Review History


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.

View examples of open peer review.

Summary

  • The initial submission of this article was received on January 17th, 2023 and was peer-reviewed by 2 reviewers and the Academic Editor.
  • The Academic Editor made their initial decision on February 15th, 2023.
  • The first revision was submitted on March 27th, 2023 and was reviewed by 2 reviewers and the Academic Editor.
  • A further revision was submitted on April 4th, 2023 and was reviewed by the Academic Editor.
  • The article was Accepted by the Academic Editor on April 5th, 2023.

Version 0.3 (accepted)

· Apr 5, 2023 · Academic Editor

Accept

Well done on addressing all Reviewers' comments.

[# PeerJ Staff Note - this decision was reviewed and approved by Jafri Abdullah, a PeerJ Section Editor covering this Section #]

Version 0.2

· Mar 30, 2023 · Academic Editor

Minor Revisions

Please act on the pending minor concerns of Reviewer 1.

·

Basic reporting

Line 76 – remove ‘in’
Line 142 – ‘filtrate’ to ‘filter’
Line 176 – ‘and’ instead of comma
Line 184 – ‘analysis’ to ‘analyze’
Line 215 – comma after GAPDH
Line 222 – Remove ‘that’
Line 283 – ‘7’ to ‘seven’
Lines 317-323: Place ‘the’ in front of ‘most’
Line 324 – ‘rest’ to ‘other’
Line 330 – I think a semicolon would make more sense instead of a comma.
Line 341– ‘prominent’ to ‘prominently’
Line 378 – ‘lesion’ to ‘lesions’
Line 379 – ‘DEG’ to ‘DEGs’
Line 466 – italicize ‘in vitro’
Line 511-512 – This sentence is repeated from the first sentence of the previous paragraph. Combine these two paragraphs into one.
Line 523 – ‘. Taking..’

Experimental design

/

Validity of the findings

/

Additional comments

Well done to the authors on addressing the comments and suggestions made by the reviewers. The manuscript reads well and explains all methods carried out and results obtained to a very satisfactory level.

·

Basic reporting

Overall, the manuscript has been greatly improved. Both Engliush and punctuation have been improved as well.

Experimental design

No Comment

Validity of the findings

-All underlying data have been provided; they are robust, statistically sound, & controlled.
-Conclusions are well stated, linked to original research question & limited to supporting results

Additional comments

After following the guidance/steps I provided on the statistical analysis using Lasso, authors discovered 28 unique genes, thus increasing the discoveries. Authors have addressd all my concerns with regards to the statistical analysis and variable selection with Lasso. Previously, I did not comment on the English language and punctuations used. However, I find great improvement after reading the revised version of the manuscript. Overall, I am happy with this version. Well done!

Version 0.1 (original submission)

· Feb 15, 2023 · Academic Editor

Major Revisions

Concerns remain about the selection of the criteria for sample selection and the way results are presented. Several points highlighted by the Reviewers concerning the validity of findings also need to be addressed.

[# PeerJ Staff Note: Please ensure that all review and editorial comments are addressed in a response letter and any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate. #]

·

Basic reporting

The English used is overall very understandable, but I have identified some lines below that may be improved:

Line 21 – CNS in full first time mentioned
Line 25 – WGCNA in full first time mentioned
Line 29 – GSEA in full first time mentioned
Line 35 – DCA and ROC in full first time mentioned
Line 39 – ‘effectively predicted’
Line 48 – ‘exists’
Line 50 – Instead of ‘exists’ better: is present/can be observed
Lines 51-52 – Instead of ‘the deep’ better: severe/extensive
Lines 56-57 – Either remove ‘more susceptible’ OR change ‘predisposes’ to ‘makes’
Line 58 – ‘in’ to ‘on’
Line 61 – Should ‘while’ be ‘but’?
Line 72 – remove ‘in’
Line 82 – Place a comma after ‘lesions’
Line 101 – ‘top’ 20
Line 103 – ‘functions’ and DEGs’ and ‘analyses’
Line 106 – ‘are’ to ‘were’
Line 116 – ‘correlation’ to ‘correlated’
Line 133 – ‘is’ to ‘was’
Line 140 – ‘greater’
Line 154 – Remove ‘each’
Line 156 – ‘is’ to ‘was’; ‘To’ to ‘For’
Line 163 – Immunohistochemical ‘staining’
Line 164 – ‘analysis’ to ‘analyze’
Line 168 – ‘and’ before ‘CCL5’
Line 170 – ‘using’ instead of ‘by an’
Line 177 – ‘incubation’ instead of ‘incubated’
Line 194 – Add ‘and’ before ‘an internal’
Line 201 – Remove ‘that’
Line 203 – ‘depicts’ or ‘The top 20... are depicted in a heatmap’
Line 211- Why have you used ‘remarkably’ here?
Line 217 – ‘turquoise’

General comment for results section – It is more common to put a figure number in brackets, rather than starting a sentence with the figure number. E.g.: The DEGs were displayed in a volcano plot (Figure 1); instead of: Figure 1 shows the DEGs in a volcano plot.

Line 223 – ‘In total’
Line 224 – modules
Line 233 – ‘tissues’ or ‘tissue samples’
Line 233 – ‘seven’
Figure 3A legend – ‘shows’
Figure 3F legend – ns is no significance
Line 237 – ‘different’
Line 239 – Instead of ‘than’ use ‘relative to’
Lines 241-243 – This sentence needs to be rewritten and explained better. I cannot understand what is correlating with what from the way it is written.
Lines 245-252 – Indicate that for each gene you write the ‘most’ positively and negatively correlated cells. There are other cells are also positively or negatively correlated.
Figure 4B legend – ‘indicates’
Figure 4C – Words are a bit too small to understand. One must zoom in to the figures too much to see better, and then clarity is lost.
Line 256 – Figure 5B in brackets
Figure 5D legend – Specify that evaluation is being done for four genes.
Line 264 – The use of ‘remarkable’ is a bit over the top..
Line 266 – Same issue with ‘remarkably’..
Line 266 – Change ‘upon’ to ‘in’
Line 267 – ‘was demonstrated’ to ‘exhibited’
Line 268 – ‘significant enrichment in..’
Line 280 – ‘results’
Line 302 – ‘and eosinophils’
Line 305 – ‘A recent..’
Line 306 – The only mention in manuscript of SPMS, so write out in full.
Line 306 – Either ‘;the result.’ Or ‘..SPMS. The result’
Line 309 – ‘..neglected. With..’
Line 310 – ‘enriched in’
Line 311 – The first mention in manuscript of EAE, so write out in full.
Line 317 – ‘in vivo’ should be italicized
Line 319 – I think ‘Like’ may be easier to understand than ‘As’ here
Line 336 – ‘TLR9 is..’
Line 338 – ‘cytokines’
Line 357 – ‘expression’
Line 362 – ‘due to’ instead of ‘for’
Line 371 – ‘leading’
Line 377 – First time BBB is mentioned write out in full.
Line 380 – ‘PGDFRB is..’
Line 385 – ‘A previous study has shown..’
Line 388 – ‘Nevertheless’ here does not make much sense
Line 399 – ‘predicted’
Line 400 – ‘could’ better than ‘would’

Experimental design

Lines 92-92: How did you finally select the chosen twenty samples? Were they randomly chosen from the ones that met the appropriate criteria?

The suitability of GSE131282 (line 128) dataset is not explained when first mentioned in line 128. I only found out how many samples are found in the dataset later on in lines 232-233. This needs to be written in the methods section. Why was it selected?

Provide some examples of what the co-expression modules represent in WGCNA. Just using colours makes it a bit abstract.

In LASSO, how is the penalty parameter chosen and what is its significance? What do figures 3D and 3E show, and what follows from this analysis?

Brief explanation of what CIBERSORT is would be useful.

Validity of the findings

Line 237 – Not all 22 were significant, right? If not, clarify how many.

Line 245 – I only see correlation analysis for four genes. What about the other three genes? Why were they omitted from this and all further analyses?

Lines 367-368 – Have any other studies also shown these ‘aberrant levels’ due to drugs taken during treatment?

Line 373 – However, the results in your mouse model mimic what one would expect in a human MS patient, correct? This means elevated levels of CCL5 and decreased levels of PDGFRB and TLR9. Can you suggest another model that may be better than a mouse?

Line 381 – How does this tie with your findings?

Line 390 – Are you referring to GSE135511? Or GSE131282?

Line 393 – 42 healthy and 37 GM lesions tissue in GSE131282, correct? Why do you believe that this is a small sample size? GSE135511 had ten and ten samples (total 20), much fewer!

Line 394 – Why may they contain bias?

You identified DEGs and performed an enrichment analysis, but what were they used for? Discussed at all?

Why are eight genes shown in figure 3F and not all ten?

Fig 4C why only four genes not eight or ten?

More discussion required on what the nomogram shows and why it can be a useful tool. The identified genes predict under what circumstances? Is it when they are underexpressed? I did not quite get how the genes can be used for MS diagnosis. Any further tests required?

You produced flower plots (Fig. 7) but only briefly discussed some pathways in lines 382-383. Elaborate some more on the relevance of the pathways and inferences that can be made re MS.

Additional comments

I incorporated all my comments in the first three sections.

·

Basic reporting

No comment

Experimental design

No comment

Validity of the findings

-The manuscript is well written and well summarised. The intro, methods and discussion are well explained. I did not find any difficulties reading the manuscript. Well done! However, I have some major concerns with regards to the methods, but I belief these can be easily addressed by the authors.

Major comments

-Authors run PPI network using 116 screened genes and identified top ten genes by intersetcion (Figure 4A). What are the criteria for choosing only top 10 genes? This is very subjective, and suboptimal. Some genes will be left out of the left out of the gene signature. Therefore, would recommend running LASSO using LOOCV before doing a PPI analysis. Instead, do the PPI network analysis on the LASSO results and then build the final gene-signature using the PPI analysis results.
-LASSO should come before the PPI analysis. The idea is that you don’t want to be throwing out genes without any justification. Let LASSO select the candidate gene-signatures and let the PPI network analysis be used to construct the final model. In that way, you will eliminate the subjective bias in choosing top genes by intersection.

-As the Sample size is extremely sample, the study is underpowered. I will suggest applying the following steps.

1) Run LASSO on Immport&Innatedb genes (Figure 4A, Venn Diagram)
2) Run LASSO on WGCNA genes (Figure 4A, Venn Diagram)
3) RUN LASSO on the combined Immport&Innatedb and WGCNA genes (Figure 4A, Venn Diagram)
4) Intersect the results from steps 1, 2, and 3 to get final genes.
5) Do PPI analysis of the results from step 4.

Note that doing these steps doesn’t solve the issue of power, however, it will eliminate the possibility of chance findings. I therefore strongly recommend the authors do these steps.

-From figure 4D, the negative log-lambda path I noticed that the vertical line (1SD of the min Lambda) reveals 6 or 5 genes are selected and not 8 genes. Where did you get 8 genes from? This require further explanation. If you added 2 more genes not selected by LASSO, then this is a systematic bias and need to be accounted for. Note: the correct criteria for choosing genes selected by LASSO when using glmnet package is the 1SD rule. That is, the second vertical line when moving backward along the negative log-lambda path (From -8 to -2).

Additional comments

Minor comments

-Some figures are not clearly visible. Resolution is very poor. Please check these.

-Why did the authors choose to use only LASSO algorithms? There may be systematic biased due to used algorithm. There are other algorithms like ELASTIC NET, which is less strict than LASSO. Also, sure independent screening (SIS) from the ncvreg package is also available and could be explored further.

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.