Sex-Specific Expression of Non-Coding RNA Fragments in Frontal Cortex, Hippocampus and Cerebellum of Rats

Round 1

Reviewer 1 Report

In the manuscript by Fiselier and colleagues entitled “Sex-specific expression of non-coding RNA fragments”, the authors profile the abundance of four major non-coding RNAs tRNA, snoRNA, snRNA and rRNA in 4 brains issues in rats, and found brain region-specific and sex-specific differences in expression of certain ncRNA.  I am unable to provide technical input on the genomic analyses, but ncRNA profiling remains scarce in the field, and I believe this study would make a good community resource for future studies.   

Major comment:

It is not clear to me how many biological replicates were processed in the study. Since the entire study is about quantifying different ncRNA populations, biological replicates are essential to make conclusions. Analyses should be confirmed in the biological replicates. 

Minor comments:

1) The study has many findings. It would be helpful to summarize the findings as a Figure or table to make it more digestible to the reader.

2) The tissue or sex-specific differences are simply due to cell type composition differences. The authors should at least discuss limitations and caveats for tissue-level analysis in the paper.

3) Typo in abstract: “ Reads mapping to lincRNAs were significantly larger in CER as compared to HIP and CER” I think should be "HIP and FC."

Author Response

In the manuscript by Fiselier and colleagues entitled “Sex-specific expression of non-coding RNA fragments”, the authors profile the abundance of four major non-coding RNAs tRNA, snoRNA, snRNA and rRNA in 4 brains issues in rats, and found brain region-specific and sex-specific differences in expression of certain ncRNA.  I am unable to provide technical input on the genomic analyses, but ncRNA profiling remains scarce in the field, and I believe this study would make a good community resource for future studies.   

Major comment:

It is not clear to me how many biological replicates were processed in the study. Since the entire study is about quantifying different ncRNA populations, biological replicates are essential to make conclusions. Analyses should be confirmed in the biological replicates. 

Our response

The data are the compilation of three biological repeats per each experimental unit (male or female). This is now mentioned in the new version of the manuscript.

Minor comments:

1) The study has many findings. It would be helpful to summarize the findings as a Figure or table to make it more digestible to the reader.

Our response

Yes, this is a great idea. We now compiled all the data into Table 2. In fact, Table 1 also covers some of the results.

2) The tissue or sex-specific differences are simply due to cell type composition differences. The authors should at least discuss limitations and caveats for tissue-level analysis in the paper.

Our response

It is difficult to be certain whether the differences are due to the differences in the ratio of various cells in the brain regions or the differences between the same cells in males and females. In the revised version, we have discussed as to what is known about the differences in cell composition in specific regions of the brain between males and females.

3) Typo in abstract: “ Reads mapping to lincRNAs were significantly larger in CER as compared to HIP and CER” I think should be "HIP and FC."

Our response

Yes, this is now corrected, thanks!

Reviewer 2 Report

This is a very interesting work on functional fragments from ncRNAs, a very recent concept that increases the span of RNAs involved in gene regulation. The authors analyzed their abundance in different regions of rat brains to find región and sex-depending differential expression of these ncRFs.

Although the topic is very interesting the organization of the work and the manuscript is at least peculiar and different relevant questions are unanswered. In my opinion, much work has to be made for this manuscript to be published in “Epigenomes”

Concerns

I would prefer to add the raw data of the pathway analysis (and GO analysis) to the main text instead of the Venn’s diagrams. I think that this would be more informative

-An important concern deals with the fact that most of the ncRFs come from miRNAs or from repeated elements (Figure 1), but the authors analyse those from tRNAs because “tRF reads were the most abundant among the four ncRF read types analyzed (line 163)”. Why they do not analyse ncRFs from miRNAs or from repeated elements?

-Size distribution of reads from miRNA-RFs is very similar to that of mature miRNAs. What are the differences among these? Could miRNA-RFs compete with mature miRNAs? Is the target distribution similar among these two groups of sRNAs?

-What are the repetitive elements? Do their RFs preferentially map to a specific sequence/structure? Do they include the internal promoters of the repetitive elements?.

-What is the basis for the differential expression of tRF-Gly?

-In general I think that it would be much more relevant to analyse miRNA-RFs, their relationship with mature miRNAs, the potential overlap of targets, and the distribution of RFs that include seed sequences.

-The conclusions section seems to have missed some text since the statement made is not a standard conclusion

-The Methods section is very poor in details, just to highlight some I wonder how the GC content determination was performed (paragraph 2.3) or which software was used to draw Venn’s diagrams. Also, it is not clear how the target prediction was performed.

Author Response

Reviewer 2

This is a very interesting work on functional fragments from ncRNAs, a very recent concept that increases the span of RNAs involved in gene regulation. The authors analyzed their abundance in different regions of rat brains to find región and sex-depending differential expression of these ncRFs.

Although the topic is very interesting the organization of the work and the manuscript is at least peculiar and different relevant questions are unanswered. In my opinion, much work has to be made for this manuscript to be published in “Epigenomes”

Concerns

I would prefer to add the raw data of the pathway analysis (and GO analysis) to the main text instead of the Venn’s diagrams. I think that this would be more informative

Our response

We actually have submitted raw data in Supplementary Files. There are hundreds of genes, so we would need pages and pages to cover it.

-An important concern deals with the fact that most of the ncRFs come from miRNAs or from repeated elements (Figure 1), but the authors analyse those from tRNAs because “tRF reads were the most abundant among the four ncRF read types analyzed (line 163)”. Why they do not analyse ncRFs from miRNAs or from repeated elements?

Our response

Technically, miRNAs are not processed into fragments, but miRNA precursors (hairpins) can be processed into isomiRs. In this report, we have studied ncRFs, fragments that are produced from mature ncRNAs. Processing of miRNA precursors is quite different by the mechanism and we plan to analyze such differences in the future.

-Size distribution of reads from miRNA-RFs is very similar to that of mature miRNAs. What are the differences among these? Could miRNA-RFs compete with mature miRNAs? Is the target distribution similar among these two groups of sRNAs?

Our response

Yes, because miRNAs are not processed to miRFs. The difference, if any, is from the loss of nucleotides from either side of the miRNA due to degradation.

-What are the repetitive elements? Do their RFs preferentially map to a specific sequence/structure? Do they include the internal promoters of the repetitive elements?.

Our response

Those fragments that mapped to repetitive elements are those that did not belong to the category of lncRNAs and piRNAs – commonly stemming from repetitive elements. Very little is known about them – at least, they do not form any coherent functional group of ncRNAs. It also appears that their average size and distribution appears to similar between males and females or different brain areas. That is why we have not analyzed them.

-What is the basis for the differential expression of tRF-Gly?

Our response

Hard to say; definitely due to the more frequent processing of tRNA-Gly. It is hard to say what is the biological significance of this. tRF-Gly was found to be the most abundant of the nuclear tRF fractions in the mouse intestine and in monkey’s brains. In addition, 5′ tRH-GlyGCC was also found to be dynamically expressed during stem cell differentiation. We have mentioned this in the discussion.

-In general I think that it would be much more relevant to analyse miRNA-RFs, their relationship with mature miRNAs, the potential overlap of targets, and the distribution of RFs that include seed sequences.

Our response

See our response above. We will definitely analyze isomiRs in the future, but because miRNAs are not processed into miRFs, we did not analyze it in this work.

-The conclusions section seems to have missed some text since the statement made is not a standard conclusion

Our response

We expended the conclusion section.

-The Methods section is very poor in details, just to highlight some I wonder how the GC content determination was performed (paragraph 2.3) or which software was used to draw Venn’s diagrams. Also, it is not clear how the target prediction was performed.

Our response

We have added details in the new version of the manuscript. We calculated the GC content of ncRNA reads by dividing the G and C count by the total nucleotide count in ncRNA reads. miRDB is an online database tool for miRNA target prediction. We entered the ncRF sequences into the miRDB web interface and miRDB returned the predicted targets. More details are in the text. Venn diagrams for overlapping gene targets were built using on the lists of unique targets using the R package VennDiagram.

Round 2

Reviewer 1 Report

The authors have appropriately addressed all of my minor concerns. However, the one major concern has not been addressed fully.

Rather than compiling or averaging the data, the analysis should be confirmed in the three biological replicates to get a better sense of the variation between replicates. Is it not possible that one replicate may be skewing the results? Such information will be masked by compiling the data into one.

Ideally this would be shown as data in figures and tables. Alternatively, the authors should state in the main text that all findings reported have been consistently observed when comparing individual replicates. 

Author Response

We added the following sentences:

In to the Results section:

In this work, we have combined the sequencing data from biological repeats into one sample; this was done to increase the number of reads for rare reads mapping to ncRFs, especially in those cases where the processing from ncRNAs was rare. There was little variation in quality of reads or distribution of any specific ncRNA group among biological repeats, thus pooling sequence data together did not affect the quality of the analysis.

In to Methods:

For the analysis of sequencing data, biological repeats were pooled together.

Reviewer 2 Report

This is an interesting manuscript on a poorly known family of small ncRNAs. The authors have make an effort to address my concerns. I would only recomend that some of the answers to my concerns be added to the text, e.g. those regardin miRNAs.

Author Response

We added two sentences:

First to the results section:

Even so miRNAs were the most abundant among all ncRNAs, they are not processed into ncRFs, thus we did not analyze this category of ncRNAs further. 

Second to conclusions:

Also, in this work we did not analyse variant miRNA transcripts, so called isomiRs, because they are not the result of processing of mature miRNAs. In the future, however, it would be important to analyze them, as they may also be produced in tissue- and sex-specific manner.