Metabolites, the small molecules generated and transformed during metabolism, provide a direct functional readout of cellular biochemistry. Mass spectrometry–based metabolomics enables the quantitative measurement of thousands of metabolites from minimal amounts of biological material, thereby supporting systems-level analyses. Through global metabolomic profiling, emerging discoveries are linking cellular pathways to biological mechanisms and reshaping our understanding of cell biology, physiology, and medicine. Although metabolomics is a relatively young field compared with genomics and proteomics, it has already delivered clinically relevant biomarkers for disease diagnosis, fundamental insights into cellular biochemistry, and important clues to disease pathogenesis.
The research of Dr. Zhu’s group focuses on the development of mass spectrometry–based metabolomics and lipidomics technologies and their applications in elucidating the mechanisms of aging and aging-related diseases. In recent years, the group’s major academic achievements can be summarized as follows.
1) Metabolomics data processing and annotation
To enable fast, robust, and convenient mass spectrometry data processing for metabolomics and lipidomics, we developed Met4DX, a versatile software tool that supports the processing of both three-dimensional LC-MS data and four-dimensional LC-IM-MS data. Met4DX provides key functions including data conversion, peak detection, retention time correction, peak grouping, MS/MS spectral assignment, and metabolite and lipid identification, among others. Met4DX is freely available on our website (http://met4dx.zhulab.cn/) (Nature Commun., 2023; J. Am. Soc. Mass Spectrom., 2024).
To support metabolite annotation, we have developed a metabolic reaction network (MRN)-based recursive algorithm (MetDNA; http://metdna.zhulab.cn) that expands metabolite annotations without the need for a comprehensive standard spectral library (Nature Commun., 2019). We demonstrated that MetDNA enables to identify 5-10 folds more metabolites than other tools from one experiment. MetDNA also supports metabolite annotation acquired with data independent acquisition (DIA) MS technology (Anal. Chem., 2019). We have futher developed a multi-layer networking approach, knowledge-guided multi-layer metabolic networking (KGMN), to support large-scale unknown metabolite annotation within MetDNA2 (Nature Commun., 2022a). Recently, we further developed a two-layer interactive networking strategy that integrates data-driven and knowledge-driven networks to enhance metabolite annotation within MetDNA3 (Nature Commun., 2025b).
2) Ion mobility-resolved metabolomics/lipidomics and single-cell technologies
We have developed a large-scale ion mobility CCS atlas AllCCS (http://allccs.zhulab.cn)(Nature Commun., 2020; Anal. Chem., 2023), which enables confident metabolite annotation, and a variety of four-dimensional (4D) metabolomics and lipidomics technologies which support the comprehensive profiling of metabolites and lipids with high accuracy and broad coverage (Bioinformatics., 2019; Anal. Chim. Acta., 2020, 2022, Anal. Chem, 2022). To demonstrate its capability for analyses of isomeric metabolites, we also developed an IM-MS based four-dimensional sterolomics technology by leveraging a machine learning-empowered high-coverage library (>2,000 sterol lipids) for accurate sterol identification (Nature Commun., 2021). Very recently, we have developed a mass spectrum-oriented computational method, namely, Met4DX, for efficiently processing ion mobility-resolved 4D untargeted metabolomics with high coverage (Nature Commun., 2023).
Ion mobility-resolved mass cytometry for single-cell metabolomics: current single-cell metabolomics approaches are limited by insufficient sensitivity, robustness, and metabolite coverage. We developed an ion mobility–resolved mass cytometry technology that integrates high-throughput single-cell injection with ion mobility–mass spectrometry for multidimensional metabolomic profiling. Combined with our computational tool, MetCell, this technology allows high-throughput analysis while achieving exceptional profiling depth, detecting over 5,000 metabolic peaks and annotating approximately 800 metabolites per cell (Nature Methods, 2026).
3) Stable-isotope tracing metabolomics
Stable-isotope tracing metabolomics allows to unravel metabolic activity quantitatively by measuring the isotopically labeled metabolites, but has been largely restricted by coverage. To address this challenge, we have developed a technology, termed MetTracer, leveraging the advantages of untargeted metabolite annotation and targeted extraction to trace the isotope labeled metabolites in complex matrices globally (Nature Commun., 2022b). In addtion, using mass spectrometry-resolved stable-isotope tracing metabolomics, we develop an isotopologue similarity networking strategy, namely IsoNet, to effectively deduce previously unknown metabolites and related metabolic reactions (Nature Commun., 2025a).
