-->PONE-D-25-23974-->-->Chrysoprase color grading with machine learning: a systematic approach-->-->PLOS One

Dear Dr. guo,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please revise the manuscript in response to the attached reviews. Details can be found below.-->-->

Please submit your revised manuscript by Mar 22 2026 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:-->

• A letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Agnieszka Konys, Ph.D.

Academic Editor

PLOS One

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please include a complete copy of PLOS’ questionnaire on inclusivity in global research in your revised manuscript. Our policy for research in this area aims to improve transparency in the reporting of research performed outside of researchers’ own country or community. The policy applies to researchers who have travelled to a different country to conduct research, research with Indigenous populations or their lands, and research on cultural artefacts. The questionnaire can also be requested at the journal’s discretion for any other submissions, even if these conditions are not met.  Please find more information on the policy and a link to download a blank copy of the questionnaire here: https://journals.plos.org/plosone/s/best-practices-in-research-reporting. Please upload a completed version of your questionnaire as Supporting Information when you resubmit your manuscript.

3. Please note that PLOS ONE has specific guidelines on code sharing for submissions in which author-generated code underpins the findings in the manuscript. In these cases, we expect all author-generated code to be made available without restrictions upon publication of the work. Please review our guidelines at https://journals.plos.org/plosone/s/materials-and-software-sharing#loc-sharing-code and ensure that your code is shared in a way that follows best practice and facilitates reproducibility and reuse.

4. We are unable to open your Supporting Information file [S2_File.mlapp]. Please kindly revise as necessary and re-upload.

5. If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

-->Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. -->

Reviewer #1: No

Reviewer #2: Yes

Reviewer #3: Partly

**********

-->2. Has the statistical analysis been performed appropriately and rigorously? -->

Reviewer #1: No

Reviewer #2: Yes

Reviewer #3: No

**********

-->3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.-->

Reviewer #1: No

Reviewer #2: Yes

Reviewer #3: Yes

**********

-->4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.-->

Reviewer #1: No

Reviewer #2: Yes

Reviewer #3: Yes

**********

-->5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)-->

Reviewer #1: I have encountered some negative comments regarding your article. I do not think it is possible to publish your article in its current form. Please review the comments carefully.

I have encountered some negative comments regarding your article. I do not think it is possible to publish your article in its current form. Please review the comments carefully.

Reviewer #2: 1. Broaden Sample Representation for Enhanced Generalizability

The study relies on 51 chrysoprase samples exclusively sourced from Australia, which limits the model’s ability to account for color variations across global deposits. Chrysoprase from regions like Tanzania, Poland, or Kazakhstan often exhibits distinct color traits due to differing geological conditions and nickel content—factors that directly influence its market valuation. Expanding the sample set to include these international sources would ensure the grading system reflects real-world diversity. Additionally, focusing solely on cabochon-cut specimens with uniform color overlooks common natural variations like subtle color zoning or minor inclusions. Incorporating such samples would test the model’s robustness in practical, non-ideal scenarios, making it more reliable for industry use.

2. Strengthen Alignment Between Model Output and Industry Practice

While the logistic regression model delivers impressive technical performance, the manuscript does not clearly connect its color grading outcomes to standard gemological workflows. Gemologists and appraisers rely on intuitive, perceptually meaningful criteria—not just numerical Lab* values—to assess color quality. For example, explaining how the “Fancy Intense,” “Fancy,” and “Fancy Deep” categories map to existing trade terminology or price differentials would bridge this gap. Supplementing with a direct comparison between the model’s results and evaluations from certified gemologists would also validate its relevance. Without this real-world context, the tool risks being seen as academically sound but impractical for day-to-day gem grading.

3. Address Practical Accessibility of the Grading Tool

The proposed color grading app requires input of Lab* values measured with an X-Rite SP62 spectrophotometer—a specialized piece of equipment most jewelers or small-scale traders do not own. This limits the tool’s widespread adoption. To enhance usability, the authors should test the model’s performance with data from more accessible devices, such as portable colorimeters or calibrated smartphone cameras. Providing guidance on how to obtain accurate color data without high-end equipment would make the tool actionable for a broader audience. Additionally, including a user-friendly interface demonstration or step-by-step guide for non-technical users would lower barriers to adoption, ensuring the research translates into tangible industry value.

4. Relevant References to Strengthen Methodological Context

To better contextualize the study’s core contribution—applying a systematic, data-driven classification framework to solve a longstanding industry challenge of subjectivity—it is recommended to cite the following two articles in the discussion section. These references align with the study’s emphasis on rigorous, cross-disciplinary methodology and standardization:

A comprehensive review on targeting diverse immune cells for anticancer therapy: Beyond immune checkpoint inhibitors（Crit Rev Oncol Hematol. 2025）This review advocates for moving beyond single-target approaches to develop multi-faceted, standardized strategies for complex biological classification—an idea directly paralleled in this study. The authors of the anticancer review highlight that relying on isolated markers (analogous to traditional subjective gem color assessment) leads to inconsistent outcomes, while integrating multiple data streams and systematic validation (like the combination of K-means clustering, Fisher discriminant analysis, and Bayesian optimization here) enhances reliability. Citing this work reinforces the broader significance of the current study’s methodology: just as the anticancer field benefits from standardized, multi-dimensional classification, gemology gains from objective, data-integrated grading that reduces human bias. This cross-disciplinary connection underscores the generalizability of systematic classification frameworks across fields with subjective evaluation challenges.

The role of cisplatin in modulating the tumor immune microenvironment and its combination therapy strategies: a new approach to enhance anti-tumor efficacy（Ann Med. 2025）This article focuses on optimizing therapeutic outcomes through strategic method combination and validation—an objective shared by the current chrysoprase grading study. The anticancer research emphasizes that combining complementary tools (e.g., drug therapy with immune modulation) addresses limitations of single-method approaches, much like how this study integrates colorimetric data, unsupervised clustering, and supervised machine learning to overcome the flaws of manual color grading. Additionally, the article stresses the importance of translating technical advancements into practical, standardized tools—mirroring this study’s development of a publicly accessible grading app. Citing this reference strengthens the discussion of the current work’s impact: both studies demonstrate how rigorous method integration and practical tool development can transform fields reliant on inconsistent, subjective assessments, making the gemology framework more relatable to broader scientific discourse on standardization.

Reviewer #3: 1.The training set is composed entirely of 676 synthetic GemDialogue-derived points while the test set is only the 51 real chrysoprase samples. This design risks domain mismatch (synthetic vs real) and over-optimistic generalization claims. The perfect classification of the 51 samples by the final LR model is surprising and may reflect leakage, overly narrow class boundaries, or lack of realistic variance in the training set. Please (a) clarify the rationale for treating all real samples as a pure test set, and (b) repeat model evaluation using cross-validation schemes that incorporate the real samples into training folds (for example: stratified k-fold where both synthetic and real samples are mixed during fold assignment, or leave-one-out on real samples). Report results of these additional evaluations and discuss differences.

2.procedure to create 2,730 overlay colors and then filter to 676 representative points requires more quantitative justification. Describe precisely (and provide code snippets or parameters) how the filtering used the L*, C*, h° bounds of the 51 samples. Discuss whether GemDialogue overlays realistically simulate cabochon optics (translucency, surface reflection, light scattering). If possible, validate the synthetic set by comparing distributions (e.g., kernel density or convex hull) of L*a*b* values between synthetic and measured samples. If synthetic data do not adequately cover real-world variability, consider augmenting the training set with measured variations (different orientations, multiple measurement locations per cabochon, different polish conditions, inclusion of proximate non-ideal samples).

3.n = 51 real samples is small for definitive claims. Provide confidence intervals or statistical tests on classifier performance (e.g., bootstrap on the 51 samples, or uncertainty estimates from repeated resampling). Report per-class sample counts for the real samples and indicate whether any class is under-represented; if so, show class-wise precision/recall on the test set. The manuscript currently reports only overall macro F1 and an assertion that all 51 were correctly assigned; show the confusion matrix and per-class metrics for the test set and provide explanation if classes have very small n.

4.The number of clusters (three) was chosen using silhouette applied to only the 51 real samples; then k-means with k=3 was applied to the combined dataset to label synthetic points. This approach is reasonable but may bias labeling toward the real-sample distribution. Make explicit the assumptions and potential limitations of this choice. Also, adopting GIA “Fancy” nomenclature (developed for diamonds) for chrysoprase requires caution — justify this transfer more explicitly (e.g., perceptual studies or prior literature applying similar labels to coloured gemstones).

5.Provide exact measurement geometry (you state SCI and D65, 2° observer) but clarify: measurement aperture size, number and location of measurement points on each cabochon (top only or multiple locations?), orientation control, whether measurements were taken through a mounting (e.g., on white card) or handheld, and repeatability (report standard deviation of triplicate measurements). Also indicate whether measurements were made on faceted or cabochon surfaces (you state cabochons but shapes differ); polishing and curvature affect measured color — discuss.

6.Hyperparameters table is useful, but please provide (in methods or supplementary) the optimization ranges used for Bayesian optimization, the Bayesian optimizer settings (acquisition function, iterations), and the random seed(s). Confirm whether reported results are deterministic across seeds. Provide code or a README in the GitHub repo explaining how to reproduce training and evaluation exactly (including MATLAB versions and toolboxes used).

7.The manuscript shows validation confusion matrices (via CV) but does not present the confusion matrix for the independent 51-sample test set other than stating perfect assignment. Include: (a) full confusion matrix for the 51 samples, (b) per-class precision/recall/F1, (c) visualization of decision boundaries in L*a*b* (or PCA) with real samples highlighted, and (d) calibration plots or probability outputs (for logistic regression) to assess classifier confidence.

**********

-->6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.-->

Reviewer #1: No

Reviewer #2: No

Reviewer #3: Yes: Dr. S M Rashidul Hasan

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

To ensure your figures meet our technical requirements, please review our figure guidelines: https://journals.plos.org/plosone/s/figures

You may also use PLOS’s free figure tool, NAAS, to help you prepare publication quality figures: https://journals.plos.org/plosone/s/figures#loc-tools-for-figure-preparation.

NAAS will assess whether your figures meet our technical requirements by comparing each figure against our figure specifications.

Attachments

Attachment

Submitted filename: REVIEWER.docx

Chrysoprase color grading with machine learning: a systematic approach

PONE-D-25-23974R1

Dear Dr. Cheung,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice will be generated when your article is formally accepted. Please note, if your institution has a publishing partnership with PLOS and your article meets the relevant criteria, all or part of your publication costs will be covered. Please make sure your user information is up-to-date by logging into Editorial Manager at Editorial Manager® and clicking the ‘Update My Information' link at the top of the page. For questions related to billing, please contact billing support.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Agnieszka Konys, Ph.D.

Academic Editor

PLOS One

Additional Editor Comments (optional):

Reviewers' comments: