprioli, R. M., Farmer, T. B., & Gile, J. (1997). Molecular imaging of biological samples: Localization of peptides and proteins using MALDI-TOF MS. Analytical Chemistry, 69, 4751–4760. [10.1021/ac970888i]
Perry, R. H., Cooks, R. G., & Noll, R. J. (2008). Orbitrap mass spectrometry: Instrumentation, ion motion and applications. Mass Spectrometry Reviews, 27, 661–699. [10.1002/mas.20186]
Ifa, D. R., Manicke, N. E., Dill, A. L., & Cooks, R. G. (2008). Latent fingerprint chemical imaging by mass spectrometry. Science, 321, 805. [10.1126/science.1157199]
Ouyang, Z., & Cooks, R. G. (2009). Miniature mass spectrometers. Annual Review of Analytical Chemistry, 2, 187–214. [10.1146/annurev-anchem-060908-155229]
March, R. E. (2009). Quadrupole ion traps. Mass Spectrometry Reviews, 28, 961–989. [10.1002/mas.20250]
Liu, J., Wang, H., Manicke, N. E., Lin, J.-M., Cooks, R. G., & Ouyang, Z. (2010). Development, characterization, and application of paper spray ionization. Analytical Chemistry, 82, 2463–2471. [10.1021/ac902854g]
Schymanski, E. L., et al. (2015). Non-target screening with high-resolution mass spectrometry: Critical review using a collaborative trial on water analysis. Analytical and Bioanalytical Chemistry, 407, 6237–6255. [10.1007/s00216-015-8681-7]
Hollender, J., Schymanski, E. L., Singer, H. P., & Ferguson, P. L. (2017). Nontarget screening with high resolution mass spectrometry in the environment: Ready to go? Environmental Science & Technology, 51, 11505–11512. [10.1021/acs.est.7b02184]
Dührkop, K., et al. (2019). SIRIUS 4: A rapid tool for turning tandem mass spectra into metabolite structure information. Nature Methods, 16, 299–302. [10.1038/s41592-019-0344-8]
Gessulat, S., et al. (2019). Prosit: Proteome-wide prediction of peptide tandem mass spectra by deep learning. Nature Methods, 16, 509–518. [10.1038/s41592-019-0426-7]
Maurer, H. H. (2021). Hyphenated high-resolution mass spectrometry—the “all-in-one” device in analytical toxicology? Analytical and Bioanalytical Chemistry, 413, 2303–2309. [10.1007/s00216-020-03064-y]
Dührkop, K., et al. (2021). Systematic classification of unknown metabolites using high-resolution fragmentation mass spectra. Nature Biotechnology, 39, 462–471. [10.1038/s41587-020-0740-8]
Young, A., Röst, H., & Wang, B. (2024). Tandem mass spectrum prediction for small molecules using graph transformers. Nature Machine Intelligence, 6, 404–416. [10.1038/s42256-024-00816-8]
Comments