You are currently viewing a new version of our website. To view the old version click .
MDPIMDPI
Search all articles
by
  • Xinli Li1,*,
  • Mingyang Liu2,* and
  • Jiaqi Sun2
  • et al.

Reviewer 1: Anonymous Reviewer 2: Anonymous

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The article entitled “An Intelligent Strategy for Colony De-replication Using Raman Spectroscopy and Hybrid Clustering” refers to the development of a method for the quickest selection of strains using Raman Spectroscopy and the application of clustering algorithms. The article is novel and the importance for the scientific community could be high; however, serious scientific flaws in material and methods and results sections were observed that should be corrected, so that the article could be published. More specifically:

  1. Introduction section should be strengthened with more information about the colony selection processes and the aim of the study. The last paragraph of the introduction summarizes the method and the results of the manuscript and should be removed.
  2. Figure 1 quality should be improved.
  3. Section 2.2: Information about the Raman spectra acquisition is missing. Which spectrometer model is used? With which kind and model of detector is the spectrometer equipped? What substrate was used for recording the Raman spectra? Also, all other parameters used, which are important for the reproducibility of the experiments should be mentioned.
  4. Lines 127-128: The website URL should be in the references section.
  5. Lines 139-143: They should be transferred to Materials and Methods section.
  6. Figure 3: The characteristic peaks of Lactococcus lactis should be mentioned and assigned to the respective modes.
  7. Line 148: Lactococcus lactis should be in Italics.
  8. The Raman spectra of the 7 different strains that were used to produce the balanced and imbalance pure-colony datasets should be provided, the spectral differences among the strains should be highlighted and the specificity of the method regarding the different strains should be established.
  9. Figures 4 and 5: The colour used for each cluster should match the colour of the respective strain, for clarity purposes.
  10. Sections 3.2 and 3.3: Method validation includes accuracy, precision, sensitivity, specificity and robustness of the method. However, in the section only the results of k-means and HCA clustering are presented.
  11. Figure 6: It is not clear if the results are from plate A, plate B or the combination of the two plates with 113 colonies. Also, it should be proven that the clusters shown in the figure are indeed different strains.
  12. The final results about the improvement of the picking efficiency are not explained if they are average results and seem to be somehow misleading. For example, for plate B only 1 more colony was identified compared to image-clustering method and less than double colonies by using the random method. These results are not as promising as the results presented in the manuscript. A reformation of the presentation of the final results should be performed.
  13. The discussion section is very weak, as there is no connection to previously conducted studies from literature.
  14. The conclusion section should not repeat the results (which are already mentioned in line 196 and line 53) and it should refer to possible applications of the developed method.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The paper uses Raman spectroscopy combined with a hybrid clustering strategy to improve colony selection efficiency in fermentation microbiology. By combining k-means with hierarchical clustering, the method significantly reduces redundancy while maintaining species coverage. Validation on pure colonies and complex mixed plates shows that picking only 12–26% of colonies achieves 80–100% species coverage, outperforming random and image-based approaches. However, there are issues that must be addressed:

 

  1. Pure-colony provides controlled testbeds. However, external validation on independent datasets is missing.
  2. Include confidence intervals or statistical significance for performance metrics.
  3. Report hyperparameter tuning details for clustering.
  4. They report on clustering metrics, but no formal statistical tests (e.g., confidence intervals) are reported.
  5. Include a comparison against additional clustering methods (e.g., PCA+ clustering, UMAP + clustering).
  6. Limitations such as sample size, lack of external validation, and potential variability in Raman acquisition should be discussed.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The article entitled “An Intelligent Strategy for Colony De-replication Using Raman Spectroscopy and Hybrid Clustering” referred to the development of a method for the quickest selection of strains using Raman Spectroscopy and the application of clustering algorithms was improved significantly after the first round of revision. However, there are still some issues to be addressed before publication:

  1. Figure 3: The characteristic peaks of Lactococcus lactis were successfully assigned in the manuscript. However, peaks E and F are not placed in the correct Raman shift in the Figure. Please amend the Figure accordingly.
  2. Figure 4: The colours of the Raman spectra are not matching the colours of the strains in Figure 4a or the clusters in Figure 4c and Figure 4d. Please modify the colours accordingly. Each Raman spectrum should have the same colour with the strain (Fig. 4a) and with the predicted clusters (Fig. 4c and 4d).
  3. Figure 5. Please harmonize also the colours as already described in the previous comment for Figure 4.
  4. Section 3.2 is entitled as “Method validation on pure colonies”, while Section 3.3 is entitled as “Validation on complex mixed plates”. However, as already emphasized in the comments of the first round, the method is not validated, which require determination of accuracy, precision, specificity and sensitivity and may mislead the reader. The titles are better to be replaced with the titles: “Application of the method on pure colonies” and “Application of the method on complex mixed plates”.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The authors correctly addressed all of my comments. 

Author Response

None

Round 3

Reviewer 1 Report

Comments and Suggestions for Authors

Two out of the four comments were addressed in the previous round. However, the inconsistency in the colours of the Figures 4a and 4b, as well as in the colours of the Figures 5a and 5b still exist, although the authors refer that they have been recoloured. For example, in Figure 4a the L1 strain is coloured orange, while the same strain in Figure 4b (Raman spectrum) is coloured grey-green. The same inconsistency could be observed also among all other strains except for the middle one L4. Please check carefully if the Raman spectra presented in Figures 4b and 5b refer to the labeled strains and the groups in Figures 4a and 5a refer to the labeled strains as well; otherwise please amend the colours to succeed consistency between the Figures 4a-4b and Figures 5a-5b.

Author Response

Comments: Two out of the four comments were addressed in the previous round. However, the inconsistency in the colours of the Figures 4a and 4b, as well as in the colours of the Figures 5a and 5b still exist, although the authors refer that they have been recoloured. For example, in Figure 4a the L1 strain is coloured orange, while the same strain in Figure 4b (Raman spectrum) is coloured grey-green. The same inconsistency could be observed also among all other strains except for the middle one L4. Please check carefully if the Raman spectra presented in Figures 4b and 5b refer to the labeled strains and the groups in Figures 4a and 5a refer to the labeled strains as well; otherwise please amend the colours to succeed consistency between the Figures 4a-4b and Figures 5a-5b.

Response: We sincerely apologize for the oversight. In Figures 4b and 5b, the order of strain labels L1–L7 was mistakenly reversed during the previous revision. The correct order should be from bottom to top (L1–L7) rather than top to bottom. This labeling has been corrected in the latest revised version. We greatly appreciate the reviewer’s careful observation.