Scholarlink

<p>The geographical provenance of agricultural products is a key determinant of quality and economic value. An untargeted metabolomics-based ML pipeline was established to authenticate the origin of Yongfu Luohan Guo. LC-QTOF/MS profiling of samples from multiple origins generated comprehensive metabolite datasets, which were analyzed using supervised ML algorithms. Multivariate analysis revealed distinct metabolic signatures differentiating Yongfu and non-Yongfu Luohan Guo. The CNN algorithm outperformed other models, achieving 100% classification accuracy using 8 key metabolites These biomarkers include two flavonoids: 5-hydroxy-3-(5-hydroxy-2,4-dimethoxyphenyl)-6-methoxy-7-[3,4,5-trihydroxy-6-(hydroxymethyl)oxan-2-yl]oxychromen-4-one and glabrol; three terpenoids: rhodojaponin III, lapidin, and acetylcimigenol arabinoside; and three alkaloids: corynoxeine, indirubin, and hydroquinidine. Region-specific metabolite accumulation trends were observed in Yongfu Luohan Guo, correlating with fruit quality attributes. This metabolomics-ML integration strategy presents a scalable solution for geographical indication authentication of diverse agricultural products.</p> Graphical Abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

LC-QTOF/MS-Based Non-targeted Metabolomics Combined with Machine Learning Algorithms for Geographical Origin Discrimination of Yongfu Luohan Guo (Siraitia grosvenorii)

  • 期刊论文

  • 发表日期：

    2026年1月31日

查看全文
  • Weichao Sun,
  • Huahong Liu,
  • Yuxin Wu,
  • Yuqi Qin,
  • Yu Zhang,
  • Wenying Jin,
  • Hong Hai,
  • Yun Zhang,
  • Jinfang Nie,
  • Haiyan Fu

摘要

The geographical provenance of agricultural products is a key determinant of quality and economic value. An untargeted metabolomics-based ML pipeline was established to authenticate the origin of Yongfu Luohan Guo. LC-QTOF/MS profiling of samples from multiple origins generated comprehensive metabolite datasets, which were analyzed using supervised ML algorithms. Multivariate analysis revealed distinct metabolic signatures differentiating Yongfu and non-Yongfu Luohan Guo. The CNN algorithm outperformed other models, achieving 100% classification accuracy using 8 key metabolites These biomarkers include two flavonoids: 5-hydroxy-3-(5-hydroxy-2,4-dimethoxyphenyl)-6-methoxy-7-[3,4,5-trihydroxy-6-(hydroxymethyl)oxan-2-yl]oxychromen-4-one and glabrol; three terpenoids: rhodojaponin III, lapidin, and acetylcimigenol arabinoside; and three alkaloids: corynoxeine, indirubin, and hydroquinidine. Region-specific metabolite accumulation trends were observed in Yongfu Luohan Guo, correlating with fruit quality attributes. This metabolomics-ML integration strategy presents a scalable solution for geographical indication authentication of diverse agricultural products.

Graphical Abstract