The Transcriptional Cell Atlas of Testis Development in Sheep at Pre-Sexual Maturity

Round 1

Reviewer 1 Report

Wu et al performed scRNA-Seq on 3 month old testis from sheep, identifying 9 somatic cell and 5 germ cell clusters. The single cell transcriptional approach is sound and provides a snapshot of the cell populations present in the testes. However, the description of the results and the conclusions of the authors are not in line with the data presented. English is overall quite poor and makes it difficult to follow the authors' reasoning. 

Major comments:

1) The authors say "spermatogenesis is a process by which male animals transmit genetic information to the next generation". This is not correct, as spermatogenesis (as the word implies) is just the process of sperm formation. This concept is repeated twice in both the introduction and the results. I'm not sure if this is a loss in translation, but the authors should be more precise.

2) At the end of the introduction, they say "it revealed dramatic changes of stem cells in the testes at 3 month-old sheep, candidate factors and pathways regulating germ and somatic cell development". The sentence is not grammatically correct, but in any case, what "changes" are the authors referring to? They analyzed only 1 time point (3 month) and only 1 condition (wild type sheep), so "changes" relative to what?

3) Line 107: "The quality and concentration of RNA were assessed using 1.5% agarose gel electrophoresis and a ThermoScientific Nanodrop". These methods detect concentration, but are not suitable to assess RNA quality.

4) Lines 184-190: the authors say that the sheep testis undergoes dramatic developmental and structural changes, and "this understudied process was characterized by profiling and analyzing the single-cell transcriptome". Again, how are the authors characterizing a developmental process when they use only 1 time point? Also, they say "germ and somatic cells within thetesticular tissue sections were observed at various developmental stages (Fig 1c)". What do the author mean? What developmental stages? Fig 1c is just a simple staining (loos H&E) and does not show anything like that. Again, the study focuses only on 1 time point and 1 condition, so I'm not sure what do they mean with developmental stages or changes.

This issue is important and keeps showing up throughout the manuscript (see above and below).

5) Lines 236-238. How were the DE genes identified? Was each cluster compared to each of the remaining 14? Are the authors saying that these numbers of genes (that go from 621 for smooth muscle cells to 1488 for Sertoli cells) are unique markers for each cluster? These numbers seem very high for being unique cell markers.

6) The authors should explain what they mean when they say (lines 273-275) that the function of undifferentiated spermatogonia was found to be similar to meiotic germ cells. First, there are some differences showing in Fig 4, but most importantly the authors seem to disregard that functional terms in GO are just classification of general functions, they don't necessarily explain the biological function of a cell. For example, two different cell types from two different organs may show the same GO term, but it doesn't mean that the 2 cell types have the same function. Maybe the authors wanted to say that spermatogonia and meiotic germ cells have similar GO MF terms?

7) Lines 281-284. It is difficult to understand this paragraph, grammar and style must be improved. Nonetheless, the authors say "the relationship between the reproductive system and their nich by GO and KEGG analysis of marker genes was conducted". What do they mean?

8) Similar to above, I don't understand what they mean by saying "the marker genes of Sertoli and Leydig cells were associated with the nucleus and transcription in other somatic cells".

9) Lines 294-296: " KEHH pathways, as displayed in figure 5b, were the ribosome pathway in testicular somatic cells, indicating that ribosomes may affect the function of testicular somatic cells". I think the authors misinterpreted completely. It just shows that these cells, as expected, are transcriptionally active. Ribosomes 'affect' the function of pretty much every cell, so I'm not sure what the meaning of that sentence may be.

10) Lines 307-309: The authors should explain this better. The trajectory analysis is interesting, but the authors fail to properly explain it. Also, the significance of Fig 6 is not shown/discussed. 

11) Lines 336-339. I do not agree with the authors when they say "we used 10x genomic sequencing platform to analyze the dynamic transcription process of testicular development". Again, they used only 1 time point and cannot address testicular development.

Similarly, when they say that they "constructed a dynamic transcription map of the cells in testicular development in mature sheep". They cannot claim that with only 1 time point. What the authors did is to construct a transcriptional map (not dynamic) of the testis. 

Again, when using only 1 time point, the result is a snapshot of sample. The authors should limit to describe what they found in terms of cell composition without claiming the identification of developmental changes of processes.

12) The discussion section is way too long, repetitive, and does not describe the data objectively. Lines 408-449 are completely unnecessary and should be deleted. Lines 485-488 repeat (again!) that they identified candidate genes for testicular development, which is not true at all. Lines 491-493: this is misleading as the authors did not address "the relationship between somatic and germ cells". Lines 495-496: not true, the authors did not identify genes by trajectory analysis. Lines 496-497: this is quite a claim and again not true, the authors did not determine the most appropriate time period for the isolation of spermatogonal stem cells. In order to do so they should have done the experiments at different time points and conclude which one was better.

13) Table 1 (line 224) is not present. 

Minor points:

1) In the methods, sections 2.1 and 2.2.1 should be combined.

2) Line 89-90: "About 400 milligrams of tissue were taken from the same region". What region?

3) Although most readers will be familiar with GO, "CC" (line 261), "MF" (line 266), and "BP" (line 286) should be spelled out.

In conclusion, the manuscript presents a cell atlas as per the title. However, all claims of developmental changes, identification of developmental marker genes, stem cells etc are not relevant and not supported by data. The authors should just stick to describing the data they have, which is (again) a snapshot of cell populations in the 3-month old sheep testis.

Author Response

Major comments:

1) The authors say "spermatogenesis is a process by which male animals transmit genetic information to the next generation". This is not correct, as spermatogenesis (as the word implies) is just the process of sperm formation. This concept is repeated twice in both the introduction and the results. I'm not sure if this is a loss in translation, but the authors should be more precise.

Response 1: Thank you for your comments. First of all, forgive me for not expressing it clearly, and modified the statement in the manuscript “Spermatogenesis is a process of sperm formation that is highly regulated by multi-factors.”

2) At the end of the introduction, they say "it revealed dramatic changes of stem cells in the testes at 3 month-old sheep, candidate factors and pathways regulating germ and somatic cell development". The sentence is not grammatically correct, but in any case, what "changes" are the authors referring to? They analyzed only 1 time point (3 month) and only 1 condition (wild type sheep), so "changes" relative to what?

Response 2: Thank you for your advice. Pardon my inaccuracy, but I’ve reorganized the language. See lines 76-78 of the manuscript. “The expression profiles of the testes at 3 month-old sheep was established. Besides, we analyzed the cell populations and marker gene identification results, and also predicted the potential functions.”

3) Line 107: "The quality and concentration of RNA were assessed using 1.5% agarose gel electrophoresis and a ThermoScientific Nanodrop". These methods detect concentration, but are not suitable to assess RNA quality.

Respond 3:  Thank you for your comments. We have reword the sentences in lines 139-140 of the text “1.5% agarose was used detected integrity of 28S and 18S bands and their ratio, the concentration was determined using a ThermoScientific NanoDrop 2000c (ThermoFisher Scientific Inc., Waltham, MA, USA). Only samples with 1.8 < OD 260/280 nm < 2.0, OD 260/230 nm ≥ 2.0, RNA Integrity Number: RIN ≥ 6.5, and 28S/18S ≥ 1.0, were used for subsequent analysis.”

4) Lines 184-190: the authors say that the sheep testis undergoes dramatic developmental and structural changes, and "this understudied process was characterized by profiling and analyzing the single-cell transcriptome". Again, how are the authors characterizing a developmental process when they use only 1 time point? Also, they say "germ and somatic cells within thetesticular tissue sections were observed at various developmental stages (Fig 1c)". What do the author mean? What developmental stages? Fig 1c is just a simple staining (loos H&E) and does not show anything like that. Again, the study focuses only on 1 time point and 1 condition, so I'm not sure what do they mean with developmental stages or changes.

This issue is important and keeps showing up throughout the manuscript (see above and below).

Respond 4: Thank you for your suggestion. First of all, please forgive my inaccuracy, we have reword the sentences in lines 183-189 of the manuscript, and we have redescribed what this means in terms of identifying the cellular composition of testicular tissue, with very strict corrections.

5) Lines 236-238. How were the DE genes identified? Was each cluster compared to each of the remaining 14? Are the authors saying that these numbers of genes (that go from 621 for smooth muscle cells to 1488 for Sertoli cells) are unique markers for each cluster? These numbers seem very high for being unique cell markers.

Response 5: Thank you for the detailed review and we feel really sorry for our carelessness. We have modified the description of identifying the DE genes to “To explore the marker genes for each cell type, we used the Findmarker function in Seurat to screen for differentially expressed genes (DEGs) in each cluster and compare them with other clusters. Genes highly expressed in a specific cell cluster are generally used as markers to identify that cell type. To enhance the accuracy and efficiency of analysis, DE genes were also considered marker genes where expression was observed in at least 10% of the cells in a cluster and |log fold change| >0.25.” In addition,

6) The authors should explain what they mean when they say (lines 273-275) that the function of undifferentiated spermatogonia was found to be similar to meiotic germ cells. First, there are some differences showing in Fig 4, but most importantly the authors seem to disregard that functional terms in GO are just classification of general functions, they don't necessarily explain the biological function of a cell. For example, two different cell types from two different organs may show the same GO term, but it doesn't mean that the 2 cell types have the same function. Maybe the authors wanted to say that spermatogonia and meiotic germ cells have similar GO MF terms?

Response 6: I’m sorry if my inappropriate description made you suspicious. As you said, what we’re trying to expressis is that spermatogonia and meiotic germ cells have similar GO MF terms. we have reword the sentences in lines 271-272 of the manuscript “In conclusion, the function of spermatogonia and meiotic germ cells have similar GO MF terms.”

7) Lines 281-284. It is difficult to understand this paragraph, grammar and style must be improved. Nonetheless, the authors say "the relationship between the reproductive system and their nich by GO and KEGG analysis of marker genes was conducted". What do they mean?

Response 7: Special thanks for your time and we apologize for our careless. We have carefully and thoroughly proofread this paragragh to correct all the grammar and typos, as follows, “Somatic cells in the testis are instrumental in spermatogenesis, we performed the GO and KEGG analysis of marker genes to investigate the different functions of somatic cells. GO analysis demonstrated that translation, the oxidation-reduction process, rRNA processing, and cellular amide metabolic processes were significantly enriched in BP part in all types of somatic cells. While mitochondria, extracellular exosomes, ribosomes, nucleoli, the Arp2/3 protein complex, and the small ribosomal subunit were enriched in CC, and the structural constituent of ribosomes, rRNA binding, mRNA binding, and translation initiation factor activity were significantly enriched in MF. The most significantly enriched GO terms of the marker genes of somatic cells are displayed in Figure 5a, which demonstrates that the marker genes of the Sertoli cells were enriched in the envelope of CC. In addition, the most significantly enriched KEGG pathways in Figure 5b were the ribosome pathway in testicular somatic cells, which indicates that ribosomes may affect the function of testicular somatic cells.” Besides, we are very sorry for our improper description, we have carefully modified the description.

8) Similar to above, I don't understand what they mean by saying "the marker genes of Sertoli and Leydig cells were associated with the nucleus and transcription in other somatic cells".

Respond 8: Thank you for your comments and we are very sorry for our improper description. We have reword the sentences.

9) Lines 294-296: " KEHH pathways, as displayed in figure 5b, were the ribosome pathway in testicular somatic cells, indicating that ribosomes may affect the function of testicular somatic cells". I think the authors misinterpreted completely. It just shows that these cells, as expected, are transcriptionally active. Ribosomes 'affect' the function of pretty much every cell, so I'm not sure what the meaning of that sentence may be.

Respond 9: Thank you for your comments and we are very sorry for our improper description. We have revised the description of figure 5b in section 3.4.2 of the manuscript

10) Lines 307-309: The authors should explain this better. The trajectory analysis is interesting, but the authors fail to properly explain it. Also, the significance of Fig 6 is not shown/discussed. 

Respond 10: First of all, forgive me for not expressing it clearly, we have completed the careful revision of the concluding statement and supplemented the A_dark and A_pale specific marker discussion section in lines 349-359 of the manuscript.

11) Lines 336-339. I do not agree with the authors when they say "we used 10x genomic sequencing platform to analyze the dynamic transcription process of testicular development". Again, they used only 1 time point and cannot address testicular development.

Respond 11: Special thanks for your time and we apologize for our improper description. We have changed the description to “We used the 10x genomic sequencing platform to conduct the expression profiles about the testis of the 3 month-old sheep.”

Similarly, when they say that they "constructed a dynamic transcription map of the cells in testicular development in mature sheep". They cannot claim that with only 1 time point. What the authors did is to construct a transcriptional map (not dynamic) of the testis. 

Again, when using only 1 time point, the result is a snapshot of sample. The authors should limit to describe what they found in terms of cell composition without claiming the identification of developmental changes of processes.

Respond: Thanks very much for your reminder. We have realized that some descriptions are inappropriate in our manuscript. We have modified the sentences according to your suggestions. Sorry for our careless again.

12) The discussion section is way too long, repetitive, and does not describe the data objectively. Lines 408-449 are completely unnecessary and should be deleted. Lines 485-488 repeat (again!) that they identified candidate genes for testicular development, which is not true at all. Lines 491-493: this is misleading as the authors did not address "the relationship between somatic and germ cells". Lines 495-496: not true, the authors did not identify genes by trajectory analysis. Lines 496-497: this is quite a claim and again not true, the authors did not determine the most appropriate time period for the isolation of spermatogonal stem cells. In order to do so they should have done the experiments at different time points and conclude which one was better.

Respond: Thanks very much for your reminder. We have realized that some descriptions are inappropriate in our manuscript. We have deleted lines 408-449 and 485-488 of the manuscript as you suggested. The discussion section of the manuscript has been revised carefully and the objective description of the results of the experiment has been added. As you said, testicular tissue is a complex organ, male spermatogenesis is regulated by many factors, and studying the relationship between somatic cells and germ cell is the next step in our work.

13) Table 1 (line 224) is not present. 

Respond: Thanks very much for your reminder. We have added Table 1 to supplementary Table 4. Sorry for our careless again.

Minor points:

1) In the methods, sections 2.1 and 2.2.1 should be combined.

Respond: Thank you for your advice. We have combined the sections 2.1 and 2.2.1 according to your suggestion.

2) Line 89-90: "About 400 milligrams of tissue were taken from the same region". What region?

Respond: Thank you for your comments. Esticular tissue acquisition information has been added to 2.1.1, see lines 90-91 of the manuscript.

3) Although most readers will be familiar with GO, "CC" (line 261), "MF" (line 266), and "BP" (line 286) should be spelled out.

Respond: Thank you for your reminder. We have changed “GO” to “Gene Ontology (GO)”, “CC” to “Cellular Component (CC)”, “BP” to “Biological Process (BP)”, and “MF” to “Molecular Function (MF)” in the manuscript as you suggested.

Author Response File: Author Response.pdf

Reviewer 2 Report

In general, I think that the quality of the raw data is good despite I have not been able to access them but the analysis of these data may be improved, first by a more detailed description and discussion about the clustering that has been able to reach one of the main achievements of this research: Identify a number of somatic and germinal cell types. And second, by improving the downstream analyses, such as GO enrichment. The english of the paper needs extensive revision, and the quality of the figures improved and labelled with bigger font size.

See the attached PDF for more detailed comments.

Comments for author File: Comments.pdf

Author Response

Introduction

• The meaning of the 1st sentence (line 35): "Pre-sexual maturity in male animals are constantly preparing for spermatogenesis." is not clear, it should be reworded.

Response: Thank you for your comments. First of all, forgive me for not expressing it clearly, and modified the statement in lines 37-39 of the text “Sexual maturity refers to the period after puberty when the male’s body and reproductive organs are further developed, reproductive function is improved and normal fertility is achieved.”

• In second sentence (lines 36-37), the plural of "testis" shoud be "testes" or ..large increase in their volume.. shoud read ..in its volume.. . and "During this period, the testis involves major changes to testis physiology. . . "repetition of "testis". Maybe it may read "..Testes involves mayor physiological changes.."

Response: Thank you for your advice. We have finished the revision according to your suggestion,

We have changed “ testis ” to “ testes ” in line 40, and reword the statement in lines 40-41 of the text “During this period, testes involves major physiological changes, including a large in-crease in their volume and the initiation of somatic cell proliferation/maturation and spermatogenesis.” We have one native English speaking teacher touch up the full text. See the report at the end of the document.

• line 56-57: ". . . four germ cell clusters were identified according to marker genes from the expression patterns of marker genes. . . " -> ". . . according to the expression pattern of marker genes. . . "

Response: Thank you for your comments. We have finished the revision in lines 61-62 of the text “four germ cell clusters were identified according to the expression patterns of marker genes,"

line 58: "spermatigonia" -> "spermatogonia"

Response: Thank you for your advice. Please forgive my misspelling of the word, which has been changed to “spermatogonia”.

Material and Methods

• The first sentence of 2.2.1, starting at line 85, is a repetition of the paragraph stating at line 78. Maybe this 2.2.1 section shoud start at "The testicular tissue was cut . . . ."

Response: Thank you for your comments. The first sentence of 2.2.1 has been removed from this article.

• this sentence (stating at 109) is not clear: "Equal quantities of RNA from each testis tissue were pooled then ribosomal RNA (rRNA) was then digested in DNA-free RNA using TruSeq stranded total RNA . . . " – is "DNAse-free RNAse"??

Response: Thank you for your advice. Please forgive my inaccuracy in describing the experimental method. We have reorganized the language in lines 108-111 of the manuscript “The total RNA of single cell suspension were extracted using TRIzol Reagent (Invitrogen, Carlsbad, CA, USA), and the complementary DNA (cDNA) was synthesized using reverse transcription (RT) regent (Takara, Dalian, Liaoning, China) following the kit instructions.”

• stating at line 140: ". . .±2 standard deviations of the mean standard deviation and the proportion of mitochondrial gene proportion less than 5%. . . " may read ". . .±2 standard deviation and a mitochondrial gene proportion less than 5%. . . "

Response: Thank you for your comments. We have reword the sentences in lines 139-140 of the text “UMI values within ±2 standard deviation and a mitochondrial gene proportion less than 5%”.

• repetition of "simmilar" at 142: ". . . Cells that were similar to others displayed similar. . . "

Response: Thank you for your advice. To further explain the principle of cell clustering, the sentences in the manuscript have been revised in lines 141-142 “Cells that were similar gene expression profiles and so were clustered together to represent a cell population according to the gene expression levels of each cell”.

• line 145: the manuscript uses numbers for bibliographic citations, but author-year was user in this line for Betch et al, 2018.

Response: Thank you for your comments. I sincerely apologize for my carelessness. I have revised the citation format of the manuscript and deleted the errors.

• 2.2.6 (stating at 157): ¿what list of genes were used as background for the GO enrichment analysis? ¿How this analysis was performed? R or python packages? web tools? etc. . .

Response: Thank you for your advice. I’m sorry that I didn’t elaborate on the GO enrichment methodology, which makes it difficult for you to understand. Enrichment analysis is done with the online tool David (https://DAVID.ncifcrf.gov/). The Ovis aries reference genome (Oar 4.0) was selected as the background gene.

Results

• Figure 1 is dispensable. It does not add any clarification over the data already stated in the text. Even 1b and 1c are not cited in the text.

Response: Thank you for your advice. F1 is a summary of the data collection and experimental workflow. F1 b is the cell viability and cell number of testicular single cell suspensions, and F1c is Histomorphological analysis of sheep testes, and Figure 1 has been added to the supplementary material s1.

• starting at 212: "The literature was earched and the cell marker database (http://biocc.hrbmu.edu.cn/CellMarker/) used to identify the cell types.." shoud be reworded (and typos corrected "earched" -> "searched")

Response: Thank you for your comments. We have reword the sentences in lines 214 – 217 in the manuscript “The literature was earched and the cell marker database (http://biocc.hrbmu.edu.cn/CellMarker/) used to identify the cell types in each cluster. Besides, we tested the expression localization of known marker genes in our cells and further predicted the cell types”

• Figs 2a and 2c are redundant, 2c shoud be enough. Labels in this figure are hardly readable and in some cases completely unreadable.

Response: Thank you for your advice. Figs 2a and 2c have been removed from this article, and labels in Figs 2b has been re-edited and typewritten to ensure it is readable.

• There are some contradictions in the markers of the different cell types, for instance Sertoli cells (line 219) where identified by SOX9 expression which is correct but in table 3, cited at line 222 for additional markers, expression of SOX9 is reported in the top 100 markers of the Spermatids cluster, but not in the Sertoli cluster.

Response: Thank you for your comments. We have re-selected the first 100 differentially expressed genes in each cell subgroup and modified the errors in table3.

• Table 3. Does the top 100 means the top 100 most expressed markers? Are them ordered by expression level in the table? or by any other criteria?

Response: Thank you for your comments. Marker genes are defined as genes that are highly expressed in the majority of cells in a given cell population and only a small percentage in the rest of the cell population, the gene was up-regulated in this cell group compared with other cell clusters. The specific Marker genes of each cell group were screened by BIMOD test, which was used to detect the difference between the designated cell group and all other cell clusters. The top 100 characteristic genes with the highest expression in each cell cluster are listed in Table 1.

• What method was used for clusterization and what sensitivity? changes in sensitivity may lead in over- or under-clustering. It is difficult to choose the more appropriate sentivity for a given dataset but maybe a tool such as SCCAF: https://github.com/SCCAF/sccaf may help to evaluate whether

the unknown cluster is a product of mis-clustering or if it really is a different new biologically significant cell type. Even if SCCAF if made to work with anndata objects from SCANPY, SEURAT objects can be converted to anndata or, alternatively, the same analysis made with SEURAT can be

performed with SCANPY.

Response: Thank you for your suggestion. The dimensionality reduction algorithms used in this experiment are MNN (mutual nearest neighbors) and T-SNE (t-distributed Neighbor Embedding, t-distributed Neighbor Embedding). Based on MNN (mutual nearest neighbors), the results of dimensionality reduction are visualized by t-SNE, and the clustering algorithm is SNN, and finally the optimal cell grouping is obtained.

We try to keep the number of subpopulations within a reasonable range by limiting the number of principal components and the resolution value in Seurat analysis.

Seurat was created as a result of our data analysis, which focused on the cellular composition of the testes of three month old hu sheep, your suggestion will be an important guideline for us to study different subpopulations in the next step. Thank you again for your constructive comments.

• Figure 3 should be supplementary. At least 3a. In the legend of the figure: "The numbers 0-14 in the X axis. . . "does it refer to 3A? there are no numbers on the X axis but the actual names of the cell types. Thus, this whole sentence should be removed.

Response: Thank you for your advice. Figure 3A has been added to Supplemental Figure s3, and the sentence of “"The numbers 0-14 in the X axis. . .” has been removed.

• It is not clear how the GO and KEGG enrichment analysis was performed. What tools where used. To what background they were compared. It may be useful to provide the actual statistical significance of the enriched terms in a supplementary table, or the raw ouput of the functions used for the analysis besides the color-coded P values in figs 4b and 5b. What is the meaning of the background colors in the lists of figs 4a and 5a? If the analysis is not correctly performed or interpreted it can lead to misscon clusions or conclussions that are not relevant. for instance, the position of the term in the hierarchy is important as more general terms include a higher number of genes by grouping genes included in descendant terms in the DAG. For instance, the conclusion at lines 295-296: "Furthermore, the most significantly enriched KEGG pathways, as displayed in Figure 5b, were the ribosome pathway in testicular somatic cells, indicating that ribosomes may affect the function of testicular somatic cells" Is a very generic term that obviously not only may affect the function of testicular somatic cells, but of any other cell type in the organism because ribosomes are involved in the syntesis of all proteins. Other descendant terms of this one in the DAG hierarchy shoud be considered, even if they are less statistically significant because they may provide more useful information about functions that are more specific to testicular somatic cells.

Response: Thank you for your comments. I’m sorry that I didn’t elaborate on the GO enrichment methodology, which makes it difficult for you to understand. Enrichment analysis is done with the online tool David (https://DAVID.ncifcrf.gov/). The Ovis aries reference genome (Oar 4.0) was selected as the background gene.

Data information for GO enrichment and KEGG is added in Supplementary Table 4. The background colors in the list of figures 4A and 5A have no special meaning, just to list the enriched items in each cell cluster.

Special thanks for your time and we apologize for our improper description. We have modified the sentences on lines 288-290 of the manuscript according to your suggestions.

• The color legend in fig 6 is a little confusing in this horizontal orientation it would be better in a vertical orientation with each pair of color cell_type in a line.

Response: Thank you for your advice. Figure 6 has been redrawn as you suggested, and added the captions on lines 311-312 of the manuscript.

• lines 320-322: "Immunohistochemical staining also confirmed these changes in expression levels of the genes in the sheep estes (Figure 7b)". Immunohistochemistry is good for a spatial localization of the expression of a gene but not for expression levels. Observing an expression signal in few cells of an histological section may confirm that a particular gene is expressed but not its level of expression.

Response: Thank you for your comments. Pardon my inaccuracy, but I’ve reorganized the language. See lines319-320 of the manuscript.

• The caption of Figure 7 has typos and syntactic errors.

Response: Thank you for your comments. Please forgive my misspelling of the word, we have reword the caption of Figure 7 in lines 323-326 of the manuscript.

Discussion

• Line 343-344: "In addition, the sequencing data were submitted to the GEO database". A reviewer token should be provided. See: https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/inf o/reviewer.html second paragraph: "Once a GEO submission is complete and has been approved by curators, the authors are invited to generate the reviewer token - this is a very quick process, the author simply has to log in to their GEO account and click the ’Reviewer access’ link located at the top of their Series record. Authors are instructed to include this token with their manuscript when they send it to the journal for review. The journal editors should ensure that reviewers receive the token. The reviewers then can enter the token on the Series record page to get access."

Response: Thank you for your advice. I am sorry that the original data was not immediately released due to the need for follow-up research. Immediate release has been requested and can be accessed through the login number in the data availability statement. The accession no: GSE184343.

• line 387: "cellular stimuli or toxic insults" I suggest to change to "cellular stimuly or toxicity" or something simmilar that do not contain "insults".

Response: Thank you for your comments. We have changed “cellular stimuli or toxic insults” to “cellular stimuly or toxicity” in line 389 according to your suggestion.

• The discussion is mostly a generic description of the diferent cell types more appropriate for a review about the cell types in the testis. Dicussion related with the actual findings in this research is vague and no clear conclusions were reached. The last paragraph, that tries to reach a final conclusion, reveals that the only relevant contribution is that scRNAseq has been able to identify a number of germ and somatic cell types and that the signaling pathways required for spermatogenesis are complex.

Response: Thanks very much for your reminder. We have realized that some descriptions are inappropriate in our manuscript. We have deleted lines 408-449 and 485-488 of the manuscript as you suggested. The discussion section of the manuscript has been revised carefully and the objective description of the results of the experiment has been added. As you said, testicular tissue is a complex organ, male spermatogenesis is regulated by many factors, and studying the relationship between somatic cells and germ cell is the next step in our work.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have answered to my comments appropriately.  I commend their effort in improving the manuscript. 

Author Response

Thank you very much for taking the time to review our manuscript and make important suggestions, as well as for your approval of our revised manuscript.

Best regards,

YiWu

Author Response File: Author Response.docx

Reviewer 2 Report

The authors have taken care of most typographical and editorial errors throughout the manuscript, although some typos remain:

line 58-59 spermatigonia

line 207 The literature was earched

I still have doubts about the validity of the enrichment analysis. The last version of DAVID 6.8 has not been updated since October 2016 and this can have an impact on the interpretation of results. See, as example, pages 193 and 194 and Fig. 2 of "The Gene Ontology Handbook":

https://0-link-springer-com.brum.beds.ac.uk/book/10.1007/978-1-4939-3743-1

Furthermore, a specific GO term is said to be over-represented when the number of proteins assigned to that term within the family of interest is enriched versus the background model. Thus, choosing the correct background model is relevant and is dependent on the dataset and the question that is being asked (see section 2.2 of the same book). The whole genome is not always an apropriate background model. For instance, is the question is to know which GO terms are enriched in testis before sexual maturity as compared to sexual mature testes, the set of all proteins expressed in testis is a more appropriate background model. It would makes no sense to include as background genes whose expression is specific for liver or brain (which are included in the whole genome background model) because they are not included in the set of studied genes.

It seems that the authors have just uploaded their list of genes to DAVID, choosed a generic background (whole goat genome) and described in the paper some of the terms that appeared in the list of enriched terms without taking into account what question they want to answer and thus not reaching any conclusion beside the description of some of the enriched terms that appears in the results.

Although the results of the enrichment analysis is a list of terms, they are actually part of a graph and we cannot ignore that structure. A general term may be over-represented becasue some, or even just one, of its descendand is over-represented and this descendant will have a more specific role in testis that the more general term and may give a more specific and relevant information. I suggest other tools such as GATOOLS that may help to interpret the results as being part of a graph:

https://github.com/tanghaibao/goatools

https://0-www-nature-com.brum.beds.ac.uk/articles/s41598-018-28948-z

I truly recommend to repeat the analysis with different backgrounds to assess different questions and to have a look at more specific GO terms in the hierarchy, descendant from the general terms found in this study. I thing GOATOOLS is a usefull tool for this purpose. Maybe performing analyses  first without propagating parent GO terms up the hierarchy to help identify the more specific terms or constructing subtrees starting from GO terms of interest.

Thanks for releasing the GSE184343 data but for the review this is not necessary. GEO provides a reviewer token that allow reviewers to access the data even if they are not publicly released.

I think the paper has improved but may still have some caveats in the GOEA analysis. The scRNA data may be useful for further studies and may deserve publication but in its present form the paper should be considered of low priority unless a reanalysis of GO terms provide more relevant conclusions.

Author Response

The authors have taken care of most typographical and editorial errors throughout the manuscript, although some typos remain:

line 58-59 spermatigonia

Response: Thank you for your advice. Please forgive my misspelling of the word, which has been changed to “spermatogonia”.

line 207 The literature was earched

Response: Thank you for your comments. We sincerely apologize for my carelessness. We have changed “ earched ” to “ sarched ” in line 207 of the manuscript.

I still have doubts about the validity of the enrichment analysis. The last version of DAVID 6.8 has not been updated since October 2016 and this can have an impact on the interpretation of results. See, as example, pages 193 and 194 and Fig. 2 of "The Gene Ontology Handbook":

https://0-link-springer-com.brum.beds.ac.uk/book/10.1007/978-1-4939-3743-1

Furthermore, a specific GO term is said to be over-represented when the number of proteins assigned to that term within the family of interest is enriched versus the background model. Thus, choosing the correct background model is relevant and is dependent on the dataset and the question that is being asked (see section 2.2 of the same book). The whole genome is not always an apropriate background model. For instance, is the question is to know which GO terms are enriched in testis before sexual maturity as compared to sexual mature testes, the set of all proteins expressed in testis is a more appropriate background model. It would makes no sense to include as background genes whose expression is specific for liver or brain (which are included in the whole genome background model) because they are not included in the set of studied genes.

It seems that the authors have just uploaded their list of genes to DAVID, choosed a generic background (whole goat genome) and described in the paper some of the terms that appeared in the list of enriched terms without taking into account what question they want to answer and thus not reaching any conclusion beside the description of some of the enriched terms that appears in the results.

Although the results of the enrichment analysis is a list of terms, they are actually part of a graph and we cannot ignore that structure. A general term may be over  represented becasue some, or even just one, of its descendand is over-represented and this descendant will have a more specific role in testis that the more general term and may give a more specific and relevant information. I suggest other tools such as GATOOLS that may help to interpret the results as being part of a graph:

https://github.com/tanghaibao/goatools

https://0-www-nature-com.brum.beds.ac.uk/articles/s41598-018-28948-z

I truly recommend to repeat the analysis with different backgrounds to assess different questions and to have a look at more specific GO terms in the hierarchy, descendant from the general terms found in this study. I thing GOATOOLS is a usefull tool for this purpose. Maybe performing analyses  first without propagating parent GO terms up the hierarchy to help identify the more specific terms or constructing subtrees starting from GO terms of interest.

Response: First of all, thank you for your constructive suggestions on GO enrichment analysis in our manuscript. GOATOOLS, a Python-based library, makes it more efficient to stay current with the latest ontologies and annotations.

In order to study the cellular composition and biological processes involved in the testis tissue of sheep during pre-sexual maturation, the GO enriched background gene was selected as the annotation document after the testis single cell sequencing. The results of a new enrichment analysis using GOATOOLS for each of clusters showed that GOATOOLS provides GO terms by median descendant count are twenty times more specific than the broad GO terms from DAVID6.8, and similar in specificity to DAVID6.8 GO terms. GOATOOLS currently uses the Fisher’s exact test to compute uncorrected P-values.

GOATOOLS were better able to find specific GO terms. For example, germ cell and Meiotic nuclear division were significantly enriched in the testis of sheep at the early stage of Pre-sexual maturation, providing insights into the development of male germ cell in sheep. Finally, we added the GOATOOLS results to the supplemental table 4, and reorganized the manuscript sections 2.2.5, 3.4.1 and 3.4.2. As you said, testicular tissue is a complex organ, male spermatogenesis is regulated by many factors, and studying the relationship between somatic cells and germ cell is the next step in our work.

Thank you again for your important suggestions, and we hope that our revision of the manuscript will meet with your approval.

Thanks for releasing the GSE184343 data but for the review this is not necessary. GEO provides a reviewer token that allow reviewers to access the data even if they are not publicly released.

Response: Thank you so much for reminding me.

Author Response File: Author Response.docx