Efficient generation of HPLC and FTIR data for quality assessment using time series generation model: a case study on Tibetan medicine Shilajit – PubMed Black Hawk Supplements

BLACK HAWK: Most trusted shilajit supplement for sleep

Published article

CONCLUSION: This study offers a novel and effective approach for researching medicinal materials with small sample sizes, and addresses the limitations of improving model performance through data augmentation strategies.
Black Hawk Supplements, best supplements in the UK

Efficient generation of HPLC and FTIR data for quality assessment using time series generation model: a case study on Tibetan medicine Shilajit - PubMed

Efficient generation of HPLC and FTIR data for quality assessment using time series generation model: a case study on Tibetan medicine Shilajit

Rong Ding et al. Front Pharmacol. .

Abstract

Background: The scarcity and preciousness of plateau characteristic medicinal plants pose a significant challenge in obtaining sufficient quantities of experimental samples for quality evaluation. Insufficient sample sizes often lead to ambiguous and questionable quality assessments and suboptimal performance in pattern recognition. Shilajit, a popular Tibetan medicine, is harvested from high altitudes above 2000 m, making it difficult to obtain. Additionally, the complex geographical environment results in low uniformity of Shilajit quality.

Methods: To address these challenges, this study employed a deep learning model, time vector quantization variational auto- encoder (TimeVQVAE), to generate data matrices based on chromatographic and spectral for different grades of Shilajit, thereby increasing in the amount of data. Partial least squares discriminant analysis (PLS-DA) was used to identify three grades of Shilajit samples based on original, generated, and combined data.

Results: Compared with the originally generated high performance liquid chromatography (HPLC) and Fourier transform infrared spectroscopy (FTIR) data, the data generated by TimeVQVAE effectively preserved the chemical profile. In the test set, the average matrices for HPLC, FTIR, and combined data increased by 32.2%, 15.9%, and 23.0%, respectively. On the real test data, the PLS-DA model’s classification accuracy initially reached a maximum of 0.7905. However, after incorporating TimeVQVAE-generated data, the accuracy significantly improved, reaching 0.9442 in the test set. Additionally, the PLS-DA model trained with the fused data showed enhanced stability.

Conclusion: This study offers a novel and effective approach for researching medicinal materials with small sample sizes, and addresses the limitations of improving model performance through data augmentation strategies.

Keywords: FTIR; HPLC; Shilajit; classification; time series generation.

PubMed Disclaimer

Conflict of interest statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Figures

FIGURE 1
FIGURE 1

Overview of our proposed VQ (i.e., tokenization) (stage 1, The encoder and the decoder are denoted by E and D respectively. STFT and ISTFT stand for Short-time Fourier Transform and Inverse Short-time Fourier Transform, respectively.).

FIGURE 2
FIGURE 2

(A) Overview of the prior model training (stage 2). (B) Overview of MaskGIT’s iterative decoding. The TopK operation is equivalent to torch.topk from PyTorch. (The dark green block represents the [MASK] token.).

FIGURE 3
FIGURE 3

Principles of dataset partitioning.

FIGURE 4
FIGURE 4

The second-order guide spectrogram of the infrared spectrum of medicinal materials, H is high quality, M is medium quality, and L is low quality.

FIGURE 5
FIGURE 5

Loss variation for training and validation sets with different data sources.

FIGURE 6
FIGURE 6

Visualization of generated results at three epochs (13, 623, and 1989) during the training process. The blue line represents the original time-series data, the yellow line represents the predicted data. x_l refers to the low-frequency component of the time-series signal, x_h refers to the high-frequency component of the time-series signal, x represents the time series data generated or predicted. (A) HPLC. (B) FTIR. (C) HPLC_FTIR.

FIGURE 7
FIGURE 7

PCA and T-SNE visual mapping of test samples (blue points) and generated samples (yellow points). (A) PCA original data direct mapping. (B) PCA mapping of data in latent space. (C) T-SNE original data direct mapping. (D) T-SNE mapping of data in latent space. (E) HPLC. (F) FTIR. (G) HPLC_FTIR.

FIGURE 8
FIGURE 8

The mean and standard deviation of the generated data and the real data, H is high quality, M is medium quality, and L is low quality. (A) HPLC. (B) Generated HPLC. (C) FTIR. (D) Generated FTIR.

FIGURE 9
FIGURE 9

Selection of the number of latent variables through 10-Fold cross-validation, the red line and the blue line are the accuracy of the model on the training set and the average accuracy on the cross-validation set under different numbers of latent variables. (A) HPLC. (B) FTIR. (C) HPLC-FTIR. (D) HPLC-TimeVQVAE. (E) FTIR-TimeVQVAE. (F) HPLC_FTIR-TimeVQVAE.

FIGURE 10
FIGURE 10

Analysis of Explained variance in X and Y Blocks in PLS-DA model. (A) HPLC. (B) FTIR. (C) HPLC-FTIR. (D) HPLC-TimeVQVAE. (E) FTIR-TimeVQVAE. (F) HPLC-FTIR-TimeVQVAE.

FIGURE 11
FIGURE 11

Confusion matrix and ROC curves on the test set, H is high quality, M is medium quality, and L is low quality. (A) HPLC. (B) FTIR; (C) HPLC-FTIR; (D) HPLC-TimeVQVAE; (E) FTIR-TimeVQVAE; (F) HPLC-FTIR-TimeVQVAE.

Similar articles

References

    1. Adib E., Fernandez A. S., Afghah F., Prevost J. J. (2023). Synthetic ECG Signal generation using probabilistic diffusion models. IEEE Access 11, 75818–75828. 10.1109/ACCESS.2023.3296542 – DOI
    1. Agarwal S. P., Khanna R., Karmarkar R., Anwer M. K., Khar R. K. (2007). Shilajit: a review. Phytotherapy Res. 21, 401–405. 10.1002/ptr.2100 – DOI – PubMed
    1. Bhavsar S. K., Thaker A. M., Malik J. K. (2016). in Chapter 51 – Shilajit. Nutraceuticals . Editor Gupta R. C. (Boston: Academic Press; ), 707–716. 10.1016/B978-0-12-802147-7.00051-6 – DOI
    1. Borràs E., Ferré J., Boqué R., Mestres M., Aceña L., Busto O. (2015). Data fusion methodologies for food and beverage authentication and quality assessment – a review. Anal. Chim. Acta 891, 1–14. 10.1016/j.aca.2015.04.042 – DOI – PubMed
    1. Borràs E., Ferré J., Boqué R., Mestres M., Aceña L., Calvo A., et al. (2016). Olive oil sensory defects classification with data fusion of instrumental techniques and multivariate analysis (PLS-DA). Food Chem. 203, 314–322. 10.1016/j.foodchem.2016.02.038 – DOI – PubMed

Grants and funding

The author(s) declare that financial support was received for the research, authorship, and/or publication of this article. This work was supported by the supported by the National Natural Science Foundation of China (Grant number: 82274208).

BLACK HAWK: Best shilajit supplement for elderly

Read the original publication:

Efficient generation of HPLC and FTIR data for quality assessment using time series generation model: a case study on Tibetan medicine Shilajit – PubMed