Talk:Preprint/Phylogenetic reconciliation

=Ross Mounce's initial review of the first submission=

A bit about myself to reassure editors, authors and other readers of my competencies: http://topicpageswiki.plos.org/wiki/Reviewer_Guide
 * I confirm that I have read the PLOS provided reviewer guide for this article type:
 * I note that adherence to Wikipedia core content policies is desired. (https://en.wikipedia.org/wiki/Wikipedia:Core_content_policies )
 * I have previously reviewed a topic page for PLOS Computational Biology.
 * I have a reasonable understanding of both Wikipedia where I edit under the username Metacladistics, and phylogenetic estimation methods and software implementations e.g. tanglegrams in Dendroscope. A list of my academic works can be found at my ORCID page: https://orcid.org/0000-0002-3520-2046

I thank the authors for even attempting to write a review in Wikipedia style – it’s a hard task and certainly not one I would attempt myself – it requires real mastery of a topic as well as understanding of Wikipedia rules and styles. There have not been many such wiki topic pages published so far…

Comments on the Conclusion
Bearing in mind that this is a review article, and is intended to be merged into EN-Wikipedia after publication, I wonder why it has a ‘Conclusions’ section? The ‘Conclusions’ section is not Wikimedia-compliant and the section would no doubt get deleted if posted over there as there are far too many unevidenced assertions such as “Reconciliation is now mature as a methodological research subject” – the cited paper (RecPhyloXML: a format for reconciled gene trees) introduces the format to the academic literature (in 2018). From an encyclopaedic point-of-view there’s no cited evidence that reconciliation is “mature” (and I wonder what ‘mature’ even means). I would instead encourage the authors to replace the ‘Conclusions’ section with a ‘Future Directions’ section as has been done for past PLOS Comp. Bio. Wiki topic page articles e.g. Viral Phylodynamics and Ancestral Reconstruction.

Comments on Table 1
Table 1 is “Pairs of biological levels compared with phylogenetic reconciliations” Please compare this with the software table in the Ancestral Reconstruction wiki topic page. I would like to see this table approached from the point of view of prospective users wanting to try-out phylogenetic reconciliation methods and data visualisation softwares & web interfaces. Please link-out to software homepages, as well as academic paper references. On who’s opinion is it that the select softwares are the “main” ones in usage in phylogenetic reconciliation research? I fear the main softwares column may be a rather subjective or arbitrary set of choices. Could the authors not take a more inclusive approach and include more options? Consider these for inclusion in the software table:

BioGeoBEARS? RASP? COMPONENT? TreeMap? PACo? COALA? Duptree? RASCAL? ParaFit? Phyldog? Mixtrem? SylvX? And the University of Leipzig software packages such as CoRe-PA and CoRE-IPL

Perhaps there needs to be a separation in the discussion between (1) models/methods and the (2) softwares that can implement some of them, and (3) the specific disciplinary applications of those methods and softwares?

For the first row of Table 1 I would quibble that ‘Geography’ is not a biological level. The title of the table is “Pairs of biological levels compared with phylogenetic reconciliations”. Either change the table title or the wording of geography to ensure there is no conflict in meaning/scope. It would be nice to have the licensing of each of the softwares indicated.

Wikipedia pages that should link here
I think there are lots of potential linkages between this work and the Ancestral Reconstruction Wikipedia page (which was also the fruit of a PLOS Comp Bio article).

General Comments
With the exception of figure 1 and figure 5 the figures are in my opinion far too complex and hard to follow. I have a strong dislike for multi-compartmental figures and I can’t help but note that figure 3 goes from 3a to 3m, thirteen different panels plus legend! I can’t think of another Wikipedia article that has a 13-compartment image.

I’m really surprised not to see the word “tanglegram” used anywhere in this work even though I perceive they are there in the figures albeit under “leaf rearrangement”. I think this is symptomatic of the underlying problem of this review – it jumps too quickly into high complexity without starting from basics, the history of the origin of these methods (e.g. Brooks Parsimony Analysis?), and simple easy to understand empirical examples from the literature e.g. a host<->parasite tanglegram example in a well-studied system e.g. chewing lice (Geomydoecus) and pocket gophers (Rodentia: Geomyidae).

Alternatively, I wonder if the authors might want to ward-off my objections by more tightly focusing the scope of their work here to ‘Gene tree species tree reconciliation’ which I think is the focus of most of their work presented here. Trying to provide figure images that encompass all the different variations becomes too abstract, too ‘meta’ to easily understand. Whereas focussing-in on specific use-cases might aid clarity through specificity? I simply struggled to understand the flow of argument between sections of this work if I’m honest, particularly because it lacks a sense of history or chronology in methods development. And who would understand the codebox bit below? Not I. It’s too much - I can't imagine many encyclopaedia articles have that much code in them.

c(U,l) =  min(        cS + min( c(U',l')+c(U",l"), c(U",l') + c(U',l"), c(U',l) + cL, c(U",l) + cL )        cD + c(U,l') + c(U,l")        cT+ min( minV(c(V,l')) +c(U,l"), minV(c(V,l"))+c(U,l'))         )

Above all I’m also a little bit concerned that this work might be a bit biased towards work done or led from Lyon and/or in France. I would hope that another reviewer with deeper knowledge or phylogenetic reconciliation literature from Leipzig (e.g. Nicolas Wieseke), Australia (e.g. Mike Charleston), or the USA (e.g. Mukul Bansal’s research group or the Libeskind-Hadas Research Group, at Harvey Mudd College) might be able to provide comment on this - does it represent the topic fairly? I think particularly the history of the development of methods and software from 1980-2010 is largely absent. Can the authors please consider writing a 'History' subsection for the article, or taking a chronological approach to introducing the concepts in the article?

Rossmounce (talk) 07:54, 3 August 2021 (PDT)

= Review by Mukul Bansal =

This topics page article on phylogenetic reconciliation does a good job of reviewing and logically organizing a lot of existing research on the subject. A particularly interesting aspect of this article is that it identifies major shared themes and ideas behind prior reconciliation-related research in different application domains. The article also clearly identifies limitations of existing work and provides insight into useful future research directions. Certain aspects of the article, including its organization, figures, and coverage of relevant research, could be improved.

Organization
While the overall organization of the article is generally fine, transitions between different sections and subsections can sometimes be abrupt. Section and subsection titles could also be improved. Specific suggestions follow:


 * The last sentence of Section 2 (“Two-level reconciliation methods …”) seems out of place. It would help to add some context around that sentence. It would also help to include some text explaining how the article is organized (at the end of Section 2).
 * Since the focus of this article is on DTL-like reconciliation models, it would help to explicitly clarify this early on. For example, you don’t really discuss Duplication-Loss-Coalescence models or hybridization models.
 * Title of Section 3: The Duplication Transfer Loss (DTL) model is mentioned for the first time in this title. This is rather abrupt. The DTL model should either be motivated earlier on, or this title should be updated. The word “Construction” in the title is also sounds odd. Perhaps something like “Historical Development of Phylogenetic Reconciliation Models” could work.
 * Title of Section 3.5: perhaps replace “Necessity to weight” by “Assigning weights to”.
 * Consider merging Sections 3.5 and 3.7.
 * Title of Section 3.8: perhaps replace “time feasibility” by “temporal feasibility”.
 * Title of Section 4: This title is not informative. Perhaps you mean to say something like “Addressing Additional Practical Considerations”?
 * Section 4 could use some motivating text describing the main challenges that arise in practice when applying DTL reconciliation (as described in detail in Sections 4.1 and 4.2).
 * Section 4.3 could become its own full section.
 * Title of Section 5: Could be shortened and made more informative.
 * Title of Section 5.3.3: Perhaps replace “multi-scale” by “multi-level” for consistency? Also, the title mentions Figure 4J, but there isn’t any part J in Figure 4.
 * Title of Section 5.3.4: This title is not very informative.

Figures
Some of the figures are confusing and could be improved. Specific suggestions follow:


 * Figure 2: The left third of the figure, showing the progressive zoom, is confusing. It could help to describe what exactly is being shown in this progressive zoom in the caption.
 * Figure 3: I know this figure is meant to be read alongside the text in Section 3, but clarity could be improved by providing a more detailed caption that summarizes what is shown in each panel.
 * Figure 4: Same suggestion as for Figure 3.

Other Relevant Research
While this article does a good job of reviewing prior work on reconciliation, I thought some very relevant citations were missing. A few such citations are listed below, along with the subsections/themes for which they may be relevant. Embarrassingly, some of these citations are to my own previous work. I encourage the authors to examine the papers below and determine independently which of them are worth citing and discussing in this article.


 * Table 1: https://doi.org/10.1186/1471-2105-11-S1-S60
 * Section 3.9: https://doi.org/10.1093/bioinformatics/btw105, https://doi.org/10.3390/a14080231
 * Section 4.1: https://doi.org/10.1089/cmb.2013.0073, https://doi.org/10.1142/S0219720012500254
 * Section 4.2.2: https://doi.org/10.1093/sysbio/syu007, https://doi.org/10.1109/TCBB.2017.2710342, https://doi.org/10.1007/978-3-030-42266-0_10, https://doi.org/10.1007/978-3-319-05269-4_33 and related previous work.
 * Section 4.2.3: https://doi.org/10.1006/mpev.1996.0071, https://doi.org/10.1080/10635150500354928, https://doi.org/10.1093/bioinformatics/14.9.819, https://doi.org/10.1006/mpev.1999.0676, https://doi.org/10.1093/bioinformatics/btn230, https://doi.org/10.1186/1471-2105-11-574, https://doi.org/10.1109/TCBB.2010.74, https://doi.org/10.1186/1748-7188-8-28, https://doi.org/10.1186/1471-2105-12-S1-S14
 * Section 4.3.1: https://doi.org/10.1145/3307339.3342168
 * Section 5.1.1: This section could use more references related to the holobiont concept.
 * Section 5.2.1: https://doi.org/10.1006/mpev.1996.0071, https://doi.org/10.1142/9789812799623_0050, https://dl.acm.org/doi/10.5555/646341.686721, https://doi.org/10.1089/cmb.2009.0139, https://doi.org/10.1093/bioinformatics/btn150, https://doi.org/10.1093/molbev/msab043

Mukul (talk) 20:08, 8 November 2021 (PST)