From fragments to function: a review of AI-driven mass spectrometry and fragment-based strategies

Abstract

Traditional Chinese medicine (TCM) possesses unique advantages in disease prevention and treatment, yet its inherent complexity and diversity pose tremendous challenges for structural elucidation, mechanism research, and bioactivity characterization. High-resolution mass spectrometry (HRMS) technology demonstrates immense potential in TCM analysis due to its high sensitivity, high resolution, high throughput, and high efficiency. However, its application in TCM research remains constrained by the lack of intelligent analytical methods and unified standardized databases. Therefore, this paper focuses on the integration of artificial intelligence (AI) and mass spectrometry, providing a systematic review of the applications and potential of AI-driven mass spectrometry analysis in structural elucidation, data resource integration, multi-omics mechanism studies, and chemical biology. Furthermore, this article emphasizes that by leveraging AI models to learn the complex mapping from chemical structures to biological functions, fragment-based characterization has emerged as the bridge connecting chemical structures with biological activities. Molecular fragments themselves serve as the core “knowledge units” that carry bioactive information. Future research will focus on establishing high-quality mass spectrometry databases for the complete chemical profiles of TCM and promoting the standardization and open sharing of mass spectrometry databases, thereby advancing the integration of AI and mass spectrometry in TCM analysis and providing new tools for TCM research.

Keywords

Artificial Intelligence / Fragment-based representation / Mass spectrometry / Structural elucidation / Traditional Chinese medicine

References

[1] Zhang B, Yang S, Guo D—a. The quest for the modernization and internationalization of traditional Chinese medicine. Engineering 2019; 5(1): 1-2.

[2] Newman DJ, Cragg GM. Natural products as sources of new drugs over the nearly four decades from 01/1981 to 09/2019. J Nat Prod 2020; 83(3): 770-803.

[3] Zhu B, Li Z, Jin Z, et al. Knowledge—based in silico fragmentation and annotation of mass spectra for natural products with MassKG. Comput Struct Biotechnol J 2024; 23: 3327-3341.

[4] Yang X—X, Gu W, Liang L, et al. Screening for the bioactive constituents of traditional Chinese medicines—progress and challenges. RSC Adv 2017; 7(6): 3089-3100.

[5] Wang X, Zhang A, Yan G, et al. UHPLC—MS for the analytical characterization of traditional Chinese medicines. TRAC Trends Anal Chem 2014; 63: 180-187.

[6] Pang B, Zhu Y, Lu L, et al. The applications and features of liquid chromatography—mass spectrometry in the analysis of traditional Chinese medicine. Evid Based Complement Alternat Med 2016; 2016: 3837270.

[7] Wang R, Zhang L, Li X, et al. Single—cell sequencing and mass spectrometry imaging reveal the multicellular compartmentalization map of plant triterpenes: a ginsenoside in Panax ginseng example. Plant Biotechnol J 2025; 23(10): 4461-4476.

[8] Duan Y—y, P—r L, T—t H, et al. Application and development of intelligent medicine in traditional Chinese medicine. Current Med Sci 2021; 41(6): 1116-1122.

[9] Zhang J, Chen X, Huang L, et al. Traditional Chinese medicine + artificial intelligence: Wuzhen consensus. Acupunct Herb Med 2025; 5(2): 134-135.

[10] Chen Z, Wang H, Li C, et al. Large language models in traditional Chinese medicine: a systematic review. Acupunct Herb Med 2025; 5(1): 57-67.

[11] Xu Y, Ma Y, Xu W, et al. A large language model for deriving spectral embeddings for accurate compound identification in mass spectrometry. Commun Chem 2025; 8(1): 326.

[12] Bilbao A. The future of a myriad of accelerated biodiscoveries lies in AI—powered mass spectrometry and multiomics integration. J Mass Spectrom 2025; 60(8): e5157.

[13] Xing S, Huan T. Radical fragment ions in collision—induced dissociation—based tandem mass spectrometry. Anal Chim Acta 2022; 1200: 339613.

[14] Alexandrov T. Spatial metabolomics and imaging mass spectrometry in the age of artificial intelligence. Ann Rev Biomed Data Sci 2020; 3(1): 61-87.

[15] Guo G, Papanicolaou M, Demarais NJ, et al. Automated annotation and visualisation of high—resolution spatial proteomic mass spectrometry imaging data using HIT—MAP. Nat Commun 2021; 12(1): 3241.

[16] Hu G, Qiu M. Machine learning—assisted structure annotation of natural products based on MS and NMR data. Nat Prod Rep 2023; 40(11): 1735-1753.

[17] Kim S, Kato I, Zhang X. Comparative analysis of binary similarity measures for compound identification in mass spectrometry—based metabolomics. Metabolites 2022; 12(8): 694.

[18] Huber F, Ridder L, Verhoeven S, et al. Spec2Vec: improved mass spectral similarity scoring through learning of structural relationships. PLoS Comput Biol 2021; 17(2): e1008724.

[19] Watrous J, Roach P, Alexandrov T, et al. Mass spectral molecular networking of living microbial colonies. Proc Natl Acad Sci USA 2012; 109(26): E1743-E1752.

[20] Wang M, Jarmusch AK, Vargas F, et al. Mass spectrometry searches using MASST. Nat Biotechnol 2020; 38(1): 23-26.

[21] Huber F, van der Burg S, van der Hooft J J, et al. MS2DeepScore: a novel deep learning similarity measure to compare tandem mass spectra. J Cheminf 2021; 13(1): 84.

[22] de Jonge N F, R LJJ, Chekmeneva E, et al. MS2Query: reliable and scalable MS(2) mass spectra—based analogue search. Nat Commun 2023; 14(1): 1752.

[23] Shen X, Wang R, Xiong X, et al. Metabolic reaction network—based recursive metabolite annotation for untargeted metabolomics. Nat Commun 2019; 10(1): 1516.

[24] Ji H, Xu Y, Lu H, et al. Deep MS/MS—aided structural—similarity scoring for unknown metabolite identification. Anal Chem 2019; 91(9): 5629-5637.

[25] Wang X, Li C, Li Z, et al. A structure—guided molecular network strategy for global untargeted metabolomics data annotation. Anal Chem 2023; 95(31): 11603-11612.

[26] Kim S, Thiessen PA, Bolton EE, et al. PubChem substance and compound databases. Nucleic Acids Res 2016; 44(D1): D1202-D1213.

[27] da Silva RR, Dorrestein PC, Quinn RA. Illuminating the dark matter in metabolomics. Proc Natl Acad Sci U S A 2015; 112(41): 12549-12550.

[28] Kanehisa M, Goto S. KEGG: Kyoto encyclopedia of genes and genomes. Nucleic Acids Res 2000; 28(1): 27-30.

[29] Allard PM, Péresse T, Bisson J, et al. Integration of molecular networking and in—silico MS/MS fragmentation for natural products dereplication. Anal Chem 2016; 88(6): 3317-3323.

[30] Lai Z, Tsugawa H, Wohlgemuth G, et al. Identifying metabolites by integrating metabolome databases with mass spectrometry cheminformatics. Nat Methods 2018; 15(1): 53-56.

[31] Chen B, Li H, Huang R, et al. Deep learning prediction of electrospray ionization tandem mass spectra of chemically derived molecules. Nat Commun 2024; 15(1): 8396.

[32] Goldman S, Li J, Coley CW. Generating molecular fragmentation graphs with autoregressive neural networks. Anal Chem 2024; 96(8): 3419-3428.

[33] Wolf S, Schmidt S, Müller—Hannemann M, et al. In silico fragmentation for computer assisted identification of metabolite mass spectra. BMC Bioinf 2010; 11: 148.

[34] Allen F, Greiner R, Wishart D. Competitive fragmentation modeling of ESI—MS/MS spectra for putative metabolite identification. Metabolomics 2015; 11(1): 98-110.

[35] Nowatzky Y, Russo FF, Lisec J, et al. FIORA: Local neighborhood—based prediction of compound mass spectra from single fragmentation events. Nat Commun 2025; 16(1): 2298.

[36] Litsa EE, Chenthamarakshan V, Das P, et al. An end—to—end deep learning framework for translating mass spectra to de—novo molecules. Commun Chem 2023; 6(1): 132.

[37] Young A, Röst H, Wang B. Tandem mass spectrum prediction for small molecules using graph transformers. Nat Mach Intell 2024; 6(4): 404-416.

[38] Wei JN, Belanger D, Adams RP, et al. Rapid prediction of electron—ionization mass spectrometry using neural networks. ACS Cent Sci 2019; 5(4): 700-708.

[39] Hong Y, Li S, Welch CJ, et al. 3DMolMS: prediction of tandem mass spectra from 3D molecular conformations. Bioinformatics 2023; 39(6): btad354.

[40] Dührkop K, Shen H, Meusel M, et al. Searching molecular structure databases with tandem mass spectra using CSI:FingerID. Proc Natl Acad Sci U S A 2015; 112(41): 12580-12585.

[41] Hoffmann MA, Nothias LF, Ludwig M, et al. High—confidence structural annotation of metabolites absent from spectral libraries. Nat Biotechnol 2022; 40(3): 411-421.

[42] Stravs MA, Dührkop K, Böcker S, et al. MSNovelist: de novo structure generation from mass spectra. Nat Methods 2022; 19(7): 865-870.

[43] Bergman HM, Lundin E, Andersson M, et al. Quantitative mass spectrometry imaging of small—molecule neurotransmitters in rat brain tissue sections using nanospray desorption electrospray ionization. Analyst 2016; 141(12): 3686-3695.

[44] Zhang Y, Buchberger A, Muthuvel G, et al. Expression and distribution of neuropeptides in the nervous system of the crab Carcinus maenas and their roles in environmental stress. Proteomics 2015; 15(23—24): 3969-3979.

[45] Zhang H, Lu KH, Ebbini M, et al. Mass spectrometry imaging for spatially resolved multi—omics molecular mapping. NPJ Imaging 2024; 2(1): 20.

[46] Palmer A, Phapale P, Chernyavsky I, et al. FDR—controlled metabolite annotation for high—resolution imaging mass spectrometry. Nat Methods 2017; 14(1): 57-60.

[47] Alexandrov T, Ovchinnikova K, Palmer A, et al. METASPACE: a community—populated knowledge base of spatial metabolomes in health and disease. BioRxiv 2019; 539478.

[48] Wadie B, Stuart L, Rath CM, et al. METASPACE—ML: context—specific metabolite annotation for imaging mass spectrometry using machine learning. Nat Commun 2024; 15(1): 9110.

[49] Tortorella S, Tiberi P, Bowman AP, et al. LipostarMSI: comprehensive, vendor—neutral software for visualization, data analysis, and automated molecular identification in mass spectrometry imaging. J Am Soc Mass Spectrom 2020; 31(1): 155-163.

[50] Yang F, Liang Z, Zhao H, et al. Mass spectral database—based methodologies for the annotation and discovery of natural products. Chin J Nat Med 2025; 23(4): 410-420.

[51] Wang M, Carver JJ, Phelan VV, et al. Sharing and community curation of mass spectrometry data with global natural products social molecular networking. Nat Biotechnol 2016; 34(8): 828-837.

[52] Horai H, Arita M, Kanaya S, et al. MassBank: a public repository for sharing mass spectral data for life sciences. J Mass Spectrom 2010; 45(7): 703-714.

[53] Xue J, Guijas C, Benton HP, et al. METLIN MS(2) molecular standards database: a broad chemical and biological resource. Nat Methods 2020; 17(10): 953-954.

[54] Wishart DS, Guo A, Oler E, et al. HMDB 5.0: the human metabolome database for 2022. Nucleic Acids Res 2022; 50(D1): D622-D631.

[55] Lee S, Hwang S, Seo M, et al. BMDMS—NP: a comprehensive ESI—MS/MS spectral library of natural compounds. Phytochem 2020; 177: 112427.

[56] Pedrioli PG, Eng JK, Hubley R, et al. A common open representation of mass spectrometry data and its application to proteomics research. Nat Biotechnol 2004; 22(11): 1459-1466.

[57] Martens L, Chambers M, Sturm M, et al. mzML—a community standard for mass spectrometry data. Mol Cell Proteomics 2011; 10(1): R110.000133.

[58] Bhamber RS, Jankevics A, Deutsch EW, et al. mzMLb: a future—proof raw mass spectrometry data format based on standards—compliant mzML and optimized for speed and storage requirements. J Proteome Res 2021; 20(1): 172-183.

[59] Damiani T, Jarmusch AK, Aron AT, et al. A universal language for finding mass spectrometry data patterns. Nat Methods 2025; 22(6): 1247-1254.

[60] Jarmusch SA, van der Hooft JJ, Dorrestein PC, et al. Advancements in capturing and mining mass spectrometry data are transforming natural products research. Nat Prod Rep 2021; 38(11): 2066-2082.

[61] Kang KB, Jeong E, Son S, et al. Mass spectrometry data on specialized metabolome of medicinal plants used in East Asian traditional medicine. Sci Data 2022; 9(1): 528.

[62] Paulhe N, Canlet C, Damont A, et al. PeakForest: a multi—platform digital infrastructure for interoperable metabolite spectral data and metadata management. Metabolomics 2022; 18(6): 40.

[63] Zhai Y, Liu L, Zhang F, et al. Network pharmacology: a crucial approach in traditional Chinese medicine research. Chin Med 2025; 20(1): 8.

[64] Wu L, Wu Q, Du X, et al. Revolutionizing pharmacological research of traditional Chinese medicine with single—cell omics technologies. Fitoterapia 2025; 186: 106846.

[65] Hui T, Zhou J, Yao M, et al. Advances in spatial omics technologies. Small Methods 2025; 9(5): e2401171.

[66] Zhang Y, Li XJ, Wang XR, et al. Integrating metabolomics and network pharmacology to explore the mechanism of Xiao—Yao—San in the treatment of inflammatory response in CUMS mice. Pharmaceuticals (Basel) 2023; 16(11): 1607.

[67] Shu L, Wang Y, Huang W, et al. Integrating metabolomics and network pharmacology to explore the mechanism of Tongmai Yangxin pills in ameliorating doxorubicin—induced cardiotoxicity. ACS Omega 2023; 8(20): 18128-18139.

[68] Pan M, Deng Y, Qiu Y, et al. Shenling Baizhu powder alleviates non—alcoholic fatty liver disease by modulating autophagy and energy metabolism in high—fat diet—induced rats. Phytomedicine 2024; 130: 155712.

[69] Zhang X, Zhang J, Zhou Z, et al. Integrated network pharmacology, metabolomics, and transcriptomics of Huanglian—Hongqu herb pair in non—alcoholic fatty liver disease. J Ethnopharmacol 2024; 325: 117828.

[70] Fan Y, Wang A, Liu Z, et al. Integrated spatial metabolomics and network pharmacology to explore the pharmacodynamic substances and mechanism of Radix ginseng—Schisandra chinensis Herb Couple on Alzheimer’s disease. Anal Bioanal Chem 2024; 416(19): 4275-4288.

[71] Ali A, Abouleila Y, Shimizu Y, et al. Single—cell metabolomics by mass spectrometry: advances, challenges, and future applications. TrAC Trend Anal Chem 2019; 120: 115436.

[72] Minakshi P, Ghosh M, Kumar R, et al. Single—cell metabolomics: technology and applications. Single—Cell Omics 2019; 1(1): 319-353.

[73] Chen H, King FJ, Zhou B, et al. Drug target prediction through deep learning functional representation of gene signatures. Nat Commun 2024; 15(1): 1853.

[74] Lu Z, Song G, Zhu H, et al. DTIAM: a unified framework for predicting drug—target interactions, binding affinities and drug mechanisms. Nat Commun 2025; 16(1): 2548.

[75] Lai H, Wang L, Qian R, et al. Interformer: an interaction—aware model for protein—ligand docking and affinity prediction. Nat Commun 2024; 15(1): 10223.

[76] Bryant P, Kelkar A, Guljas A, et al. Structure prediction of protein—ligand complexes from sequence information with Umol. Nat Commun 2024; 15(1): 4536.

[77] Deng Y, Yao Y, Wang Y, et al. An end—to—end deep learning method for mass spectrometry data analysis to reveal disease—specific metabolic profiles. Nat Commun 2024; 15(1): 7136.

[78] Torun FM, Winter S V, Doll S, et al. Transparent exploration of machine learning for biomarker discovery from proteomics and omics data. J Proteome Res 2023; 22(2): 359-367.

[79] Bai P, Miljković F, John B, et al. Interpretable bilinear attention network with domain adaptation improves drug—target prediction. Nat Mach Intell 2023; 5(2): 126-136.

[80] Shah PM, Zhu H, Lu Z, et al. DeepDTAGen: a multitask deep learning framework for drug—target affinity prediction and target—aware drugs generation. Nat Commun 2025; 16(1): 5021.

[81] Zhang Y, Wang Y, Wu C, et al. Drug—target interaction prediction by integrating heterogeneous information with mutual attention network. BMC Bioinf 2024; 25(1): 361.

[82] Cheng Z, Zhao Q, Li Y, et al. IIFDTI: predicting drug—target interactions through interactive and independent features based on attention mechanism. Bioinformatics 2022; 38(17): 4153-4161.

[83] McNutt AT, Francoeur P, Aggarwal R, et al. GNINA 1.0: molecular docking with deep learning. J Cheminform 2021; 13(1): 43.

[84] Zhang JD, Xue C, Kolachalama VB, et al. Interpretable machine learning on metabolomics data reveals biomarkers for Parkinson’s disease. ACS Cent Sci 2023; 9(5): 1035-1045.

[85] Li Y, Liu X, Zhou J, et al. Artificial intelligence in traditional Chinese medicine: advances in multi—metabolite multi—target interaction modeling. Front Pharmacol 2025; 16: 1541509.

[86] Holvey RS, Erlanson DA, de Esch I JP, et al. Fragment—to—lead medicinal chemistry publications in 2023. J Med Chem 2025; 68(2): 986-1001.

[87] Xu W, Kang C. Fragment—based drug design: from then until now, and toward the future. J Med Chem 2025; 68(5): 5000-5004.

[88] Guo ZR. FBDD and drugs originated from FBDD. Acta Pharm Sin 2023; 58(12): 3490-3507.

[89] Zhang K, Yang X, Wang Y, et al. Artificial intelligence in drug development. Nat Med 2025; 31(1): 45-59.

[90] Yang Z, Shi S, Fu L, et al. Matched molecular pair analysis in drug discovery: methods and recent applications. J Med Chem 2023; 66(7): 4361-4377.

[91] Seo S, Lim J, Kim WY. Fragment—based molecular generative model with high generalization ability and synthetic accessibility. arXiv preprint arXiv:2111.12907 2021.

[92] Vásquez AF, Muñoz AR, Duitama J, et al. Non—extensive fragmentation of natural products and pharmacophore—based virtual screening as a practical approach to identify novel promising chemical scaffolds. Front Chem 2021; 9: 700802.

[93] Jinsong S, Qifeng J, Xing C, et al. Molecular fragmentation as a crucial step in the AI—based drug development pathway. Commun Chem 2024; 7(1): 20.

[94] Wang F, Liigand J, Tian S, et al. CFM—ID 4.0: more accurate ESI—MS/MS spectral prediction and compound identification. Anal Chem 2021; 93(34): 11692-11700.

[95] Hu M, Yang L, Twarog N, et al. Continuous collective analysis of chemical reactions. Nature 2024; 636(8042): 374-379.

[96] Cao L, Guler M, Tagirdzhanov A, et al. MolDiscovery: learning mass spectrometry fragmentation of small molecules. Nat Commun 2021; 12(1): 3718.

[97] Huang K, Xiao C, Glass LM, et al. MolTrans: molecular interaction transformer for drug—target interaction prediction. Bioinformatics 2021; 37(6): 830-836.

[98] Wang Q, Wang Q, Huang Q, et al. Five—layer—funnel filtering mode discovers effective components of Chinese medicine formulas: Zhishi—Xiebai—Guizhi decoction as a case study. Phytomedicine 2024; 129: 155678.

[99] Xie A, Zhang Z, Guan J, et al. Self—supervised learning with chemistry—aware fragmentation for effective molecular property prediction. Brief Bioinform 2023; 24(5): bbad296.

[100] Liu H, Zhu B, Nie S, et al. Advancing ADMET prediction through multiscale fragment—aware pretraining with MSformer—ADMET. Brief Bioinform 2025; 26(5): bbaf506.

[101] Zhu B, Liao J, Liu H, et al. MSformer: a meta—structure based interpretable framework for representation learning of natural products. Anal Chem 2025; 97: 25144.

[102] Lin S, Wang Y, Zhang L, et al. MDF—SA—DDI: predicting drug—drug interaction events based on multi—source drug fusion, multi—source feature fusion and transformer self—attention mechanism. Brief Bioinform 2022; 23(1): bbab421.

[103] Lanka G, Banerjee S, Adhikari N, et al. Fragment—based discovery of new potential DNMT1 inhibitors integrating multiple pharmacophore modeling, 3D—QSAR, virtual screening, molecular docking, ADME, and molecular dynamics simulation approaches. Mol Divers 2025; 29(1): 117-137.

[104] Liu Y, Zhang P, Che C, et al. SDDSynergy: learning important molecular substructures for explainable anticancer drug synergy prediction. J Chem Inf Model 2024; 64(24): 9551-9562.

[105] Wu JN, Wang T, Chen Y, et al. t—SMILES: a fragment—based molecular representation framework for de novo ligand design. Nat Commun 2024; 15(1): 4993.

[106] Li M, Zeng M, Zhang H, et al. Biological activity predictions of ligands based on hybrid molecular fingerprinting and ensemble learning. ACS Omega 2023; 8(6): 5561-5570.

[107] Chen B, Pan Z, Mou M, et al. Is fragment—based graph a better graph—based molecular representation for drug design? A comparison study of graph—based models. Comput Biol Med 2024; 169: 107811.

[108] Chen YJ, Zeng HS, Jin HL, et al. Applications of mass spectrometry imaging in botanical research. Adv Biotechnol (Singap) 2024; 2(1): 6.