Work

Characterization and Authenticity Identification of Caryophylli flos from Different Geographical Origins Using HS-GC-IMS Combined with Deep Learning

Year: 2026
Type: article
Abstract: Caryophylli flos (CF) is an herbal medicine with medicine and food homology in China, and has been cultivated in multiple regions. At present, effective and reliable methods to identify CF samples from different origins are lacking. In this study, gas chromatography-ion mobility spectrometry (GC-IMS) was applied to analyze volatile components (VOCs) of CF samples, and a total of 54 VOCs were identified. CF samples from Guangxi exhibited relatively higher levels of VOCs than those from other regions. Moreover, eleven differential VOCs were identified, such as (E,E)-2,4-hexadienal, 3-hydroxy-2-butanone, (+)-limonene, β-myrcene and α-pinene. Subsequently, an integrated strategy that combines GC-IMS 2D chromatograms with deep learning algorithms, specifically convolutional neural networks (CNN) and residual networks (ResNet), was proposed. In a preliminary internal hold-out test, this approach achieved an identification accuracy reaching 100% for CF geographical origin, surpassing conventional machine learning models; this value is not an unbiased estimate of generalization. Collectively, this work provides a reference for VOC studies on CF samples from different origins and establishes a rapid, simple, and low-cost strategy for accurate identification of geographical origin in China for CF samples, thereby ensuring the quality uniformity of CF samples. In the future, more samples from geographical sources will be included to improve the generalization ability of the model.
Source: Foods
Language: English
Field: Chemistry
Open Access status: gold
Estimated APC paid: $3,499
OpenAlex ID: W7234408306
FWCI:
Cites: 45
Cited by:
Related to: 10