Review Reports - Chemometric Approaches for Identification of Herbal Medicinal Products

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

There are many unclearly parts in the manuscript. The authors need to explain what samples are used to generate the PCA analysis. The sample sizes for HMPs are very small, therefore the PCA model may not be useful for product variation due to different country origins and production seasons.

In many of the figures, when mouse is hovered over the figure, the message “AI-generated content may be incorrected.” appeared. The authors mentioned using OriginPro 2021 for statistical analysis. Please explain why this message appears many times.
Please describe clearly the pharmacological group of the HMPs for Figure 3.
Please describe clearly the pharmacological group of the HMPs for Figure 6.
Please describe clearly the samples used for PCA modeling in Figure 7.
Please describe clearly the pharmacological group of the HMPs for Figure 8.
Please describe clearly the samples used for PCA modeling in Figure 10.
Please describe clearly the samples used for PCA modeling in Figure 12.
Only FTIR spectra is shown for 14 samples, please show the FTIR spectra for all the samples.
UV spectra is not shown. Please show the spectra in the manuscript.
Fluorescence spectra is not shown. Please show the spectra in the manuscript.
XRF spectra is not shown. Please show the spectra in the manuscript.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This study demonstrates the instrumental analysis and chemometric evaluation of several medical products prepared from calendula flowers, eucalyptus leaves, oak bark, thyme herbs, valerian rhizomes with roots, motherwort herbs, and hawthorn berries. Tinctures and dried extracts were prepared, and then the products were analyzed using several analytical methods, including FTIR spectrometry, UV spectrophotometry, Spectrofluorimetry, and X-ray fluorescence. Additionally, attempts were made to perform 2D-DLS measurements on some tinctures. The authors used PCA for pattern recognition/qualification between different samples. The manuscript requires substantial revision in several key areas to improve its clarity, structure, and scientific rigor. Both the content and presentation should be refined to ensure that the study’s objectives, methodology, and conclusions are clearly and coherently conveyed.

Comments:

In general, the study integrates the results of multiple analytical methods, which is highly progressive and represents a truly modern approach compared to traditional HPLC techniques. However, the description and implementation of the chemometric methods appear rather superficial. PCA is one of the most widely used pattern recognition techniques. What data preprocessing methods were applied to the raw spectral data? Have the authors tried other methods, such as PLS-DA?
The PCA method effectively demonstrates that clear differentiation can be observed between distinct botanical genera in the score plots. However, what happens when the samples are more similar—for instance, when dealing with botanically related genera or when the product quality is compromised due to dilution or substitution of certain components with other materials or due to a change of other CQAs?
Furthermore, no information is provided regarding the validation of the PCA models. Without validation, how reliable are these models for actual quality control applications? It would be advisable to test this by mixing or diluting samples to assess model robustness and effectiveness.
What is the likelihood of obtaining false-positive results under such conditions?
In what arrangement were the 2D-DLS measurements performed? It is not easy to imagine how the laser passes through the packaged bottle and how the detection happened.
How can it be explained that the value of the d₁ descriptor can either decrease or increase compared to 70% alcohol, depending on the type of product?

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This study shows limited external validation and small sample diversity, focusing on a narrow range of species and manufacturers. The reliance on laboratory-prepared tinctures may not reflect industrial-scale variability. Also, the paper can improve its statistical robustness and quantitative comparison with conventional chromatographic standards. Overall, it represents a promising yet preliminary advancement toward chemometric-based quality control of herbal medicines.

Figures and Tables are appropriate but their interpretations to align with the objective are superficially explained. Refer to individual comments of figures and tables.

The “Results” section tells what happened, not why it matters or how it connects to existing research. Therefore a separate “Discussion” section is required to interpret findings beyond data presentation, to compare with existing literature, to acknowledge limitations and to connect to real-world implications.

Abstract:
The abstract lacks clarity on specific findings, such as accuracy rates, validation outcomes, or comparative advantages over conventional techniques. It can be improved by including quantitative results, a clear statement of significance or practical impact, and a concise conclusion emphasizing the method’s reliability and industrial applicability. A smoother narrative flow between methods and outcomes will also enhance readability and comprehension.

Introduction:

Line 43-44: Establish a link why these traditional methods (pharmacopeial parameters like ethanol content and methanol impurities) fall short to motivate exploration of broader, chemometric-based identification methods.
Line 57-59: Add a connection for the need of elemental profiling to the paper’s broader goal of developing reference-free identification methods.

2.1. Materials:

Line 90-91: Connect pharmacological groupings to the study’s purpose e.g., it could note why these specific herbs were selected, perhaps for their chemical diversity or availability of tincture forms.
Line 103-110: This paragraph reads more like a pharmacopoeial procedure than a methodological narrative. Add a few linking sentences to clarify why both commercial and lab tinctures were compared.

2.2.1. Principal Component Analysis (PCA):

Line 148-155: The paragraph is misaligned in focus and drifts into certification details that can be condensed. The long discussion of international calibration feels unimportant to the PCA method itself. Add a sentence about why such calibration strengthens data reliability.

Results:

This paragraph is rephrasing earlier material from the introduction rather than transitioning into results. In a results section, readers expect at least some mention of data trends, validation findings, or outcomes from the PCA or DLS analyses.

3.1. Processing Spectroscopic Data Using PCA for the Identification of herbal medicinal products and medicinal tinctures

Line 270-272: The mention of “blind samples” is good for specificity testing but it lacks detail on how many blind samples were tested or statistical confidence in classification. Also, include quantifying predictive accuracy or misclassification rate to substantiate PCA’s reliability as an identification tool.

3.2. Express identification of tinctures using the 2D-DLS method:

Line 361-362: Include reasoning to explain why six out of ten descriptors were chosen as the threshold. Was this empirically determined or statistically validated?

Conclusion:

The conclusion feels more like a recap than a synthesis. It lacks a sense of broader implication or forward direction. For example, mention how this work could be integrated into regulatory practice or industrial QA systems, or what limitations or future improvements might enhance chemometric reliability across more diverse HMPs.

Figures and Tables:

Figure 1: Mention how particle size uniformity influences signal quality or reproducibility in PCA datasets.
Figures 3-7: Include a sentence or two interpreting how characteristic peaks (1400, 931, 723 cm⁻¹) might correspond to molecular groups distinguishing species.
Figures 8-10: Include a brief rationale for why these particular wavelengths dominate (e.g., associations with chromophoric or fluorophoric compounds common to certain botanical groups). For the XRF interpretation in Figures 9–10), mention of how elemental differences biologically relate to plant types (e.g., high K/Cl ratio in Valeriana species).
Figure 11: Include interpretive note linking observed elemental variance (e.g., Fe, Zn, Mn levels) to possible phytochemical or soil correlations.
Figure 12: Include a quantitative reference (e.g., “clusters separated by >95% confidence interval”)
Figures 13-14: Include a comment on how DLS descriptors relate to particle density, color, or microstructure to highlight the novelty of the approach.
Table 1: Provide a short comparative note (e.g., “Among all descriptors, d1 and R showed highest discriminative stability”) to turn this from a theoretical insert into a meaningful analytical tool.
Table 2: Include an explanation linking variations in K, Cl, or Fe levels to PCA clustering results or to potential quality control parameters.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The authors proposed an interesting approach to using chemometric methods to identify herbal medicinal products.

In the Introduction, the authors concisely described the problem and the need for the presented research.

My comments:

Part 2.1 – "various manufacturers," meaning what? If the manufacturers of reagents are specified in the research methodology, the origin of the test samples is perhaps even more important. I understand that these are not standard substances. If the samples came from specific manufacturers, was this done with their consent? It isn't something you can buy in a herbal shop or pharmacy. Or are they ready-made pharmaceutical preparations?
Table 1 – Is a caption under the table necessary here? Descriptors d1, d2, and d3 are described in the table. However, there are no parameters r1, r2, r3, or R, which are also mentioned below the table.
Are there any Supplementary Materials where I can review the obtained descriptor values? It might be worthwhile to supplement the work with them.
Figure 3 – It would be a good idea to increase the resolution of the figure, as it is difficult to read in this format.
Where are the descriptors from Table 1 used? The PCA described later in the work uses wavenumber variables, and there is no information about the descriptors mentioned above. Please clearly indicate the need for these descriptors.
Data availability statement – the data are not presented in the article, as the authors wrote, but they are being analysed. Hence, my earlier question regarding their availability in the Supplementary Materials.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have provided satisfactory responses to the reviewers’ questions and comments. Accordingly, the figures, methodological descriptions, as well as the introduction and conclusions, have been improved and made more comprehensible. Furthermore, the discussion presented in Section 4 further clarifies the essence of the work in light of the results.

Reviewer 3 Report

Comments and Suggestions for Authors

All comments have been addressed. The revised version is improved and sounds more scientifically accurate.