Whole-Genome Alignment: Methods, Challenges, and Future Directions

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This is an extremely well written article, that sheds light on various WGA algorithms. Its very well organized and each tool is well explained. However, there are some important aspects that have not been elaborated.

Major Comments

The authors have designed this review to talk about various tools, however they have not elaborated about the working principal as well as the mathematical model behind the tool. I would strongly suggest to include such details.

Performance of most of the tools are on different datasets. I would suggest the authors to perform small scale benchmarking for all the tools to compare their performance and then shed light on various strength and weaknesses of the tool. Without this aspect, this article will merely be a basic introduction to various available tools, which might not be very helpful scientifically.

The authors also need to go through the article to make some lengthy explanations more compact. I would suggest using tables to talk about various versions of a tool (e.g., MUMmer). 

I also want the authors to make a comprehensive table that compares the performance, features and drawbacks of various tools. This will be a handy summary for the readers to go for the tool, what meets their requirements.

Minor Comments

Figure 1 is blur, please use better resolution picture

Not a requirement, but I would like the authors to improve figure design, to be more meaningful

 

Comments on the Quality of English Language

The language quality is good. No grammatical errors are there. Its very well written. However, I must say that It can be improved by cutting short length sentences as well as compact explanations.

Author Response

Dear Reviewer,

Thank you for your detailed feedback. Based on your suggestions, here's an overview of the enhancements we've made to the manuscript:

1. Inclusion of New Tools:

We have added Progressive Cactus and GraphAlign to our analysis to expand the range of tools evaluated. This update ensures that our review covers the latest and most relevant tools in the field of whole genome alignment.

We have also included an alignment analysis between human and mouse genomes to better illustrate the capabilities of the different tools.

1. Enhanced Visualization and Analysis:

New figures have been incorporated to illustrate the complexity and execution times of the tools, complementing the existing visualizations. Additionally, we've included traditional dot plots and a Circos plot to provide insights into genome comparisons and highlight similarities between human chromosomes, as demonstrated by Progressive Cactus.

2. mproved Performance Evaluation:

We have detailed the processing times and alignment percentages for each tool, including the newly added ones. This comprehensive evaluation allows for a better assessment of tool performance across various scenarios.

3. Augmented Manuscript with Additional Analyses:

The manuscript now includes additional graphs and analyses comparing the execution times and complexity among the tools evaluated. These enhancements aim to provide a richer, more informative resource for the scientific community, helping researchers make informed decisions about whole genome alignment tools.

Overall, these revisions significantly enhance the manuscript's value by providing a more thorough analysis of whole genome alignment tools, integrating the latest technologies, and improving both the visualization and detailed analysis of the results. We believe these improvements make our review an invaluable resource for researchers in the field of bioinformatics.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors present a review of the different methodological approaches and tools available to perform whole genome alignment (WGA). The different approaches for alignment available before ~2017 are set out clearly, but most of the newer WGA tools (released ~2020 onwards) are only briefly mentioned at the end of the paper. To ensure this review makes a significant advance and is a helpful resource for the bioinformatics community, each of these newer tools should also be described with regard to their methodological approach and performance. One very valuable part of the review is the benchmarking of the alignment results and processing time required for each tool to perform two specific alignments (a human genome vs human genome alignment, and a C. elegans vs bakers yeast alignment). This section should be extended by including a third alignment (e.g. human genome vs mouse genome), and also by including the results obtained for all three alignments for each of the newer tools. This review would then become a powerful resource to help researchers decide which of the currently available tools best suits their whole genome alignment needs. 

 

Some specific comments:

Can you clarify what you mean on line 164 "Comparison between Humans and Human Chromosomes"? Do you mean alignment of two different human genomes? 

Line 290 -  Figure 5 shown here (the Lagan multi-sequence alignment) appears again at Line 386 where it is called Figure 7? Is there different image that should have been at Figure 7?

I'm confused why you mention Lagan on line 326 - aren't points (i) - (v) describing how Mauve works? 

For consistency, Mercator should get its own heading on Line 490 (all the other programs have their own numbered heading)

On line 550 you introduce a new section on programs that use Hashing. Minimap2 uses hashing as an initial step in constructing its indexes, so it should be in this section too?

You briefly mention Cactus and Progressive Cactus on lines 632-633 and  again on line 659. If these programs are a good advance in producing accurate alignments quickly, they should also be detailed in full as for the other tools you fully describe!  Same for MULTIZ, Cloud Aligner, ParaGraph, Deep Align, and EAGER2 which you introduce in section 4.  If this is a review of the latest and best tools for multiple alignment, then you need to include descriptions for all of these more recent tools!  GS Align should also be compared (https://doi.org/10.1186/s12864-020-6569-1 )

 

All these more recent and "better" tools should also be included in your comparative analysis.   I think Section "5. Comprehensive analysis of genomic comparison tools" is the best part of this paper - it is very valuable to the reader to hear how different tools performed when doing the exact same genome alignment tasks. Definitely repeat this section (the human genomes comparison and the C.elegans vs yeast comparison) using these newer tools that you mentioned in Section 4 (Cactus, Progressive Cactus, MULTIZ, Cloud Aligner, ParaGraph, Deep Align, EAGER2) and also GS Align. Please include a discussion of their successes / challenges too. You should also include an alignment comparison where the expected alignment is not as high as between human - human (100%) and not as low as C. elegans vs yeast (only 12% alignment). Something like human vs mouse would be great (genomes are around 90% syntenic)? 

 

Table 2 and Table 3 can be condensed into a single table of results showing the processing times each tool took to complete each alignment (human-human, C. elegans v yeast, and human-mouse) as well as the aligned percentages. The newer tools' results (Cactus, Progressive Cactus, MULTIZ, Cloud Aligner, ParaGraph, Deep Align, EAGER2, and GS Align) should all also be included in the table.  With these additions to your manuscript, this will be a very useful review for the scientific community. 

 

It is great that your code is available on Github - please also update this with the code you use for the alignments using the newer WGA tools.

 

Your references section has many duplications present, ie:

References 18 and 19 are the same paper

References 24 and 25 are the same paper

References 64, 65 and 66 are the same paper

References 75 and 82 are the same paper

You will need to go carefully through the whole paper correcting this. 

 

Comments on the Quality of English Language

The English is ok, but would benefit from some proofreading - especially watch out for tense/plurals, and sentence fragments that don't read smoothly. 

Author Response

Dear Reviewer,

Thank you for bringing this to our attention. We have made the necessary updates to the review based on your suggestion. Specifically, we've added a human vs. mouse alignment scenario for Progressive Cactus, Mummer, and Graphalign, thus expanding our benchmarking section to include these additional alignments. This enhancement allows for a more comprehensive evaluation of the tools' performance across different species.

We believe these updates significantly increase the utility of our review, making it a valuable resource for the bioinformatics community. It assists researchers in selecting the most suitable whole genome alignment tool for their specific research needs.

 

#Comment 2: Some specific comments:

Can you clarify what you mean on line 164 "Comparison between Humans and Human Chromosomes"? Do you mean alignment of two different human genomes? 

AU: The comparison mentioned was conducted between individual chromosomes from different human genomes, rather than between entire genomes. It specifically focused on aligning chromosomes to each other to identify genetic variations between them.

 

#Comment 3: Line 290 - Figure 5 shown here (the Lagan multi-sequence alignment) appears again at Line 386 where it is called Figure 7? Is there different image that should have been at Figure 7?

AU: We apologize for the confusion caused by the mislabeling of the figures. This was indeed an error in the document. I have reviewed and corrected the sequence of all figures to ensure they are now presented in the correct order with appropriate labels.

 

#Comment 4: I'm confused why you mention Lagan on line 326 - aren't points (i) - (v) describing how Mauve works? 

AU: We have corrected it.

 

#Comment 5: For consistency, Mercator should get its own heading on Line 490 (all the other programs have their own numbered heading)

AU: I have added it.

 

#Comment 6: On line 550 you introduce a new section on programs that use Hashing. Minimap2 uses hashing as an initial step in constructing its indexes, so it should be in this section too?

AU: Thank you for your observation. While it’s true that Minimap2 incorporates hashing as part of its indexing process, it primarily relies on a technique called minimizers, which are a form of spaced hashing. We focused the hashing section on tools that predominantly use hashing algorithms as a fundamental part of their alignment methodology. Since Minimap2's use of hashing is just one element of its broader indexing strategy, we categorized it within the section on indexing techniques. We appreciate your attention to detail and the opportunity to clarify this distinction.

 

#Comment 7: You briefly mention Cactus and Progressive Cactus on lines 632-633 and  again on line 659. If these programs are a good advance in producing accurate alignments quickly, they should also be detailed in full as for the other tools you fully describe!  Same for MULTIZ, Cloud Aligner, ParaGraph, Deep Align, and EAGER2 which you introduce in section 4.  If this is a review of the latest and best tools for multiple alignment, then you need to include descriptions for all of these more recent tools!  GS Align should also be compared (https://doi.org/10.1186/s12864-020-6569-1 )

AU: Thank you for your feedback. We have expanded our coverage of Progressive Cactus, GraphAlign, and other new tools discussed in section 4, ensuring detailed descriptions are provided for each. This enhances our review of the latest multiple alignment methods. Additionally, we included a human vs. mouse alignment scenario for Progressive Cactus and GraphAlign, along with the previously mentioned tools. This update allows for a comprehensive comparison of alignment performance across different species, improving the review's utility for researchers in bioinformatics. 

 

#Comment 8: All these more recent and "better" tools should also be included in your comparative analysis.   I think Section "5. Comprehensive analysis of genomic comparison tools" is the best part of this paper - it is very valuable to the reader to hear how different tools performed when doing the exact same genome alignment tasks. Definitely repeat this section (the human genomes comparison and the C.elegans vs yeast comparison) using these newer tools that you mentioned in Section 4 (Cactus, Progressive Cactus, MULTIZ, Cloud Aligner, ParaGraph, Deep Align, EAGER2) and also GS Align. Please include a discussion of their successes / challenges too. You should also include an alignment comparison where the expected alignment is not as high as between human - human (100%) and not as low as C. elegans vs yeast (only 12% alignment). Something like human vs mouse would be great (genomes are around 90% syntenic)? 

Table 2 and Table 3 can be condensed into a single table of results showing the processing times each tool took to complete each alignment (human-human, C. elegans v yeast, and human-mouse) as well as the aligned percentages. The newer tools' results (Cactus, Progressive Cactus, MULTIZ, Cloud Aligner, ParaGraph, Deep Align, EAGER2, and GS Align) should all also be included in the table.  With these additions to your manuscript, this will be a very useful review for the scientific community. 

AU: Thank you for your feedback. To enhance our review's comprehensiveness, we've combined Tables 2 and 3 into a single results table and included processing times for human-human and human-mouse alignment scenarios. Despite challenges with inaccessible websites and compatibility issues with CentOS 7, we successfully installed and tested Progressive Cactus and Graph Align, incorporating their performance metrics into the review.

We've also enriched the manuscript with additional graphs and analyses, aiming to create a more informative resource for the scientific community. These revisions significantly improve the review's utility, helping researchers make well-informed decisions about whole genome alignment tools for their projects.

 

#Comment 9: It is great that your code is available on Github - please also update this with the code you use for the alignments using the newer WGA tools.

 AU: We have included them in the GitHub repository.

 

#Comment 10: Your references section has many duplications present, ie:

References 18 and 19 are the same paper

References 24 and 25 are the same paper

References 64, 65 and 66 are the same paper

References 75 and 82 are the same paper

You will need to go carefully through the whole paper correcting this. 

AU: We have fixed them.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors summarize available whole genome alignment tools, some current challenges and their view of future directions. However, the purpose of this article could easily achieved by publishing on a course, blog or university website. It could be a good book chapter as well. No scientific evaluation or description of scientific output data from the software discussed, is provided. Therefore I think this article is not suitable for publication in Applied Sciences. I also noticed some inappropriate and self citations (e.g., References 1 and 2)

Author Response

Dear Author,

Thank you for your insightful comments and suggestions. We appreciate your perspective and agree that additional scientific evaluation would enhance the manuscript. To address this, we plan to include a comparative analysis of the output data from the various whole genome alignment tools discussed, such as Progressive Cactus and GraphAlign. This addition ensures that our review covers the latest and most relevant tools, further enriched by an alignment analysis between human and mouse genomes to provide deeper insights into their performance.

We believe that synthesizing the available tools with a critical evaluation of their performance significantly contributes to the field, extending beyond the scope of educational materials like course content or blog posts. Our aim is to equip researchers with the knowledge to select the most appropriate tools for their specific needs, advancing genomic sciences.

We will also review all citations to ensure they are appropriate and unbiased, removing any that might be perceived as self-serving. This will ensure our work remains a neutral, informative, and scientifically rigorous contribution to the genome alignment literature.

We trust that these revisions will address your concerns and better align the manuscript with the standards of Applied Sciences.

Reviewer 4 Report

Comments and Suggestions for Authors

Dear Editor and authors,

This review addresses Whole Genome Alignment (WGA) algorithms and tools. In bioinformatics, this technique is important for comparing genomes and detecting genetic variants. Additionally, they can also be used to understand evolution. From a computational point of view, Whole Genome Alignment (WGA) can have a high processing cost since comparisons are made between large quantities of sequences. Therefore, different algorithms and techniques can be used.

 

Overall, this review is interesting and has great potential to attract readers and citations when published. I noticed some formatting errors. Hence, I recommend that the authors do a thorough review (in the minors section I include some errors that I found).

 

Accordingly, here are some suggestions that could improve the text:

 

I think the authors should include an initial section describing classical algorithms such as the Smith-Waterman alignment (local) and the Needleman-Wunsch algorithm (global). These algorithms are briefly cited in some parts (e.g., lines 135 and 270).

 

Also, the authors should highlight more visualization techniques for comparing complete genomes.

 

For example, one of the first tools presented by the authors is MUMmer. There is a traditional plot for comparing genomes, which uses x and y axes to represent genomes and plots a line indicating regions in common (see https://bioinformatics.stackexchange.com/questions/16281/how-to-modify-dot- plot-in-mummer-3-for-bacteria-comparative-genomics).

 

Other examples are tools like Contiguator and even some Biopython packages (see for example this figure: https://pt.wikipedia.org/wiki/Biopython#/media/Ficheiro:Compara%C3%A7%C3%A3o_de_sintenia_entre_genomas. png).

 

Another interesting example is the heatmap used to indicate similarities between genomes (see https://www.researchgate.net/publication/350067650_Journal_of_Biomolecular_Structure_and_Dynamics_ISSN_Print_Pan-genomic_analyses_of_47_complete_genomes_of_the_Rickettsia_genus_and_prediction_of_new_vaccine_targets_and_virulence_ factors_of_the_species/figures?lo=1).

 

Also, the authors could include the colored comparative blocks generated by Mauve (https://genomeintelligence.org/?p=1157).

 

In summary, the authors could include more visualizations, plots, and illustrations. 

 

In the section "3.1. Performance characteristics", cite the complexity of the algorithms (big O notation).

 

In section 5, there was a lack of images demonstrating the results of the comparisons. They can be included in the supplementary materials or included in the GitHub repository that the authors shared (https://github.com/BacemDataScience/WholeGenomeAlignment).

 

 

Finally, check the result of the iThenticate report. Some sentences are identical to the original articles. Although the citations are included in the text, the authors could have paraphrased these excerpts.

 

---

 

Minors:

Please, include a space between the words and the references. For example:

where you see: "genomes[1]." 

change to: "genomes [1]."

 

 

Line 69:  I think there is a typing mistake here: "such as SibeliaZ, BubbZ,[13], and the inno

line 117: I think there is a typing mistake here: "]. Researchers a"

Line 211: I think there is a typing mistake here: "tropy and mutual information )[29]"

Line 230: I think there is a typing mistake here: "[31] .By aligning se" change to "[31]. By aligning se"

Line 314: ".Th algorithm constructs" change to ". The algorithm constructs"

Line 347: "regions outside " change to "regions outside."

Line 394: "rearrangements [42] It is often" change to " rearrangements [42]. It is often"

Lines 520-522-525-536: padronize a escrita de "de Bruijn"

Line 535: "per site[43] This" => "per site [43]. This"

Author Response

Dear Reviewer,

#Comment 1: Dear Editor and authors,

This review addresses Whole Genome Alignment (WGA) algorithms and tools. In bioinformatics, this technique is important for comparing genomes and detecting genetic variants. Additionally, they can also be used to understand evolution. From a computational point of view, Whole Genome Alignment (WGA) can have a high processing cost since comparisons are made between large quantities of sequences. Therefore, different algorithms and techniques can be used.

Overall, this review is interesting and has great potential to attract readers and citations when published. I noticed some formatting errors. Hence, I recommend that the authors do a thorough review (in the minors section I include some errors that I found).

Accordingly, here are some suggestions that could improve the text:

I think the authors should include an initial section describing classical algorithms such as the Smith-Waterman alignment (local) and the Needleman-Wunsch algorithm (global). These algorithms are briefly cited in some parts (e.g., lines 135 and 270).

AU: Thank you for your suggestion and attention to detail. We have included a section in our manuscript that describes classical algorithms such as the Smith-Waterman (local) and the Needleman-Wunsch (global) alignments. These foundational methods are mentioned throughout our review to provide context for sequence alignment techniques.

We believe that a comprehensive overview of both classical and modern alignment algorithms is crucial for understanding the evolution and diversity of methodologies in bioinformatics. We are pleased that your suggestion aligns with our approach and hope this addition enhances the manuscript's value to the scientific community.

#Comment 2: Also, the authors should highlight more visualization techniques for comparing complete genomes.

For example, one of the first tools presented by the authors is MUMmer. There is a traditional plot for comparing genomes, which uses x and y axes to represent genomes and plots a line indicating regions in common (see https://bioinformatics.stackexchange.com/questions/16281/how-to-modify-dot- plot-in-mummer-3-for-bacteria-comparative-genomics).

Other examples are tools like Contiguator and even some Biopython packages (see for example this figure: https://pt.wikipedia.org/wiki/Biopython#/media/Ficheiro:Compara%C3%A7%C3%A3o_de_sintenia_entre_genomas. png).

Another interesting example is the heatmap used to indicate similarities between genomes (see https://www.researchgate.net/publication/350067650_Journal_of_Biomolecular_Structure_and_Dynamics_ISSN_Print_Pan-genomic_analyses_of_47_complete_genomes_of_the_Rickettsia_genus_and_prediction_of_new_vaccine_targets_and_virulence_ factors_of_the_species/figures?lo=1).

Also, the authors could include the colored comparative blocks generated by Mauve (https://genomeintelligence.org/?p=1157).

In summary, the authors could include more visualizations, plots, and illustrations. 

AU: Thank you for your feedback. We have expanded the visualization section of our manuscript to include a wider variety of illustrative methods, including a Circos plot those highlights similarities between human chromosomes based on Progressive Cactus output. This approach enhances our presentation of comparative genomics across different species, showcasing the versatility of visualization techniques in elucidating genomic relationships.

We value your input and believe these enhancements substantially improve the visual component of our manuscript, giving readers a clearer understanding of genome comparisons through diverse illustrative examples.

 

#Comment 3: In the section "3.1. Performance characteristics", cite the complexity of the algorithms (big O notation).

AU: Thank you for your suggestion to include the complexity of algorithms (big O notation) in the 'Performance characteristics' section. We have already detailed the complexity for each algorithm in its description. By providing the computational demands using big O notation, we offer readers a deeper insight into each alignment method's performance characteristics. This enhancement allows for a more thorough evaluation of the tools discussed in the review. 

#Comment 4: In section 5, there was a lack of images demonstrating the results of the comparisons. They can be included in the supplementary materials or included in the GitHub repository that the authors shared (https://github.com/BacemDataScience/WholeGenomeAlignment).

 AU: Thank you for your feedback regarding the need for images demonstrating the results of the comparisons in section 5. We've taken your suggestion into consideration and have included more output and log files in the Github repository.

#Comment 5: Finally, check the result of the iThenticate report. Some sentences are identical to the original articles. Although the citations are included in the text, the authors could have paraphrased these excerpts.

AU: While citations are included in the text, we acknowledge that better paraphrasing could have been employed to avoid verbatim repetition. We have used Plagiarism Checker X

#Comment 6: Minors:

Please, include a space between the words and the references. For example:

where you see: "genomes[1]." 

change to: "genomes [1]."

Line 69:  I think there is a typing mistake here: "such as SibeliaZ, BubbZ,[13], and the inno

line 117: I think there is a typing mistake here: "]. Researchers a"

Line 211: I think there is a typing mistake here: "tropy and mutual information )[29]"

Line 230: I think there is a typing mistake here: "[31] .By aligning se" change to "[31]. By aligning se"

Line 314: ".Th algorithm constructs" change to ". The algorithm constructs"

Line 347: "regions outside " change to "regions outside."

Line 394: "rearrangements [42] It is often" change to " rearrangements [42]. It is often"

Lines 520-522-525-536: padronize a escrita de "de Bruijn"

Line 535: "per site[43] This" => "per site [43]. This"

AU: We have revised the document to include spaces between words and references, corrected typing mistakes, standardized the spelling of "de Bruijn," and ensured consistency in English usage throughout the text.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The changes have been incorporated. 

Author Response

Thank you for acknowledging the changes made to the manuscript. We are glad to hear that the additional content has improved the overall quality and informativeness of the article. Your feedback has been invaluable in shaping the final version of our work.

Reviewer 2 Report

Comments and Suggestions for Authors

in general I was pleased with the additional content that was added into the manuscript.

 

The most important question that remains for the authors is if they actually got the program GraphAlign to run properly (on the Graph Align test dataset used to confirm a correct install, or some other test data that has been successfully used with Graph Align previously). At present, the manuscript reads as if they didn’t get it to work, but it is unclear if this because they are using a fairly old computation setup (Centos7, which is to be discontinued on June 30 2024) and couldn’t get the program to run at all, or if the program GraphAlign really is not capable of performing the human genome alignment.

If the GraphAlign program was not able to run correctly on a test dataset using the authors’ computational setup, then it’s section in Part 5 should be removed.

Comments on the Quality of English Language

The English is ok, but would benefit from some proofreading

Author Response

Thank you for acknowledging the changes made to the manuscript. We are glad to hear that the additional content has improved the overall quality and informativeness of the article.

Regarding the GraphAlign program, we did manage to install it successfully. However, despite multiple attempts, we encountered persistent failures when running the analyses. It seems that the computational demands, particularly when aligning large genomes such as human vs human and human vs mouse, may have contributed to these issues. We have added this information in the manuscript to provide a comprehensive explanation of our findings.

Your feedback has been invaluable in shaping the final version of our work.

 

Reviewer 3 Report

Comments and Suggestions for Authors

Thank you for addressing my comments. This article is now much more informative and valuable.

Author Response

Thank you for acknowledging the changes made to the manuscript. We are glad to hear that the additional content has improved the overall quality and informativeness of the article. Your feedback has been invaluable in shaping the final version of our work.

Reviewer 4 Report

Comments and Suggestions for Authors

The authors addressed most of my suggestions and concerns. I recommend the article for publication. I ask that you pay attention to some minor details (described below). Also, I didn't understand Figure 14. Please check whether it is necessary to include it.

 

Minors:

Line 201 - where you see "O(n*logn)"; is it base 2? If yes, I recommend changing it to "O(n lg n)"

Line 202 - "𝑛n" and "𝑚m", change to "𝑛" and "𝑚"

Line 482 - O(n³)

Line 522 - O(n lg n)

Line 556 - O(n³)

Line 594 - O(n²)

Line 645 - O (n² lg n)

Author Response

Thank you for acknowledging the changes made to the manuscript. We are glad to hear that the additional content has improved the overall quality and informativeness of the article. Your feedback has been invaluable in shaping the final version of our work.

We have carefully considered your suggestions and addressed the minor details.

Regarding Figure 14, we believe it is informative as it explains the execution time for 6 alignment tools. We have provided further clarification in Section 5.5 to ensure its relevance and usefulness in the context of the manuscript and explain that three tools failed to provide the expected alignment.

 

Thank you again for your thorough review and valuable feedback.