Non-Destructive Carotenoid Quantification in Leaves via Raman Spectroscopy: Optimizing Treatment for Linear Discriminant Analysis
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915464952676352 |
|---|---|
| author | Park, Miri Somborn, Annette Schlehuber, Dennis Lim, Hyun Jeong Keuter, Volkmar |
| author_facet | Park, Miri Somborn, Annette Schlehuber, Dennis Lim, Hyun Jeong Keuter, Volkmar |
| contents | This study introduces a novel method for quantifying challenging carotenoids in leaf tissues, which typically produce less stable signals than fruits, grains, and roots, by applying Linear Discriminant Analysis (LDA) modeling to interpret Raman spectroscopy data. The model's performance was assessed across different spectral preprocessing techniques (smoothing, normalization, baseline correction) and through various subsets of relevant Raman shifts. To generate a broad range of carotenoid contents, genetically modified Arabidopsis thaliana mutants with controlled synthesis and more conventional Spinacia oleracea samples under dark and salt stress were utilized, allowing for the evaluation of model robustness and practical applicability. Transition scores for Arabidopsis thaliana reached 77.27-95.45% in all quantifications, while Spinacia oleracea showed 75-83.33% in 3- and 4-level modeling, demonstrating the LDA model's strong potential for effective application. Among the spectral preprocessing, smoothing had the greatest impact on model performance, enhancing results for Arabidopsis thaliana but showing better outcomes without smoothing for Spinacia oleracea. Overall, this study highlights the potential of LDA modeling combined with Raman spectroscopy as a robust and non-destructive tool for metabolite quantification in herbal plants, with promising applications in agricultural monitoring and quality control. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_10057 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Non-Destructive Carotenoid Quantification in Leaves via Raman Spectroscopy: Optimizing Treatment for Linear Discriminant Analysis Park, Miri Somborn, Annette Schlehuber, Dennis Lim, Hyun Jeong Keuter, Volkmar Quantitative Methods This study introduces a novel method for quantifying challenging carotenoids in leaf tissues, which typically produce less stable signals than fruits, grains, and roots, by applying Linear Discriminant Analysis (LDA) modeling to interpret Raman spectroscopy data. The model's performance was assessed across different spectral preprocessing techniques (smoothing, normalization, baseline correction) and through various subsets of relevant Raman shifts. To generate a broad range of carotenoid contents, genetically modified Arabidopsis thaliana mutants with controlled synthesis and more conventional Spinacia oleracea samples under dark and salt stress were utilized, allowing for the evaluation of model robustness and practical applicability. Transition scores for Arabidopsis thaliana reached 77.27-95.45% in all quantifications, while Spinacia oleracea showed 75-83.33% in 3- and 4-level modeling, demonstrating the LDA model's strong potential for effective application. Among the spectral preprocessing, smoothing had the greatest impact on model performance, enhancing results for Arabidopsis thaliana but showing better outcomes without smoothing for Spinacia oleracea. Overall, this study highlights the potential of LDA modeling combined with Raman spectroscopy as a robust and non-destructive tool for metabolite quantification in herbal plants, with promising applications in agricultural monitoring and quality control. |
| title | Non-Destructive Carotenoid Quantification in Leaves via Raman Spectroscopy: Optimizing Treatment for Linear Discriminant Analysis |
| topic | Quantitative Methods |
| url | https://arxiv.org/abs/2504.10057 |