Genome-Wide Comparative Analysis of Transposable Elements by Matrix-TE Method Revealed Indica and Japonica Rice Evolution
Round 1
Reviewer 1 Report
Authors developed an interesting method which upgrades the software LTR_finder, aiming to identify truncated and intact LTR/TEs, compared to LTR_finder which was developed for the identification of intact LTR/TEs only.
At first I have to say that I'm not expert in transposable elements, so it is possible the my understanding is not correct. Therefore if you find question unappropriate please let me know, otherwise I recommend to include additional explanations what will make the manuscript easier to follow and without doubts.
Here are questions and comments:
lines 62-63: provide an explanation why GPDF was used.
line 76 - how 1500 bp threshold for the ORFs was determined?
line 81 - which methods were used for the construction of phylogenetic trees by MEGA software?
lines 110-112: "... they "were (should here be "were"?) extracted using the reference sequences...". I don't understand how did you extract sequences of the clusters based on the phylogenetic tree. Couldn't sequences of the same clusters be extracted based on sequence identity matrix (based on the numbers indicating the similarities between the sequences, obtained with BioEdit)?
line 120: "... 1520 Nip ORFs...". Some of the LTR TEs can have more than one ORF. Correct? So you were not counting TEs, but the ORFs?
Could only tblastn and protein sequences from LTR transposable elements be used? If I understand correctly, ORF sequences were used for searching with blastn. Which proteins were coded by these ORF?
What are characteristics of TE clusters - type1, type2, typeRT, typePHA - are these some formal groups of LTR TEs (defined with LTR_finder?) or you gave these groups the names?
Did you take into consideration the length of blastn local alignments?
Do you plan to publish the program scripts and the data together with instructions to make method available to others as well?
Thank you.
Author Response
Reply to reviewer 1:
Dear reviewer,
Thank you very much for your kindly comments and suggestions, it is really nice to correct the issues and make significant improvement for the manuscript accroding to your opinions. And here are my reply point by point:
Question: lines 62-63: provide an explanation why GPDF was used.
Answer: We found the random distribution of SNPs in TE ORFs (Figure 4A,B), so the GPDF model was used to fit the TE peak curves, because GPDF model was matched with stochastic nucleotide substitution events. And the explanation was added to the manuscript. Line 65-67.
Question: line 76 - how 1500 bp threshold for the ORFs was determined?
Answer: the full length TE ORFs could encode five proteins including gag, pr, RT, RH and int, but the ORFs shorter than 1500bp would probably encode less than one protein, so 1500 bp was set as the shreshold. Acctually, no significan cluster signal (identity > 95%) was observed in the super-matrix for the regions with TE ORFs shorter than 1500bp.
Question: line 81 - which methods were used for the construction of phylogenetic trees by MEGA software?
Answer: the molecular phylogenetic trees of the ORFs were generated by MEGA software with Construct Neighbor-Joining Tree method. And the information was added to the manuscritp. Line 85-86.
Question: lines 110-112: "... they "were (should here be "were"?) extracted using the reference sequences...". I don't understand how did you extract sequences of the clusters based on the phylogenetic tree. Couldn't sequences of the same clusters be extracted based on sequence identity matrix (based on the numbers indicating the similarities between the sequences, obtained with BioEdit)?
Answer: yes, you are right. ‘they’ should be ‘were’ there, it was a clerical error and was corrected in the manuscript. And the ORF clusters were extracted based on the sequence identities in the matrix, the discription was corrected in the manuscript. Line 116-118.
Question: line 120: "... 1520 Nip ORFs...". Some of the LTR TEs can have more than one ORF. Correct? So you were not counting TEs, but the ORFs?
Answer: yes, you are right. Acctually, the full length Ty1/Copia and Ty3/Gypsy TE ORFs could encode five proteins including gag, pr, RT, RH and int. So the full length TE ORFs were considered as the intact TEs.
Question: Could only tblastn and protein sequences from LTR transposable elements be used? If I understand correctly, ORF sequences were used for searching with blastn. Which proteins were coded by these ORF?
Answer: yes, you are right. The full length TE ORF sequences were used as the reference sequences for searching and scanning the genomes with BLASTN. And the transposases were coded by these ORFs.
Question: What are characteristics of TE clusters - type1, type2, typeRT, typePHA - are these some formal groups of LTR TEs (defined with LTR_finder?) or you gave these groups the names?
Answer: the TE clusters typeRT and typePHA are the formal groups with detailed characteristics and annotations. But the TE clusters type1 and type2 are not the formal groups with just fuzzy annotations, so we gave the names type1 and type2, probably they would have the formal names in the future studies by other groups.
Question: Did you take into consideration the length of blastn local alignments?
Answer: yes, we have normalized the blastn results by ORFs length. Because all the intact and truncated TE ORFs with various length could be scanned out by blastn method, so all the sequences were normalized to 30 bp length as one unit, the detailed description was in Reference 3 by Huang et al.
Question: Do you plan to publish the program scripts and the data together with instructions to make method available to others as well?
Answer: yes, we will put the scripts and data together with instructions on website to make this method available to other groups.
Reviewer 2 Report
The purpose of the article is reported a detailed description developed a pipeline named Matrix-TE to comprehensively evaluate the differentiation of intact and truncated LTR/TEs in Indica and Japonica rice throughout the whole genomes with a special eye on centromeric regions. Overall, the data indicate that the optimized Matrix-TE approach may be used to specifically analyze TE content, family evolution and time of TE insertions. The article, however, must be improved in terms of writing since some grammar and syntax errors are present in the manuscript. They should address the subject and critically review the information from the literature.
My suggestions:
The authors need to revise the title of the paper in a more meaningful way.
The abstract is written in a way lacks logic. It should highlight the salient findings more critically.
Keywords are present in the title, choose others.
Introduction need more convincing rational for this article.
The introduction has long paragraphs, I suggest reducing the size of the paragraphs.
The results has long paragraphs, I suggest reducing the size of the paragraphs.
The discussion is poorly written hence, needs rewriting. The discussion should be further strengthened by adding some more relevant papers. The literature search is insufficient, only few related research papers in the past three years are cited, add the latest research results appropriately. See the below links if you think it will benefit your discussion.
Rewrite the conclusion! It needs to be much improved.
Author Response
Reply to reviewer 2:
Dear Reviewer,
Thank you so much for your precious comments and suggestions. We appreciate your opinions to make significant improvements for our manuscript. And below are my reply point by point:
Question: The authors need to revise the title of the paper in a more meaningful way.
Answer: yes, we revised the title. And now the title is “Genome-wide comparative analysis of transposable elements by Matrix-TE method reveled Indica and Japonica rice evolution”.
Question: The abstract is written in a way lacks logic. It should highlight the salient findings more critically.
Answer: we made modifications in the abstract at line 11-12, line 14, line17-19.
Question: Keywords are present in the title, choose others.
Answer: we chose new keywords: transposon, divergence. line 23
Question: Introduction need more convincing rational for this article. The introduction has long paragraphs, I suggest reducing the size of the paragraphs.
Answer: we made modifications for the introduction and deleted some redundant descriptions. line 25-45, line 46, line 62, line 65-67.
Question: The results has long paragraphs, I suggest reducing the size of the paragraphs.
Answer: yes, we made modifications in the reulsts and deleted some redundant descriptions. line 110-111, line 116-118, line 121, line 126, line 187.
Question: The discussion is poorly written hence, needs rewriting. The discussion should be further strengthened by adding some more relevant papers. The literature search is insufficient, only few related research papers in the past three years are cited, add the latest research results appropriately. See the below links if you think it will benefit your discussion. Rewrite the conclusion! It needs to be much improved.
Answer: we revised the discussion section and added the latest relevant paper published by Zhang et al. (Reference 38). line 226, line 231, line 240, line 243, line 249, line 260, line 264-265. Actually, the latest publication successfully applied GPDF model to interpret the evolution of autopolyploids in Saccharum species. line 410-413.
