All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
All remaining concerns of the reviewer were addressed and revised manuscript is acceptable now.
[# PeerJ Staff Note - this decision was reviewed and approved by Paula Soares, a PeerJ Section Editor covering this Section #]
No comment.
I thank the authors for carefully addressing all of my concerns which were not clear after the first round of revision.
No comment.
No comment.
Please address concerns of reviewer #2 and make the necessary amendments to the manuscript.
All my concerns have been addressed.
All my concerns have been addressed.
All my concerns have been addressed.
No comment
I thank the authors for clarifying some issues raised during the initial revision.
However, from methodological point of view I do not see a valid reason why 3 normal and tumor samples should be analysed separately and then finding an overlap of DEGs. Standard procedure is to analyse all samples together (thus you have 3 biological replicates in your data) and for RNA-seq data having 3 samples is a minimum requirement. You can perform your analysis both with DESeq2 or edgeR, and then filter your significant DEGs if having adjusted p-value less then 0.05 and/or include a potential fold change cut-off. If your current results from individual analysis (which I do not recommend) are robust, the same genes should be obtained as top hits also from the standard RNA-seq differential expression analysis pipeline. Based on your PCA plot, your samples are clearly separating on PC1 meaning that you have a clear difference between the normal and tumor samples, which in turn definitely goes in favor to use the standard analysis pipeline. I strongly recommend to do the standard pipeline and to compare it with your current unstandardized separate DE analysis pipeline.
Regarding the confusion on using TCGA normal adjacent tissue vs GTEx healthy tissue as a control, please be aware that all 3 TCGA normal adjacent tissue samples are in fact paired with cancer samples (meaning it is the same experimental design as yours) - samples ending in _01 represent cancer samples, whereas those with _11 represent normal adjacent samples. All 3 are present in TCGA as both cancer and normal adjacent tissue samples. Because of it, you can perform the same analysis with TCGA cancer and normal adjacent tissue samples as with your own data for validation purposes.
No comment
Please address the concerns of all reviewers and revise the manuscript accordingly.
**PeerJ Staff Note:** Please ensure that all review, editorial, and staff comments are addressed in a response letter and that any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate.
This manuscript was developed in a professional manner with clear and accurate writing. Literature references were well-selected to support the idea of this manuscript. Figures and tables were clear and well ploted. My major concern is:
1. Use a diagram to summarize the distinguished 10 DEG genes, especially for five of which never reported before;
2. Replot figure 6 to be more approachable.
3. Re-write the discussion portion. Finding of the 5 unreported genes is very interesting. However, the current discussion lacks of insights and feels more like a "summary".What can be the potential clinical application of this finding?
Experimental design is well-organized. Though suffering from small sample size, it genereates results with great interest.
The impact of this research is limited. It does generate new knowledge but lack good analysis of it.
In this manuscript, Xu et al reported 10 differentially expressed genes (DEGs) during human cervical tumorgenesis. They collected specimens from 3 patients after radical hysterectomy. Using next generation sequencing, they compared DEG and found 5 increased- and 42 decreased expression of genes. By performing several kinds of functional enrichment analysis, they further revealed that changes in extracellular matrix organization and cell cycle checkpoints were the two most essential processes and responsible for cervical tumorgenesis. More importantly, these DEGs were also found in the public cancer databases, suggesting their importance in cancer therapy. Finally, they confirmed analytic result by RT-qPCR.
Overall, this is a well conducted study, which characterizes a set of predicting genes to determine the process of cervical cancer. The paper is well written, and the figures are clear.
no comment
no comment
For the Ackownledgements part, a clear blank was shown between "generous" and "domain". Please correct it.
In the manuscript titled "Identification of 10 Differentially Expressed Genes in Tumorigenesis of Cervical Cancer via Next Generation Sequencing," Jia Xu et al. performed bulk RNA-seq on tumor and adjacent normal tissues from three patients with cervical cancer to identify the top DEGs. The story is interesting and presented clearly. The experimental design seems reasonable, the analysis appears correct, and the data supports the claims. Some specific comments are as follows:
1) I suggest reorganizing the introduction section to include more about gene expression and tumorigenesis aspects of cervical cancer as background.
2) Please rephrase the introduction and discussion to ensure logical transitions between sentences. Currently, sentences stand well alone but do not connect to form coherent paragraphs. For example, lines 70-75, the use of 'However, therefore' does not reflect the logical relationship between the sentences.
3) In lines 65-66, you probably will need a reference for this claim.
4) In lines 78-80, I think both Sanger sequencing and NGS obtain sequence data through the synthesis of complementary chains.
5) Lines 277-281 seem more appropriate for the introduction rather than the discussion.
6) In lines 292-345, detailed information on each of the 10 genes may be unnecessary in the discussion section. It might be more valuable to discuss how these genes provide direction and insights for future studies on tumorigenesis.
7) Discussing the ROC analysis and its application as clinical biomarkers or diagnostics would be beneficial. Currently, the ROC is not mentioned at all in your discussion.
8) It would also be valuable to discuss the survival curves.
9) Discuss why some of the 10 genes may show opposite results in different cases, such as MASP1 and ACKR1 in case 9.
10) More detailed figure legends are needed, especially for Figure 6 (lines 673-675). What is used as the error bar? What do the stars indicate?
1) In lines 105-107, it is unclear how the three technical replicates were achieved. Did you sequence the same RNA three times, or did you extract RNA from the same tissue three times?
2) In lines 117-121, more information is needed about the sequencing method. Did you perform 150 bp paired-end sequencing? Was there polyA-enrichment or rRNA depletion? What was the sequencing depth? Any quality control for the reads? How were the raw reads processed and mapped to which reference genome? How was the sequencing coverage? Such information is crucial.
3) In line 181, GAPDH is used as the internal normalizer for RT-qPCR; however, based on my experience and reading, the expression of GAPDH itself may vary between tumor and non-tumor tissues. Have you checked the expression changes of GAPDH in your RNA-seq data? If it changes, I would recommend using another normalizer for RT-qPCR.
4) In lines 91-102, please indicate whether there are any inclusion or exclusion criteria for the cases included in the study.
5) In line 135, I am confused about the 'All the DEGs' mentioned. Did you use the union of DEGs identified in the three patients to do the functional enrichment analysis?
1) In lines 247-253, only the AUC is reported for the ROC analysis. However, as far as I know, the ROC also provides other important information, such as insight into specificity. Please indicate whether it can be applied here.
2) How comparable is your sequencing data to the published data in GTEx and TCGA? In terms of the number of genes, is there a significant overlap between the up/down-regulated genes identified by your study and their data? If not, can you explain? Will the top DEGs identified by them also be identified by your data?
3) I would recommend combining the results from different cases in Figure 6 to obtain a population-level view of the changes in the DEGs. A boxplot or violin plot would be suitable for presenting this data.
This article evaluates the potential biomarkers of cervical cancer based on RNA-seq data. Major strength of this article is extensive data validation, however there are several points that can be improved.
No need to have introductory section on NGS and Sanger sequencing, as NGS is now a standard method. It would be beneficial to add more background on what is known with regards to some traditional biomarkers for cervical cancer and why have you decided to choose RNA-seq as your method of choice to answer your primary question.
Please describe how paired adjacent normal tissue was collected and checked that they are normal.
Please add more details regarding this statement: "Three technical replicates were completed simultaneously." as there is nothing stated about this later on.
Why have you decided to do RNA-seq on only 3 samples when you have collected in total 18 patients with cervical cancer?
Please provide more details why you have chose to analyse each patient separately and then as an "integrated" analysis of all three cancer vs three normal adjacent tissues samples (vs. just making one analysis with all three patients together)? Also, please describe why choosing an arbitrary threshold of top 500 genes, doing an intersection and then choosing only top 10 genes to make validation using publicly available data? Why not using the complete list of DEGs to do so? And when selecting your genes with these thresholds did you make sure that you are only comparing them if they have the same direction of effect? Also, why did you select only downregulated genes as your final selection?
Why did you use both edgeR and DESeq2?
Did you include age as covariate into your model? Since it is known that gene expression changes as we age.
When using TCGA data, you should have downloaded the count data and perform DE analysis yourself and then compare the results with your own data. That way you could have TCGA normal adjacent tissue as your control samples which would directly correspond to your experimental design. It is known that GTEX healthy samples are different control samples than TCGA normal adjacent tissue.
In addition, since your patients are in stage I and II, did you consider performing analysis with TCGA data only with stages I and II to correspond to your cohort?
Have you considered checking GEO for other cohorts with cervical cancer and normal samples data to include in your validation step?
Was functional enrichment analysis performed on all DEGs or only on a specific subset of DEGs?
Unclear why some part of the analysis was performed in Graphpad when everything else was done in R. Would be better that it is uniformed (regarding visualizations).
Taking all your results together, it should make sense to try out some machine learning classification model (using both publicly available data and your own) to test how good your final selection of genes (potential biomarkers) are in discriminating between the cancer and normal tissue sample.
The manuscript was written in clear and professional English. Sufficient literature references and background were provided. For the figures in this manuscript, there are some space to improve:
1. In Figure 3, please indicate in figure caption: What’s the meaning of error bar, e.g. SD or SE or IQR? And what’s the meaning of the middle black bar, e.g. mean or median?
2. In Figure 6, please clarify in figure caption: What’s the meaning of error bar, e.g. SD or SE? Also, please indicate the corresponding p level of various number of “*”.
This study identifies promising DEGs in cervical cancer with potential diagnostic and prognostic value, including novel candidates like C1QTNF7, HSPB6, GSTM5, IGFBP6, and F10.
However, I have several concerns about the statistical analysis parts in this manuscript.
1. Both edgeR and DEseq2 support two-group and multi-group comparison. Why did the authors use edgeR for two group comparison and DEseq2 for multi-group comparison?
2. I recommend to indicate what statistical methods were used in the caption of each figure.
The results and conclusions of this manuscript were well stated and linked to research question. More clarifications related to statistical analysis need to be added to make the results and conclusions promising.
NA
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.