天津医药 ›› 2026, Vol. 54 ›› Issue (11): 1150-1156.doi: 10.11958/20260401

• 临床研究 • 上一篇    下一篇

基于代谢组学筛选结核病脂质代谢物及相关代谢通路

连家伟1(), 孙高峰2, 王泉3, 李奇凤2,△()   

  1. 1 新疆医科大学基础医学院病原生物学教研室 (邮编 830017)
    2 新疆维吾尔自治区第七人民医院/北京儿童医院新疆医院/新疆维吾尔自治区儿童医院/新疆维吾尔自治区儿科研究所科教科
    3 新疆维吾尔自治区第六人民医院/新疆维吾尔自治区传染病医院检验科
  • 收稿日期:2026-02-03 修回日期:2026-06-19 出版日期:2026-11-15 发布日期:2026-11-10
  • 通讯作者: 李奇凤 E-mail:17776934397@163.com;liqiqi521@sina.com
  • 作者简介:连家伟(2000),男,硕士在读,主要从事肺结核方面研究。E-mail:17776934397@163.com
  • 基金资助:
    新疆维吾尔自治区重点研发任务专项(2024B03021-2);国家自然科学基金地区科学基金项目(82460003)

Screening of lipid metabolites and related metabolic pathways in tuberculosis based on metabolomics

LIAN Jiawei1(), SUN Gaofeng2, WANG Quan3, LI Qifeng2,△()   

  1. 1 Department of Pathogen Biology, School of Basic Medical Sciences, Xinjiang Medical University, Urumqi 830017, China
    2 Department of Science and Education, the Seventh People's Hospital of Xinjiang Uygur Autonomous Region / Beijing Children's Hospital Xinjiang Hospital / Xinjiang Uygur Autonomous Region Children's Hospital / Xinjiang Uygur Autonomous Region Institute of Pediatrics
    3 Department of Clinical Laboratory, the Sixth People's Hospital of Xinjiang Uygur Autonomous Region / Xinjiang Uygur Autonomous Region Infectious Disease Hospital
  • Received:2026-02-03 Revised:2026-06-19 Published:2026-11-15 Online:2026-11-10
  • Contact: LI Qifeng E-mail:17776934397@163.com;liqiqi521@sina.com

摘要:

目的 通过整合公共转录组数据与血浆脂质组学结果,筛选结核病(TB)相关的潜在脂质代谢标志物,并分析其相关代谢通路。方法 基于基因表达综合数据库(GEO)中TB转录组数据筛选差异表达基因,并与基因卡(GeneCards)数据库中脂质代谢相关基因取交集,获得交集基因。同时,收集74例结核病患者(结核组)与106例健康志愿者(健康对照组),采用超高效液相色谱-高分辨质谱联用(UHPLC-HRMS)技术检测2组血浆脂质代谢物。对交集基因与差异脂质代谢产物进行联合通路分析,用最小绝对收缩和选择算子(LASSO)回归和随机森林算法交叉筛选出潜在差异脂质代谢产物,并通过受试者工作特征(ROC)曲线评估其区分能力。结果 共获得473个交集基因,主要富集于脂肪酸代谢、甘油磷脂代谢及鞘脂代谢等通路;血浆脂质组学鉴定出54个差异脂质代谢产物(35个上调、19个下调),主要富集于脂肪酸氧化、类固醇生成及胆汁酸生物合成等通路;联合通路分析筛选出20条共同关联通路。经LASSO回归和随机森林算法交叉筛选获得3-氧代-7α,12α-二羟基-5β-胆烷酸、溶血磷脂酰胆碱18:2、束状酸C、5α-四氢皮质醇、磷脂酰肌醇36:4共5个潜在脂质代谢物,ROC曲线显示其曲线下面积(AUC)均>0.9。结论 筛选出的5个与TB相关的潜在脂质代谢物涉及类固醇激素代谢、胆汁酸代谢及甘油磷脂代谢等通路。

关键词: 结核, 代谢组学, 脂质代谢, 生物标志物, 转录组学

Abstract:

Objective To identify potential lipid metabolic biomarkers associated with tuberculosis (TB) and analyze their related metabolic pathways by integrating public transcriptomic data with plasma lipidomics profiles. Methods Differentially expressed genes were screened from tuberculosis transcriptomic data in the Gene Expression Omnibus (GEO) database and intersected with lipid metabolism-related genes from the GeneCards database to obtain overlapping genes. Meanwhile, 74 TB patients (the tuberculosis group) and 106 healthy controls (the healthy control group) were recruited. Plasma lipid metabolites were detected using ultra-high-performance liquid chromatography-high-resolution mass spectrometry (UHPLC-HRMS). On this basis, joint pathway analysis of overlapping genes and differential lipid metabolites was performed, and potential differential lipid metabolites were screened by cross-validation using least absolute shrinkage and selection operator (LASSO) regression and random forest algorithm, with their discriminatory performance evaluated by receiver operating characteristic (ROC) curves. Results A total of 473 overlapping genes were obtained, which were mainly enriched in fatty acid metabolism, glycerophospholipid metabolism, sphingolipid metabolism and related pathways. Plasma lipidomics identified 54 differential lipid metabolites (35 upregulated and 19 downregulated), mainly enriched in fatty acid oxidation, steroidogenesis and bile acid biosynthesis pathways. Joint pathway analysis revealed 20 commonly associated pathways. Five potential lipid metabolic biomarkers were screened by cross-validated least absolute shrinkage and selection operator (LASSO) regression and random forest analysis:3-oxo-7α, 12α-dihydroxy-5β-cholanoic acid, lysophosphatidylcholine 18:2 (LysoPC 18:2), fasciculic acid C, 5α-tetrahydrocortisol (5α-THF) and phosphatidylinositol 36:4 (PI 36:4), with ROC curves demonstrating AUC values all >0.9. Conclusion Five potential lipid metabolic biomarkers associated with TB are identified, involving steroid hormone metabolism, bile acid metabolism, glycerophospholipid metabolism and related pathways.

Key words: tuberculosis, metabolomics, lipid metabolism, biomarkers, transcriptomics

中图分类号: