EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Multimodal Transformer Fusion of PTR‐ToF‐MS Volatiles and Targeted Non‐Volatile Metabolites Enables Data‐Driven Grading of Baimudan White Tea

Dandan Zhang, Suk-Hwan Hong, Xiaojing Li

Rapid Communications in Mass Spectrometry · 30.09.2026 · 40 (18)

DOI: 10.1002/rcm.70122 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Rapid Communications in Mass Spectrometry, online erschienen am 28.06.2026. Verfasst von Dandan Zhang und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of International Tea Industry and Culture, Pusan National University, Gyeongnam, South Korea. Hauptthemen laut MEDLINE: Tea, Volatile Organic Compounds, Mass Spectrometry, Camellia sinensis. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Rationale

Commercial grading of Baimudan white tea still relies largely on sensory evaluation and would benefit from a rapid, objective analytical approach. This study investigated whether PTR‐ToF‐MS volatile fingerprints combined with targeted non‐volatile metabolite measurements could support data‐driven grading of four official Baimudan grades.

Methods

Headspace volatiles were profiled by proton‐transfer‐reaction time‐of‐flight mass spectrometry (PTR‐ToF‐MS), and non‐volatile metabolites were quantified by HPLC, amino acid analysis, and soluble sugar assay. A total of 120 samples (30 per grade) were divided into a training set ( n = 96) and an independent prediction set ( n = 24). A two‐stream Transformer encoded the volatile and non‐volatile modalities separately and fused them for four‐class classification.

Results

The multimodal model achieved 95.8% accuracy (23/24) and a macro‐F1 score of 0.958 on the independent prediction set, with a single misclassification between adjacent high grades. Repeated stratified five‐fold cross‐validation within the training set showed stable performance. SHAP and attention analyses indicated that high‐grade teas were associated with floral/sweet volatile ions and higher amino acids and soluble sugars, whereas lower grades were associated with greener/woody volatile ions and kaempferol‐related markers.

Conclusions

Within the current dataset, integrating PTR‐ToF‐MS volatile fingerprints with targeted non‐volatile metabolites provides an objective and chemically interpretable approach for instrument‐assisted grading of Baimudan white tea.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

TeaVolatile Organic CompoundsMass SpectrometryCamellia sinensisAmino AcidsChromatography, High Pressure LiquidCarbohydrates

Stichwörter der Autor:innen

Baimudan white teaPTR‐ToF‐MSnon‐volatile metabolitesquality gradingvolatile compounds

Substanzen

TeaVolatile Organic CompoundsAmino AcidsCarbohydrates

Autor:innen und Institutionen

  • Department of International Tea Industry and Culture, Pusan National University, Gyeongnam, South Korea.
  • Fujian Business University, Fuzhou, Fujian, China.
  • College of Environment and Resources, Fuzhou University, Fuzhou, Fujian, China.

Zitieren

APA

Zhang, D., Hong, S. H., & Li, X. (2026). Multimodal Transformer Fusion of PTR‐ToF‐MS Volatiles and Targeted Non‐Volatile Metabolites Enables Data‐Driven Grading of Baimudan White Tea. Rapid Communications in Mass Spectrometry, 40(18). https://doi.org/10.1002/rcm.70122

Vancouver

Zhang D, Hong SH, Li X. Multimodal Transformer Fusion of PTR‐ToF‐MS Volatiles and Targeted Non‐Volatile Metabolites Enables Data‐Driven Grading of Baimudan White Tea. Rapid Communications in Mass Spectrometry. 2026;40(18). doi:10.1002/rcm.70122

BibTeX
@article{Zhang2026,
  author = {Zhang, Dandan and Hong, Suk-Hwan and Li, Xiaojing},
  title = {Multimodal Transformer Fusion of PTR‐ToF‐MS Volatiles and Targeted Non‐Volatile Metabolites Enables Data‐Driven Grading of Baimudan White Tea},
  journal = {Rapid Communications in Mass Spectrometry},
  year = {2026},
  volume = {40},
  number = {18},
  doi = {10.1002/rcm.70122}
}
Kostenlos starten →