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.
Well done. I have been working in East Asia for the last 7 weeks, which accords for the delay in deciding on your revised manuscript.
Now after a careful line-by-line comparison of the pen-marked original manuscript with your blue-marked revised manuscript plus the author's responses to earlier comments, it is clear that this is a much stronger report than when we started many months ago. I have read the new set of outside reviewer's comments but I have decided that a further revision is not required
Hopefully it is now clearer to potential readers and more useful for the forensic analysis community needing to use hair samples.
[# PeerJ Staff Note - this decision was reviewed, discussed and approved by Vladimir Uversky and Pedro Silva, a PeerJ Section Editor covering this Section #]
Literature reference inaccurately interpreted.
The authors use an approach (2D gel) that prevents analysis of most of the protein. What remains for comparison is only a tiny fraction of the total.
The comparison (small differences between large numbers) depends on accurate normalization. Linearity of the "volume"measurements is not addressed. It is hard to know whether the measurement is saturated in regions of the gel with the highest protein concentration.
The original goal and the results, if believable, are interesting. However, the shortcomings undermine the conclusion.
This manuscript is considerably improved, especially with respect to correction of some misinformation in the previous version. Several points need further attention, however.
1] An important advantage of the present approach is the claim of higher protein yield using the alkaline extraction method. Lines 70-71 refer to extraction of a “substantially higher percentage of hair shaft proteins within only 2 hrs”. Lines 207-208 refer to yields “generally higher compared to the methods of Lee et al (2006)….” Line 267 refers to their “newly developed protocol that enhanced the extraction yield of proteins from the human hair shaft….” Line 210 cites “an average extraction yield of 63.56 +/- 6.59”, presumably % of the starting amount, but does not explain how this figure was obtained. How much protein the method of Lee et al (2006) would extract after 2 hr was not reported, but close reading of the methods section of that paper reveals that “the final insoluble material comprised 13.3 +/- 3.9% of the total protein”, which means about 86% was soluble. Thus the presumed higher yield of the alkaline SDS method is not substantiated by the cited literature. The speed of the method could be an advantage but is canceled by coupling it with an unnecessary 2D gel separation.
2] The authors still refer to the 2D gels as giving “highly resolved profiles” (line 36). The resolution may well be higher than before, but the protein available to be analyzed is still only a tiny fraction of the total. “A larger scale analysis … using antibodies” could indeed “increase the robustness of the results” (line 323-325), but this ignores the likelihood that a direct mass spectrometric analysis (no 2D gels) is going to be more accurate than western blotting.
3] Figure 1 has two problems. First, it strains credulity that spot 7 can be analyzed with the accuracy claimed. Poorly resolved (smeary) 2D gel spots often contain more than one protein. Second, the source of the protein in panel A is not given.
4] Although the methods are greatly improved, with considerable detail now included, verification of linearity of the volume measurements is not addressed. This is critical for the normalization of spot amounts for comparison of the different samples and would be inaccurate if, for example, the measurement saturates in regions of the highest protein content.
5. The rationale for limiting the crime scenes to “old” ones (line 25) is unclear.
There is both good news to report to you concerning your manuscript “Potential use of human hair shaft keratin peptide signatures to distinguish gender and ethnicity“ by Nurdiena Mohamed Nasir, Jumriah Hiji, Jaime Jacqueline Jayapalan and Onn Haji Hashim; and a strong need for more effort in bringing this to a form suitable for publication in PeerJ. The current problems seem to your outside peer reviewers overwhelming; this knowledgeable editor is just not sure.
We obtained 2 excellent peer reviews, with comments below for your consideration and response. Both outside reviewers are knowledgeable (as you can see from their in depth technical comments) and encouraging that your method (if it can be made clear to readers and developed further in directions stated) might indeed be useful for human identification, often in forensic situations. I, as a somewhat more outside Editor, who has however followed proteomics from its early stages with 2D gel electrophoresis initially with total E. coli cellular proteins maybe 40 years ago and newer methods with mass spec identification of proteins following 1D column chromatography (my laboratory confirmed one of our DNA sequence-derived amino acid sequences by total analysis of the tryptic peptides obtained from purified protein, maybe 20 years ago). However, I have always worked on DNA and proteins from bacterial and never from humans.
In addition, I read and marked the manuscript line by line in pen. Perhaps my marked manuscript attached here as a PDF file is the most useful additional help to the authors. Below, I have added a 3rd Editor’s set of informed comments. I will not repeat points made by the outside reviewers, since basically we all agree. 2D gels is a long-gone older technique now universally replaced by liquid chromatography directly coupled to mass spectroscopy. The reviewer is correct. Nevertheless, we cannot change that today, so my overall goal is to clarify and simply your presentation so that you and other groups can move forward.
How can we do that? Please read and address the criticisms and suggestions of your 2 outside peer reviewers.
Again, if I understand correctly, the reviewers are correct that 2D gel electrophoresis for such work is basically gone and has no future.
Then I will need to decide myself whether the changes made are sufficient to move forward to publication or not. Reviewer #1 thinks that is not possible and more modern work, beyond 2D gels and with data demonstrating that the initial protein extraction from the hair was quantitative – e.g. maybe >90%. Reviewer #2 is more encouraging, explaining what the problems are and asking to see how you respond. And I am the most encouraging of your 3 readers, with still more comments and an attached PDF file of the manuscript pages that I marked in pen when reading. It might be possible to move forward without further experimental work, but I cannot promise that considering the comments today. I hesitate to send this back to the 2 excellent outside reviewers since they have worked hard and all 3 of us see the same basic problems. There are no important differences in what we see, but only how much we can encourage you and move this effort forward.
Yours sincerely,
Simon Silver
Editor, PeerJ
The background contains incorrect statements and misinterpretations of the literature. These need correction. See details below.
The 2D gel method is inferior to contemporary methods. It dramatically limits the scope of the analysis. Several aspects of the methods require more detail. See details below.
If the authors want the reader to believe the conclusions, a more streamlined and quantitative approach is needed. The alkaline SDS extraction eliminates too high a fraction of the total protein and 2D gels require too much manipulation of the samples as well as making most of the sample unusable.
This manuscript has a laudable goal, and the results, if believable, would be a considerable step forward. However, the manuscript has serious shortcomings that must be addressed and is not convincing. The main problem is the use of 2D gels. This is an outmoded technology and is totally unjustified in this instance. The authors are limited to comparisons of only the small number of peptides that are separated cleanly, thereby unable to use the vast majority of the protein (not quantitated but probably >>90%). Enthusiasm and confidence are substantially eroded by incorrect statements.
[1] The Background and Discussion claim erroneously that the hair shaft is useless for forensic investigation if it does not contain nuclear DNA. The authors appear unaware of recent reports that protein profiling of the hair shaft shows promise in this direction, and the genetically variant peptides in the proteome are likely to provide definitive identification (Proteomics 17(13-14), 2017. doi: 10.1002/pmic.201600462). Moreover, the Introduction is mistaken in that next generation sequencing of mtDNA is not limited to STRs and can be quite discriminating (Genes 9(1), 2018. pii: E49. doi: 10.3390/genes9010049).
[2] The authors cite Yamauchi et al (2008) for their assertion that the hair shaft is 80% keratin. Actually, this reference does not do a determination but reports the work of others that the cortex is 80% of the hair shaft.
[3] The authors use their recently published (2016) method of alkaline extraction of hair, claiming to get higher yields. No evidence is given in the present manuscript for higher yields, and the reference reports a yield of <50%. Other methods, subjecting the entire hair shaft protein to digestion, have yields of 90%.
[4] In the Methods, the authors say they recorded phenotypic characteristics and excluded chemically-treated hair. The reader needs to know how they determined which hair was chemically treated. Visual inspection alone is not satisfactory. This is critical because a high proportion of females bleach, perm or color their hair. Not excluding some of those samples could seriously bias the results they report and could account for their reported male/female and subpopulation differences.
[5] How a supernatant was obtained by pulverizing undissolved hair shaft fractions (line 113) is not clear.
[6] How the focused strips were “equilibrated using DTT and iodoacetamide” (line 30) is not clear, nor how this did not lead to collapse of the pH gradient. Moreover, if the purpose is to alkylate the cysteines, the subsequent reduction and alkylation of gel plugs (line 162-163) is redundant. If it is not redundant, perhaps due to incomplete alkylation earlier (line 30), then this could be a source of variability in IEF and thus in the spot densities that are analyzed and thereby negate the results.
[7] The Discussion states that SDS and urea “are known to cause deamidation and disruption of disulfide bridges”, citing Adav et al, 2018. Actually, Adav et al referred to deamidation from degradative/ageing processes, not denaturant treatment, and disrupting disulfide bonds is assisted by denaturants but requires a reducing agent such as DTT.
to distinguish gender and ethnicity” reports to show that there are proteins in hair that could be biomarkers associated with gender and bioancestry (based primarily on abundance). The work is preliminary and may have some value. But a bit more data are needed.
The authors have shown there is an abundance difference for some identified proteins. The abundance may be real. To support their findings they present x fold differences observed on the 2d gels. There needs to be some discussion about the robustness of the analysis. There are a lot of proteins that overlap and some areas on the gel are quite dense with protein. These areas appear that they may no longer be linear in signal. If so, then it seems suspect about rendering any abundance values. Can the authors provide support for the use such data and/or discuss the limitations.
The authors use the significant difference in abundance to infer gender and bioancestry. Yet there are probabilities, likelihoods, or confidence intervals placed on the observations. It is important to address uncertainty when offering a comparative analytical test.
Lines 53-55 – the authors cite a 2006 reference for little or no nuclear DNA. They need to search the literature better as such work was reported a decade earlier. Moreover, the statement that mtDNA cannot be used for genotyping is simply wrong. There are many papers describing the use of mtDNA. It may not be as powerful as nuclear DNA but has been used in many forensic cases as well as evolutionary studies.
Line 84 – I looked up the Wong et al paper and it is not the primary paper describing the alkaline lysis procedure. The method is cited (by the same authors) within that paper. It is important to cite the original article. However, alkaline solubilization of hair is hardly novel. It was first described by Budowle, B. and Acton, R. T. In: A technique for the detection of variable electrophoretic patterns of hair proteins. Electrophoresis 2:333-334, 1981. The authors should cite studies that have preceded their own work. Also, lysis seems an incorrect description as that refers to breaking open cells. The protein is likely being solubilized from the matrix. Perhaps the authors could correct terminology.
Line 126 – provide %C and buffers used for second dimension mass separation; provide the pH range or ampholytes used for the first dimension.
Line 138 – cite the silver stain method
Line 252 – the hair is not useless if no nuclear DNA, as mtDNA can be used. The mtDNA has higher discrimination power than the protein profiles described by the authors and it does provide bioancestry (from the maternal line).
Generally okay, but insufficient. Not rigorous regarding addressing statistical uncertainty.
Methods need more citation or details.
May not be statistically sufficient - need to address uncertainty and support for abundance assessments
The authors have shown there is an abundance difference for some identified proteins. The abundance may be real. To support their findings they present x fold differences observed on the 2d gels. There needs to be some discussion about the robustness of the analysis. There are a lot of proteins that overlap and some areas on the gel are quite dense with protein. These areas appear that they may no longer be linear in signal. If so, then it seems suspect about rendering any abundance values. Can the authors provide support for the use such data and/or discuss the limitations.
The authors use the significant difference in abundance to infer gender and bioancestry. Yet there are probabilities, likelihoods, or confidence intervals placed on the observations. It is important to address uncertainty when offering a comparative analytical test.
Lines 53-55 – the authors cite a 2006 reference for little or no nuclear DNA. They need to search the literature better as such work was reported a decade earlier. Moreover, the statement that mtDNA cannot be used for genotyping is simply wrong. There are many papers describing the use of mtDNA. It may not be as powerful as nuclear DNA but has been used in many forensic cases as well as evolutionary studies.
Line 84 – I looked up the Wong et al paper and it is not the primary paper describing the alkaline lysis procedure. The method is cited (by the same authors) within that paper. It is important to cite the original article. However, alkaline solubilization of hair is hardly novel. It was first described by Budowle, B. and Acton, R. T. In: A technique for the detection of variable electrophoretic patterns of hair proteins. Electrophoresis 2:333-334, 1981. The authors should cite studies that have preceded their own work. Also, lysis seems an incorrect description as that refers to breaking open cells. The protein is likely being solubilized from the matrix. Perhaps the authors could correct terminology.
Line 126 – provide %C and buffers used for second dimension mass separation; provide the pH range or ampholytes used for the first dimension.
Line 138 – cite the silver stain method
Line 252 – the hair is not useless if no nuclear DNA, as mtDNA can be used. The mtDNA has higher discrimination power than the protein profiles described by the authors and it does provide bioancestry (from the maternal line).
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.