eLife文章，今天终于尘埃落定，编辑让我们VOR了。回想起这篇文章，光投稿就折腾了一年时间。从New Phytologist（送审，一个大修改，一个拒稿）， MBE（送审), JIPB, Plant Comm。最后，实在不想折腾了，并想体验一下eLife的新模式，就投了eLife。两周后，送审，但修改，就来来回回修改了4次。下面把每个期刊的意见粘出来，供朋友们参考：

New Phytologist：

NPH-MS-2023-42523 Positive selection contributes to rapid evolution of male-biased genes in dioecious Trichosanthes pilosa (Cucurbitaceae)by Zhao, Lei; Zhou, Wei; He, Jun; Li, De-Zhu; Li, Hong-Tao

I have received two reviews of your manuscript and have read it carefully myself.  Unfortunately, neither reviewer was enthusiastic about your manuscript, with one recommending "reject" and the other "major revision".  Reviewer 1 identifies a number of deficiencies, including lack of statement of specific research goals or hypotheses, improper tissue sampling, and improper interpretation of accelerated evolutionary rates.  Reviewer 2 indicates that you do not compare your results to previous studies and do not clarify what your novel results are, if any.  Because I agree with the reviewers' comments, I am sorry to inform you that I am declining to publish your manuscript.I regret I cannot bring you better news.

.Decision: Reject

Referee: 1 

This is a descriptive study of gene expression between the sexes in a dioecious plant in the family Cucurbitaceae that presumably has genetic sex determination (though the study  species is said only to have “young sex chromosomes”). There are no clear research goals or questions (apart from gaining “insights into the evolution of sexual dimorphism”). It is, however, correct that there are currently few studies in plants of expression patterns and sequence evolution of sex-biased genes. The species displays some sexually dimorphic traits but they are not clearly described (the text says “racemose, early-flowering and caducous males, but solitary, late-flowering and long-lived females”, but there are no pictures to explain what these look like — Figure 1 shows inflorescences but doesn’t relate the pictures to the text).Three female and three male Trichosanthes pilosa plants were studied. As in previous studies, sex differences in gene expression are most pronounced in flower buds (it thus seems strange to write that this result is intriguing, line 330). It is not clear which “sex-biased” genes are actually expressed only in buds or flowers of one sex (in other words, are sex-specific). It would be better to distinguish genes with sex-specific expression from ones with sex bias. Sex-specific genes may reflect genes expressed strictly in sex-specific organs, such as tapetum. The study compares these with mature flowers, whereas previous work often compared bud and leaf transcriptomes. The kinds of genes detected by the KEGG analysis of the biased gene sets are similar to what is often found in similar tissues.It is not clear what the difference between flower buds and mature flowers tells us.The ms also describes tests for rapid evolution of the sequences of some genes (presumably sex biased ones) and attempts to attribute evolutionary forces that might be responsible. Specifically, observed higher rates of sequence evolution were found for male-biased genes in buds compared to female-biased or unbiased genes. The conclusion (in my wording) that this “could result from either positive or sexual selection acting on beneficial alleles which may function in response to abiotic and biotic stress and/or male-male competition” is not a real conclusion, as these include most selective possibilities, and rapid evolution can also reflect loss of selective constraints (as the text mentions as bit later). In other words, writing this merely repeats that some possible sign of selection or neutral evolution has been seen. More thought is needed about the results before it would be justified to claim that the “results illuminate the molecular evolution of male-biased genes and potential selective forces during early reproductive development and significantly advance the understanding of the patterns and forces driving the evolution of sexual dimorphism in dioecious plants”.The English is poor and needs serious work by a native English speaker who understands the science. The writing can be considerably shortened to omit unnecessary repeats (such as that the results provide “insights”), and stick to the important points. The present version is rambling and derivative, appearing to repeat what has been said in previously published papers, whether about animals or plants. Although the evolution of sex differences in animals and plants may share interesting characteristics, the ms does not mention anything specific, and reads as if concepts from one kind of organism (e.g. rapid evolution of some genes) should apply to the other, without explaining any reason for this belief, or why some genes might be expected to evolve faster than others (e.g. because they function in “arms race” situations), and whether any sub-sets of the genes studied might belong to such a category. The Introduction can easily be cut to about half its present length, and be made clearer. The text starting in line 114 can be greatly shortened, as it merely repeats the basic possibilities mentioned above, but fails to make clear why sex-biased genes might differ from any other kind of genes. The examples from animals that are quoted are not helpful in making this clear, and should be omitted, and even the brown alga example is not described in a way that relates the results to sex bias, and the examples from plants in the next paragraph (line 129) is also not related to the possible causes of selection. Another general problem with the text is very vague writing where it should be explicit. For example “young sex chromosomes” is too vague. The age should be described with some actual information, ideally Y-X sequence divergence for synonymous sites across a specified number of genes (with some error estimate). Similarly, it is not clear enough to write that “Most dioecious plants do not possess sex chromosomes”. I think the authors mean “heteromorphic sex chromosomes”, but around half of dioecious plants so far studied do have some sex chromosome heteromorphism, so, while the statement is formally correct, it does not give an accurate idea of what is currently understood.The approach used was RNASeq followed by de novo assembly. The genomic locations of the sequences are therefore unknown, precluding study of questions that might be interesting, such as whether the sex-linked region carries a higher density of genes with sex biased expression than other genome regions. It is not very clear whether the “unigenes” are complete genes, and whether they are single-copy genes, though BUSCO analysis suggests that many are not (line 317 mentions that 895 genes were assessed as ‘complete’ and 239 as ‘fragmented’, so the total number of complete genes is small). One-to-one orthologous groups (OGs) were inferred using OrthoFinder analysis of several other Trichosanthes species and another, presumably Cucurbit, species L. cylindrica, whose genus name is not given, and its relationship is not explained (Momordica charantia, Cucumis melo, and several Cucurbita species are mentioned in the Results section, similarly without any of the information needed). The Methods section does not make clear whether partial sequences were included in the PAML analyses of synonymous and non-synonymous substitutions. The genome sizes are not mentioned, but I believe that some of these plants have large genomes, and determining orthology without being able to make use of synteny might be problematic, and/or non-functional duplicate copies (for example in pericentromeric regions) may be hard to distinguish from functional copies.PAML-generated median ω estimates were just under 20% higher for male- than female-biased genes, but female-biased and unbiased genes did not differ.  Two of 343 genes showing male-bias in floral buds had values > 1. Tests such as MacDonald-Kreitman tests need to be applied before any conclusions are possible about whether any set of genes, e.g. those with male-bias, show more signals of positive selection compared with others, or might be under relaxed purifying selection. The authors should look at other similar attempts to test such ideas, for example Obbard et al. (2006). Natural selection drives extremely rapid evolution in antiviral RNAi genes. Current Biology, 16(6), 580-585.  Possible contamination with non-functional copies also needs to be considered, including in relation to the codon usage analyses. These analyses also fail to consider the possibility that genes may be located in different genome regions, and these could differ in GC content. Line 444 states that male-biased genes had much higher median dS values in than female-biased or unbiased genes in some of the tests, and this is not explained, but ought to be understood before tests for selection can be applied. One possibility is that high dS values reflect undetected gene duplications in the not very recent past.The Discussion is also much too long, given that few, if any, firm conclusions emerge. It the text that repeats parts of the Introduction were removed, plus repeats of the results (which have just been described and are also in the Summary and repeated in the Conclusions, which is unnecessary), the section contains nothing much about what the study tells us about any interesting question.

Referee: 2 

 general comments:The article is interesting but hard to read primarily because of language issues. Some methodological problems must be addressed in a revised version, which seems possible. Most importantly, however, the discussion is deficient in that it does not clarify how the new results compare to previous studies of evolutionary rates for sex-biased genes in dioecious plants. What is novel, what has been seen before? Detailed comments:35    empirical instead of experimental. Experimental suggests manipulation, whereas the present study is a survey/obsvervational.36    Contrary to this statement, male and female are compared, not flower buds and mature flowers.38    "evidence" instead of "evidences"43f.    The most cautious explanation for faster sequence evolution, namely relaxed purifying selection, must not be left out.43f.    Opposing positive with sexual selection is a nonsensical statement; the direction of selection (positive/negative) is not related to the driver of selection (sexual/natural/ etc).Intro64/65 trivial statement: of course SBGs are located on autosomes or sex chromosomes; no other significant genomic regions exist (organelles?)66: insinuates a generalisation which is not true: many SBGs may not be involved in development of SD traits. Also, gene expression is itself a trait..70: "gene expressions" is an incorrect pluralization74 ff: False statement. Hermaphroditic plants do not necessarily have bisexual flowers, see moneocy. Thus the false dichotomy: hermaphroditic flowers vs. dioecy. Better: bi-sexual versus uni-sexual individuals.77: This statement is unfounded. Instead, there is a lack of data on potential sex chromosomes in most dioecicous plants.77f: language.80ff: I disagree; there are plenty of other approaches suitable for the study of selection on sexual dimorphism in dioecious plants.88: "a wider range"90f: language99: proteins100f: confusing/confused/trivial statement: "Expression of sex-biased genes differed from expression of unbiased genes."106: confusion of subject/object of the sentence.107: Wrong. In Scharmann et al (2020), data on 10 species of Angiosperms is presented, + many further studies exist.107ff: Unclear language and logic. Why are these types of comparisons different and why they are relevant?114ff: Incoherent language and logic. For example: "adaptive fixation of beneficial alleles induces positive selection". It could work they other way round..123ff: Structure. Revisiting the same point as before, but now concluding something else.131: comparedOK I'll stop here giving detailed comments, there are too many language problems. Focussing on logic/technical now.146: relaxazation is not a real word165f: Reference Ming et al. 2011 does in fact not contain information on Trichosanthes pilosa, but lists three references for the cytogenetics of some other Trichosanthes spp. I doubt the adjective "young" can meanignfully be applied in this case; there is no data on the age.# Materials & Methods184f: I think the naming schemes of the samples during analyses are not at all relevant for the readers.198: Was assembly done with reads from all samples, or at least from both sexes? If not, there could be a bias against genes expressed in only one sex.203: Which BUSCO set was this?Plants growing in open environments generally have microorganisms on their surface and perhaps inside of their tissues. Which measures were taken to exclude possible contaminants (micro-organisms) from the read data and/or the expression quantification?212ff: Was a threshold used to exclude genes that had no expression in any sample? (in most datasets there are genes with zero or nearly-zero read counts, which could be artefacts). Excluding these would be important to evaluate the proportion of all expressed genes that are sex-biased.216/217: failed to specify which categories were tested against another? I guess male against female separately for each developmental stage?217: P-values were adjusted how?218f: Please specifiy how the "background" for the enrichment tests was set up: this should be the annotations (GOs / KEGGs) for all genes used in differential expression tests. Enrichment results are erroneous if an unspecific background is used.235: Which tool was used for alignment, and were any measures taken to avoid false alignments / filter against poorly aligned regions?248: Does conflation of the two dioecious species bias the results? Omega is estimated for the dioecious clade (T. pilosa, T. anguina) and then compared between the genes categorised as per sex-biased expression in T. pilosa. How great is the risk that omega is substantially affected by the sequence of T. anguina, for which the sex-bias status is actually unknown?253: Unclear why Benjamini-Hochberg procedure is useful here. Multiple testing is not rampant with just the three pairwise comparisons!?224ff: section "Evolutionary rate analyses". Parts of this section are astonishingly similar in structure and phrasing, including some mistakes ("codon bias usage" instead of "codon usage bias", "CondonFreq" instead of "CodonFreq") and included references, to the paper Catalan et al (2018). The source is cited, though.246: "codon bias usage" -> "codon usage bias".    reading up: What is the idea behind analyses of codon bias?    - codon bias = preferential use of some codons over their synonyms.    - "codon bias is largely thought to be due to weak selection favoring the use of codons that are most efficiently and accurately translated (Akashi 1994, 1995; Carlini and Stephan 2003). Selection intensity for codon usage bias, therefore, is expected to vary among genes. Presumably, highly expressed genes have more codon bias because selection for translational efficiency and accuracy is stronger in these genes."            from: "Hambuch TM, Parsch J. Patterns of synonymous codon usage in Drosophila melanogaster genes with sex-biased expression. Genetics. 2005"    - Hambuch & Parsch find that Drosophila male-biased genes have less codon bias than f-biased and unbiased genes, which are similar.    - their interpretation:    together with omega: selective constraint on syn subst is measureable as codon bias; low dS may drive higher omega.## continue without deep understanding of this.256: codon usage bias# Results312: 25118 out of 59051 is 42.4%, not 80%315: rephrase to just report the conventional BUSCO scores; the main score is 78.8% if I understand correctly?321ff: please add the proportion / percentage of sex-biased genes among all expressed genes.334ff, Figure 2 c-d: Why are the median expression values of sex-biased genes relevant at all? What does it reveal?343f: Enrichment of functions of sex-biased genes over ... what? The functions of all expressed genes in the test, or else..? What is the "background" against which these enrichments are found?In general, I find the (four times repeated) statements of structure "... enriched in metabolic and signaling pathways associated with/related to stigma and pistil development, including ..." vague, perhaps even misleading? Is there a term "stigma and pistil development", and is it enriched? If not, then on what is the claimed 'association' between "stigma and pistil development" and the listed KEGG terms made? Is this just 'intuition'?This should be phrased more cautiously.363: Unfounded claim; there is no data for the metabolic profiles of the tissues! Rather, the distinct gene expression profiles allow to speculate that the metabolic profiles are also different..382f: The sentence has a language problem; I do not understand it.472: language problem?# Discussion489: again invoking the false distinction between positive and sexual selection496f: Unclear.500: It is not evident to me from Figure 1 that there the buds are more sexually dimorphic than the mature flowers - please explain. How can the later developmental stage be less sexually dimorphic than its preceding developmental stage?502-508: unclear. What is meant by "narrow expression pattern"?509-518: unclear.526f: Misleading wording: "we found that some male-biased genes ... respond to nutrient deficiency ..." - this study presents no data on the function of genes!531: I do not think this speculation about climate adaptation is very helpful or meaningful, and it is confusingly worded. Rather, evolutionary rates of sex-biased genes could have something to do with sexual reproduction?551: Wording: the genes did almost certainly not undergo adaptation 'in this study'...559: More clearly state this as speculation. The evolutionary changes could as well have other effects and causes.The discussion entirely lacks comparisons of the new Trichosanthes results with those from previous studies of the same scope, i.e. sex-biased gene expression in flowers of other dioecious plants (e.g. Silene, Salix, Rumex, etc)!

MBE:

01-Jun-2023MS: MBE-23-0363Title:  Positive selection and relaxed purifying selection contribute to rapid  evolution of male-biased genes in a dioecious flowering plantThank  you for submitting your manuscript to Molecular Biology and Evolution  (MBE). We regret to inform you that your manuscript did not receive high  enough priority for publication after an in-depth review by the editors  and the peer reviewers. In general, MBE seeks to publish research,  methods, and resources of broad significance in molecular evolutionary  biology. Even when external reviewers find a manuscript to be  scientifically and technically sound, the ultimate priority for  publication is based on the novelty and impact of the work. Specific  comments from the Associate Editor and external reviewers are included  below.While we will not be able to accept your manuscript for  publication in MBE, we would like to offer you the opportunity to  transfer your article to our sister journal, Genome Biology and  Evolution (GBE).  Like MBE, GBE is an open-access journal that  specializes in the dissemination of cutting-edge research in molecular  evolution. We encourage you to transfer your manuscript to GBE and take  advantage of the significant benefits it offers you, including:

• Time-saving submission transfer

• An engaged editorial board of professional scientists

• Constructive and timely review process

• Consideration of your existing MBE reviews

• The potential for expedited publication

• Open-Access

Associate EditorEditors’ comments to the author:Thank  you for submitting your work for consideration. It has now been seen by  two experts in the field, whose comments appear below. Although they  both found the work to be scientifically sound, they also both expressed  concern about the suitability of this work for MBE. My own assessment  is similar to theirs - it seems to be sound work, but not work that  advances the field to the degree expected for work published in MBE.  Thus, while I cannot recommend publication in MBE, I am happy to refer  this work for publication in GBE.

Reviewer: 1

This  is a rather straightforward study of sex-biased gene evolution in a  dioecious plant, Tichosanthes pilosa. The authors are correct that such  studies are far more common in animals, and plants are under-studied in  this regard. Unusually, they include a developmental component,  examining both developing and mature flowers, and this, to my knowledge,  has only been done in animals (e.g. Perry et al. MBE 2014; Parker et  al. Heredity 2022; Mank et al. MBE 2010). In contrast to studies in  animals, which have found that sex-bias increases in both number of  sex-biased genes and in degree of expression difference, the authors  here find the opposite. This is perhaps the most surprising finding of  the study, and should be emphasized more.I have no quibbles with  the analysis or results. I am not sure whether they alone are  sufficiently novel to merit publication in MBE. On one hand plants are  understudied with regard to sex-biased gene evolution, however the study  in its current form provides another species-specific datapoint rather  than a major conceptual advance. This decision I ultimately leave to the  AE.Minor comments:Line 89 – It is important to note that  the delineation of sex-bias was very low in the Cossard and Scharmann  papers – they only used a statistical threshold and did not include a  fold-change threshold. This means that the genes they called  “sex-biased” are often expressed at very similar, although statistically  significant, levels. Given that dN/dS correlates with increased  fold-change (see Harrison et al. PNAS 2015), it is perhaps not terribly  surprising that the low levels of sex-bias in Cossard and Scharmann did  not show elevated rates of evolution.Line 198 – the section on rates  of evolution. It would be very helpful to break down dN/dS rates to  separate differences in rates of dN versus dS. In other words, are  elevated rates of dN/dS due to higher dN, lower dS or both?Line 202 –  I’m not sure that the low numbers of sex-biased genes found in mature  flowers (45 female-biased, 13 male-biased) are sufficient for a robust  test of different rates of evolution. There is quite a lot of variance  in dN/dS estimates and therefore large numbers of genes in each group  are needed.Line 239– “After comparing the A model and null model, we  discovered that 39 out of 343 OGs (11.34%) exhibited strong evidence of  having certain sites that evolved under positive selection based on  foreground 2b ω value, LRT P and BEB value (Fig. 5 and Table S6). As an  additional approach, we utilized the aBSREL and BUSTED methods that were  implemented in HyPhy v.2.5 software, and detected significant evidence  of positive selection.” Even many experts in the field will have a  difficult time navigating the acronyms in these two sentences. Worth  spelling out each one in the first instance

.Reviewer: 2

Zhao  et al. use RNA-seq to investigate sex-biased gene expression in the  dioecious plant Trichosanthes pilosa (Cucurbitaceae). They compare  transcriptomes of female and male floral buds and mature flowers. They  detected a high proportion of sex-biased genes in floral buds (7-10% of  expressed genes), but a much lower proportion in flowers (The  authors present new data and address a relevant question in molecular  evolutionary genomics in a new taxon. Overall, I think the study is well  done and I don't see any flaws in the analysis. My main concern in that  the results are consistent with what has been seen in other taxa,  where, for example, there have been many reports over the past decade  (or more) of rapid evolution of male-biased genes due to positive  selection and/or relaxed purifying selection. While the inclusion of a  dioecious plant broadens this pattern across a wider taxonomic range, it  does not reveal anything that is fundamentally new regarding this  topic. For this reason, I am not convinced that the manuscript reaches  the very high level of significance and broader impact required for  publication in MBE. The data are certainly worth publishing, but might  be better suited to a more specialized journal.The authors don't  provide an explanation for their codon bias results. In fact, they find  that dS is significantly higher in male-biased genes (line 420). This  is difficult to reconcile with the observed greater codon bias in  male-biased genes. Could it be that there is a mutational bias? For  example, if there is a mutational bias towards A or T and male-biased  genes are under less purifying selection at synonymous sites, they will  become more AT-rich at these sites, which could results in greater codon  bias when measured by ENC. To test this, the authors could compare %GC  at synonymous sites of the different groups of genes.

eLife (第一次返回后的意见，其他几次请看   https://elifesciences.org/reviewed-preprints/89941v2

eLife assessment This  valuable paper examines gene expression differences between male and  female individuals over the course of flower development in the  dioecious angiosperm Trichosantes pilosa. The authors show that  male-biased genes evolve faster than female-biased and unbiased genes,  which is frequently observed in animals but this is the first report of  such a pattern in plants. In spite of the limited sample size, the  reviewers found the evidence to be mostly solid and the methods  appropriate for a non-model organism. The resources produced will be  used by researchers working in the Cucurbitaceae, and the results  obtained advance our understanding of the mechanisms of plant sexual  reproduction and its evolutionary implications: as such they will  broadly appeal to evolutionary biologists and plant biologists.

Public Reviews: Reviewer #1 (Public Review): The  evolution of dioecy in angiosperms has significant implications for  plant reproductive efficiency, adaptation, evolutionary potential, and  resilience to environmental changes. Dioecy allows for the  specialization and division of labor between male and female plants,  where each sex can focus on specific aspects of reproduction and  allocate resources accordingly. This division of labor creates an  opportunity for sexual selection to act and can drive the evolution of  sexual dimorphism. In the present study, the authors investigate  sex-biased gene expression patterns in juvenile and mature dioecious  flowers to gain insights into the molecular basis of sexual dimorphism.  They find that a large proportion of the plant transcriptome is  differentially regulated between males and females with the number of  sex-biased genes in floral buds being approximately 15 times higher than  in mature flowers. The functional analysis of sex-biased genes reveals  that chemical defense pathways against herbivores are up-regulated in  the female buds along with genes involved in the acquisition of  resources such as carbon for fruit and seed production, whereas male  buds are enriched in genes related to signaling, inflorescence  development and senescence of male flowers. Furthermore, the authors  implement sophisticated maximum likelihood methods to understand the  forces driving the evolution of sex-biased genes. They highlight the  influence of positive and relaxed purifying selection on the evolution  of male-biased genes, which show significantly higher rates of  non-synonymous to synonymous substitutions than female or unbiased  genes. This is the first report (to my knowledge) highlighting the  occurrence of this pattern in plants. Overall, this study provides  important insights into the genetic basis of sexual dimorphism and the  evolution of reproductive genes in Cucurbitaceae. There are, however, parts of the manuscript that are not clearly described or could be otherwise improved. -  The number of denovo-assembled unigenes seems large and I would like to  know how it compares to the number of genes in other Cucurbitaceae  species. The presence of alternatively assembled isoforms or assembly  artifacts may be still high in the final assembly and inflate the  numbers of identified sex-biased genes. - It is interesting that  the majority of sex-biased genes are present in the floral buds but not  in the mature flowers. I think this pattern could be explored in more  detail, by investigating the expression of male and female sex-biased  genes throughout the flower development in the opposite sex. It is also  not clear how the expression of the sex-biased genes found in the buds  changes when buds and mature flowers are compared within each sex. -  The statistical analysis of evolutionary rates between male-biased,  female-biased, and unbiased genes is performed on samples with very  different numbers of observations, therefore, a permutation test seems  more appropriate here. - The impact of pleiotropy on the  evolutionary rates of male-biased genes is speculative since only two  tissue samples (buds and mature flowers) are used. More tissue types  need to be included to draw any meaningful conclusions here. 

Reviewer #2 (Public Review): Summary: This  study uses transcriptome sequence from a dioecious plant to compare  evolutionary rates between genes with male- and female-biased expression  and distinguish between relaxed selection and positive selection as  causes for more rapid evolution. These questions have been explored in  animals and algae, but few studies have investigated this in dioecious  angiosperms, and none have so far identified faster rates of evolution  in male-biased genes (though see Hough et al. 2014 https://doi.org/10.1073/pnas.1319227111). Strengths: The  methods are appropriate to the questions asked. Both the sample size  and the depth of sequencing are sufficient, and the methods used to  estimate evolutionary rates and the strength of selection are  appropriate. The data presented are consistent with faster evolution of  genes with male-biased expression, due to both positive and relaxed  selection. This is a useful contribution to understanding the  effect of sex-biased expression in genetic evolution in plants. It  demonstrates the range of variation in evolutionary rates and selective  mechanisms, and provides further context to connect these patterns to  potential explanatory factors in plant diversity such as the age of sex  chromosomes and the developmental trajectories of male and female  flowers. Weaknesses: The presence of sex chromosomes is a  potential confounding factor, since there are different evolutionary  expectations for X-linked, Y-linked, and autosomal genes. Attempting to  distinguish transcripts on the sex chromosomes from autosomal  transcripts could provide additional insight into the relative  contributions of positive and relaxed selection. 

Reviewer #3 (Public Review): The  potential for sexual selection and the extent of sexual dimorphism in  gene expression have been studied in great detail in animals, but hardly  examined in plants so far. In this context, the study by Zhao, Zhou et  al. al represents a welcome addition to the literature. Relative  to the previous studies in Angiosperms, the dataset is interesting in  that it focuses on reproductive rather than somatic tissues (which makes  sense to investigate sexual selection), and includes more than a single  developmental stage (buds + mature flowers). The main  limitation of the study is the very low number of samples analyzed, with  only three replicate individuals per sex (i.e. the whole study is built  on six individuals only). This provides low power to detect  differential expression. Along the same line, only three species were  used to evaluate the rates of non-synonymous to synonymous  substitutions, which also represents a very limited dataset, in  particular when trying to fit parameter-rich models such as those  implemented here. A third limitation relates to the absence of a  reference genome for the species, making the use of a de novo  transcriptome assembly necessary, which is likely to lead to a large  number of incorrectly assembled transcripts. Of course, the production  of a reference transcriptome in this non-model species is already a  useful resource, but this point should at least be acknowledged  somewhere in the manuscript. Each of these shortcomings is  relatively important, and together they strongly limit the scope of the  conclusions that can be made, and they should at least be acknowledged  more prominently. The study is valuable in spite of these limitations  and the topic remains grossly understudied, so I think the study will be  of interest to researchers in the field, and hopefully inspire further,  more comprehensive analyses.