Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Sep 19;4(9):101172.
doi: 10.1016/j.xcrm.2023.101172. Epub 2023 Aug 30.

Population serum proteomics uncovers a prognostic protein classifier for metabolic syndrome

Xue Cai  1 , Zhangzhi Xue  1 , Fang-Fang Zeng  2 , Jun Tang  3 , Liang Yue  4 , Bo Wang  5 , Weigang Ge  5 , Yuting Xie  4 , Zelei Miao  3 , Wanglong Gou  3 , Yuanqing Fu  3 , Sainan Li  4 , Jinlong Gao  4 , Menglei Shuai  3 , Ke Zhang  3 , Fengzhe Xu  3 , Yunyi Tian  3 , Nan Xiang  5 , Yan Zhou  4 , Peng-Fei Shan  6 , Yi Zhu  7 , Yu-Ming Chen  8 , Ju-Sheng Zheng  9 , Tiannan Guo  10
Affiliations

Population serum proteomics uncovers a prognostic protein classifier for metabolic syndrome

Xue Cai et al. Cell Rep Med. .

Abstract

Metabolic syndrome (MetS) is a complex metabolic disorder with a global prevalence of 20%-25%. Early identification and intervention would help minimize the global burden on healthcare systems. Here, we measured over 400 proteins from ∼20,000 proteomes using data-independent acquisition mass spectrometry for 7,890 serum samples from a longitudinal cohort of 3,840 participants with two follow-up time points over 10 years. We then built a machine-learning model for predicting the risk of developing MetS within 10 years. Our model, composed of 11 proteins and the age of the individuals, achieved an area under the curve of 0.774 in the validation cohort (n = 242). Using linear mixed models, we found that apolipoproteins, immune-related proteins, and coagulation-related proteins best correlated with MetS development. This population-scale proteomics study broadens our understanding of MetS and may guide the development of prevention and targeted therapies for MetS.

Keywords: DIA-MS; apolipoproteins; blood proteomics; diabetes; high-density lipoproteins; machine learning; metabolic syndrome; population proteomics; prospective prediction; triglycerides.

PubMed Disclaimer

Conflict of interest statement

Declaration of interests Y. Zhu and T.G. are shareholders of Westlake Omics, Inc. B.W., W.G., and N.X. are employees of Westlake Omics, Inc. A patent related to this study has been submitted to the China National Intellectual Property Administration.

Figures

None
Graphical abstract
Figure 1
Figure 1
Study design GNHS study population: 7,890 samples (4,796 in the discovery cohort, 3,094 in the validation cohort) were collected from 3,840 individuals. Altogether, we generated 18,970 DIA-MS data. After data cleaning, the protein matrices (containing 4,637 samples and 438 proteins for the discovery cohort and 3,067 samples and 413 proteins for the validation cohort) were used for data mining and machine-learning modeling.
Figure 2
Figure 2
QC analyses and dataset pre-processing (A) Pearson correlation of the MS-QC samples from the discovery and validation cohorts. (B) Batch effect evaluation using principal-component analysis (PCA) and the QC samples from the discovery and validation cohorts. (C) Pearson correlation of the technical and biological replicates from the discovery and validation cohorts. Samples were processed twice as biological replicates and were acquired twice by MS as technical replicates.
Figure 3
Figure 3
Machine-learning model for predicting the risk of developing MetS (A) Case and control samples for the machine-learning model. (B) Workflow of the machine-learning model built with quantitative proteomics data and clinical features. (C) Shapley importance of variables in the model. (D) Receiver operating characteristic (ROC) plot of the machine-learning model.
Figure 4
Figure 4
Performance of new potential protein biomarkers of MetS and their related pathways (A) Dysregulated proteins highly associated with MetS were selected using linear mixed models. The inner circle represents dysregulated proteins of the baseline model, while the outer circle represents dysregulated proteins of the dynamic model. The cutoff of dysregulated proteins has been set at B-H adjusted p value < 0.05. (B) Enriched pathways analyzed by IPA of 175 and 194 significantly dysregulated proteins. The size of circle represents the −log10(p value) and the color represents the ratio, which is calculated by dividing the number of analysis-ready molecules by the total number of molecules within the pathway. (C) Top (|β| > 0.2) significantly dysregulated proteins associated with MetS using the linear mixed model. (D) Expression of typical dysregulated proteins in discovery cohort and validation cohort. B-H adjusted p values of these proteins are < 0.001.
Figure 5
Figure 5
Dysregulated proteins in different types of MetS (A) Volcano plot of the dysregulated proteins in TG-based MetS. The cutoff of dysregulated proteins has been set at B-H adjusted p value < 0.05, |log2(fold change)| > 0.2. (B) Volcano plot of the dysregulated proteins in HDL-based MetS. The cutoff of dysregulated proteins has been set at B-H adjusted p value < 0.05, |log2(fold change)| > 0.2. (C) Expression of the dysregulated proteins in TG-based MetS and non-MetS. (D) Expression of the dysregulated proteins in HDL-based MetS and non-MetS. See also Figure S2.
Figure 6
Figure 6
Dysregulated proteins in the development of MetS (A) Volcano plot and boxplots of the dysregulated proteins in the development of waist-based MetS. The cutoff of dysregulated proteins has been set at B-H adjusted p value < 0.05, |log2(fold change)| > 0.2. (B) Volcano plot and boxplots of the dysregulated proteins in the development of TG-based MetS. The cutoff of dysregulated proteins has been set at B-H adjusted p value < 0.05, |log2(fold change)| > 0.2. (C) Volcano plot and boxplots of the dysregulated proteins in the development of SBP-/DBP-based MetS. The cutoff of dysregulated proteins has been set at B-H adjusted p value < 0.05, |log2(fold change)| > 0.2. (D) Boxplots of the overlapping core dysregulated proteins in the development of waist-based MetS, TG-based MetS, and SBP-/DBP-based MetS. The numbers between the paired box plots represent the B-H adjusted p values.
See this image and copyright information in PMC

References

    1. Expert Panel on Detection, E., and Treatment of High Blood Cholesterol in, Expert Panel on Detection Evaluation and Treatment of High Blood Cholesterol in Adults Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III) JAMA. 2001;285:2486–2497. doi: 10.1001/jama.285.19.2486. - DOI - PubMed
    1. Alberti K.G.M.M., Eckel R.H., Grundy S.M., Zimmet P.Z., Cleeman J.I., Donato K.A., Fruchart J.-C., James W.P.T., Loria C.M., Smith S.C., et al. Harmonizing the Metabolic Syndrome. Circulation. 2009;120:1640–1645. doi: 10.1161/CIRCULATIONAHA.109.192644. - DOI - PubMed
    1. International Diabetes Federation: The IDF consensus worldwide definition of the metabolic syndrome.
    1. Ranasinghe P., Mathangasinghe Y., Jayawardena R., Hills A.P., Misra A. Prevalence and trends of metabolic syndrome among adults in the asia-pacific region: a systematic review. BMC Publ. Health. 2017;17:101. doi: 10.1186/s12889-017-4041-1. - DOI - PMC - PubMed
    1. do Vale Moreira N.C., Hussain A., Bhowmik B., Mdala I., Siddiquee T., Fernandes V.O., Montenegro Júnior R.M., Meyer H.E. Prevalence of Metabolic Syndrome by different definitions, and its association with type 2 diabetes, pre-diabetes, and cardiovascular disease risk in Brazil. Diabetes Metabol. Syndr. 2020;14:1217–1224. doi: 10.1016/j.dsx.2020.05.043. - DOI - PubMed

Publication types