Author response:
The following is the authors’ response to the original reviews.
Reviewer #1 (Public review):
Summary:
The manuscript "Lifestyles shape genome size and gene content in fungal pathogens" by Fijarczyk et al. presents a comprehensive analysis of a large dataset of fungal genomes to investigate what genomic features correlate with pathogenicity and insect associations. The authors focus on a single class of fungi, due to the diversity of lifestyles and availability of genomes. They analyze a set of 12 genomic features for correlations with either pathogenicity or insect association and find that, contrary to previous assertions, repeat content does not associate with pathogenicity. They discover that the number of proteincoding genes, including the total size of non-repetitive DNA does correlate with pathogenicity. However, unique features are associated with insect associations. This work represents an important contribution to the attempts to understand what features of genomic architecture impact the evolution of pathogenicity in fungi.
Strengths:
The statistical methods appear to be properly employed and analyses thoroughly conducted. The manuscript is well written and the information, while dense, is generally presented in a clear manner.
Weaknesses:
My main concerns all involve the genomic data, how they were annotated, and the biases this could impart to the downstream analyses. The three main features I'm concerned with are sequencing technology, gene annotation, and repeat annotation.
We thank the reviewer for all the comments. We are aware that the genome assemblies are of heterogeneous quality since they come from many sources. The goal of this study was to make the best use of the existing assemblies, with the assumption that noise introduced by the heterogeneity of sequencing methods should be overcome by the robustness of evolutionary trends and the breadth and number of analyzed assemblies. Therefore, at worst, we would expect a decrease in the power to detect existing trends. It is important to note that the only way to confidently remove all potential biases would be to sequence and analyze all species in the same way; this would require a complete study and is beyond the scope of the work presented here. Nevertheless some biases could affect the results in a negative way, eg. is if they affect fungal lifestyles differently. We therefore made an attempt to explore the impact of sequencing technology, gene and repeat annotation approach among genomes of different fungal lifestyles. Details are described in Supplementary Results and below. Overall, even though the assembly size and annotations conducted with Augustus can sometimes vary compared to annotations from other resources, such as JGI Mycocosm, we do not observe a bias associated with fungal lifestyles. Comparison of annotations conducted with Augustus and JGI Mycocosm dataset revealed variation in gene-related features that reflect biological differences rather than issues with annotation.
The collection of genomes is diverse and includes assemblies generated from multiple sequencing technologies including both short- and long-read technologies. Not only has the impact of the sequencing method not been evaluated, but the technology is not even listed in Table S1. From the number of scaffolds it is clear that the quality of the assemblies varies dramatically. This is going to impact many of the values important for this study, including genome size, repeat content, and gene number.
We have now added sequencing technology in Table S1 as it was reported in NCBI. We evaluated the impact of long-read (Nanopore, PacBio, Sanger) vs short-read assemblies in Supplementary Results. In short, the proportion of different lifestyles (pathogenic vs. nonpathogenic, IA vs non-IA) were the same for short- and long-read assemblies. Indeed, longread assemblies were longer, had a higher fraction of repeats and less genes on average, but the differences between pathogenic vs. non-pathogenic (or IA vs non-IA) species were in the same direction for two sequencing technologies and in line with our results. There were some discrepancies, eg. mean intron length was longer for pathogens with long-read assemblies, but slightly shorter on average for short-read assemblies (and to lesser extent GC and pseudo tRNA count), which could explain weaker or mixed results in our study for these features.
Additionally, since some filtering was employed for small contigs, this could also bias the results.
The reason behind setting the lower contig length threshold was the fact that assemblies submitted to NCBI have varying lower-length thresholds. This is because assemblers do not output contigs above a certain length, and this threshold can be manipulated by the user. Setting a common min contig length was meant to remove this variation, knowing that any length cut-off will have a larger effect on short-read based assemblies than long-read-based assemblies. Notably, genome assemblies of corresponding species in JGI Mycocosm have a minimum contig length of 865 bp, not much lower than in our dataset. Importantly, in a response to a comment of previous reviewer, repeat content was recalculated on raw assembly lengths instead of on filtered assembly length.
I have considerable worries that the gene annotation methods could impart biases that significantly affect the main conclusions. Only 5 reference training sets were used for the Sordariomycetes and these are unequally distributed across the phylogeny. Augusts obviously performed less than ideally, as the authors reported that it under-annotated the genomes by 10%. I suspect it will have performed worse with increasing phylogenetic distance from the reference genomes. None of the species used for training were insectassociated, except for those generated by the authors for this study. As this feature was used to split the data it could impact the results. Some major results rely explicitly on having good gene annotations, like exon length, adding to these concerns. Looking manually at Table S1 at Ophiostoma, it does seem to be a general trend that the genomes annotated with Magnaporthe grisea have shorter exons than those annotated with H294. I also wonder if many of the trends evident in Figure 5 are also the result of these biases. Clades H1 and G each contain a species used in the training and have an increase in genes for example.
We have applied 6 different reference training sets (instead of one) precisely to address the problem of increasing phylogenetic distance of annotated species. To further investigate the impact of chosen species for training, we plotted five gene features (number of genes, number of introns, intron length, exon length, fraction of genes with introns) as a function of branch length distance from the species (or genus) used as a training set for annotation. We don’t see systematic biases across different training sets. However, trends are very clear for clades annotated with fusarium. This set of species includes Hypocreales and Microascales, which is indeed unfortunate since Microascales is an IA group and at the same time the most distant from the fusarium genus in this set. To clarify if this trend is related to annotation bias or a biological trend, we compared gene annotations with those of Mycocosm, between Hypocreales Fusarium species, Hypocreales non-Fusarium species, and Microascales, and we observe exactly the same trends in all gene features.
Similarly, among species that were annotated with magnaporthe_grisea, Ophiostomatales (another IA group) are among the most distant from the training set species. Here, however, another order, Diaporthales, is similarly distant, yet the two orders display different feature ranges. In terms of exon length, top 2 species in this training set include Ophiostoma, and they reach similar exon length as the Ophiostoma species annotated using H294 as a training set. In summary, it is possible that the choice of annotation species has some effect on feature values; however, in this dataset, these biases are likely mitigated by biological differences among lifestyles and clades.
Unfortunately, the genomes available from NCBI will vary greatly in the quality of their repeat masking. While some will have been masked using custom libraries generated with software like Repeatmodeler, others will probably have been masked with public databases like repbase. As public databases are again biased towards certain species (Fusarium is well represented in repbase for example), this could have significant impacts on estimating repeat content. Additionally, even custom libraries can be problematic as some software (like RepeatModeler) will include multicopy host genes leading to bona fide genes being masked if proper filtering is not employed. A more consistent repeat masking pipeline would add to the robustness of the conclusions.
We have searched for the same species in JGI Mycocosm and were able to retrieve 58 genome assemblies with matching species, with 19 of them belonging to the same strain as in our dataset. Overall we found no differences in genome assembly length. Interestingly, repeat content was slightly higher for NCBI genome assemblies compared to JGI Mycocosm assemblies, perhaps due to masking of host multicopy genes, as the reviewer mentioned. By comparing pathogenic and non-pathogenic species for the same 19 strains, we observe that JGI Mycocosm annotates fewer repeats in pathogenic species than Augustus annotations (but trends are similar when taking into account 58 matching species). Given a small number of samples, it is hard to draw any strong conclusions; however, the differences that we see are in favor of our general results showing no (or negative) correlation of repeat content with pathogenicity.
To a lesser degree, I wonder what impact the use of representative genomes for a species has on the analyses. Some species vary greatly in genome size, repeat content, and architecture among strains. I understand that it is difficult to address in this type of analysis, but it could be discussed.
In our case the use of protein sequences could underestimate divergence between closely related strains from the same species. We also excluded strains of the same species to avoid overrepresentation of closely related strains with similar lifestyle traits. We agree that some changes in the genome architecture can occur very rapidly, even at the species level, though analyzing emergence of eg. pathogenicity at the population level would require a slightly different approach which accounts for population-level processes.
Reviewer #2 (Public review):
Summary:
In this paper, the authors report on the genomic correlates of the transition to the pathogenic lifestyle in Sordariomycetes. The pathogenic lifestyle was found to be better explained by the number of genes, and in particular effectors and tRNAs, but this was modulated by the type of interacting host (insect or not insect) and the ability to be vectored by insects.
Strengths:
The main strength of this study lies in the size of the dataset, and the potentially high number of lifestyle transitions in Sordariomycetes.
Weaknesses:
The main strength of the study is not the clarity of the conclusions.
(1) This is due firstly to the presentation of the hypotheses. The introduction is poorly structured and contradictory in some places. It is also incomplete since, for example, fungusinsect associations are not mentioned in the introduction even though they are explicitly considered in the analyses.
We thank the reviewer for pointing this out. We strived to address all comments and suggestions of the reviewer to clarify the message and remove the contradictions. We also added information about why we included insect-association trait in our analysis.
(2) The lack of clarity also stems from certain biases that are challenging to control in microbial comparative genomics. Indeed, defining lifestyles is complicated because many fungi exhibit different lifestyles throughout their life cycles (for instance, symbiotic phases interspersed with saprotrophic phases). In numerous fungi, the lifestyle referenced in the literature is merely the sampling substrate (such as wood or dung), which doesn't mean that this substrate is a crucial aspect of the life cycle. This issue is discussed by the authors, but they do not eliminate the underlying uncertainties.
We agree with the reviewer that lack of certainty in the lifestyle or range of possible lifestyles of studied species is a weakness in this analysis. We are limited by the information available in the literature. We hope that our study will increase interest in collecting such data in the future.
Reviewer #3 (Public review):
Summary:
This important study combines comparative genomics with other validation methods to identify the factors that mediate genome size evolution in Sordariomycetes fungi and their relationship with lifestyle. The study provides insights into genome architecture traits in this Ascomycete group, finding that, rather than transposons, the size of their genomes is often influenced by gene gain and loss. With an excellent dataset and robust statistical support, this work contributes valuable insights into genome size evolution in Sordariomycetes, a topic of interest to both the biological and bioinformatics communities.
Strengths:
This study is complete and well-structured.
Bioinformatics analysis is always backed by good sampling and statistical methods. Also, the graphic part is intuitive and complementary to the text.
Weaknesses:
The work is great in general, I just had issues with the Figure 1B interpretation.
I struggled a bit to find the correspondence between this sentence: "Most genomic features were correlated with genome size and with each other, with the strongest positive correlation observed between the size of the assembly excluding repeats and the number of genes (Figure 1B)." and the Figure 1B. Perhaps highlighting the key p values in the figure could help.
We thank the reviewer for pointing out this sentence. Perhaps the misunderstanding comes from the fact that in this sentence one variable is missing. The correct version should be “Most genomic features were correlated with genome size and with each other, with the strongest positive correlation observed between the genome size, the genome size excluding repeats and the number of genes (Figure 1B)”. Also, the variable names now correspond better to those shown on the figure.
Reviewer #1 (Recommendations for the authors):
The authors have clearly done a lot of good work, and I think this study is worthwhile. I understand that my concerns about the underlying data could necessitate rerunning the entire analysis with better gene models, but there may be another option. JGI has a fairly standard pipeline for gene and repeat annotation. Their gene predictions are based on RNA data from the sequenced strain and should be quite good in general. One could either compare the annotations from this manuscript to those in mycocosm for genomes that are identical and see if there are systematic biases, or rerun some analyses on a subset of genomes from mycocosm. Indeed, it's possible that the large dataset used here compensates for the above concerns, but without some attempt to evaluate these issues, it's difficult to have confidence in the results.
We very appreciate the positive reception of our manuscript. Following the reviewer’s comments we have investigated gene annotations in comparison with those of JGI Mycocosm, even though only 58 species were matching and only 19 of them were from the same strain. This dataset is not representative of the Sordariomycetes diversity (most species come from one clade), therefore will not reflect the results we obtained in this study. To note, the reason for not choosing JGI Mycocosm in the first place, was the poor representation of the insect-associated species, which we found key in this study. In general, we found that assembly lengths were nearly identical, number of genes was higher, and the repeat content was lower for the JGI Mycocosm dataset. When comparing different lifestyles (in particular pathogens vs. non-pathogens), we found the same differences for our and JGI Mycocosm annotations, with one exception being the repeat content. In the small subset (19 same-strain assemblies), our dataset showed the same level of repeats between the two lifestyles, whereas JGI Mycocosm showed lower repeat content for pathogens (but notably for all 58 species, the trend was same for our and JGI Mycocosm annotations). None of these observations are in conflict with our results where we find no or negative association of repeat content with pathogens.
The figures are very information-dense. While I accept that this is somewhat of a necessity for presenting this type of study, if the authors could summarize the important information in easier-to-interpret plots, that could help improve readability.
We put a lot of effort into showing these complicated results in as approachable manner as possible. Given that other reviewers find them intuitive we decided to keep most of them as they are. To add more clarification, we added one supplementary figure showing distributions of genomic traits across lifestyles. Moreover, in Figure 5, a phylogenetic tree was added with position of selected clades, as well as a scatterplot showing distributions of mean values for genome size and number of genes for those clades. If the reviewer has any specific suggestions on what to improve and in which figure, we’re happy to consider it.
Reviewer #2 (Recommendations for the authors):
I have no major comments on the analyses, which have already been extensively revised. My major criticism is the presentation of the background, which is very insufficient to understand the importance or relevance of the results presented fully.
Lines are not numbered, unfortunately, which will not help the reading of my review.
(1) The introduction could better present the background and hypotheses:
(a) After reading the introduction, I still didn't have a clear understanding of the specific 'genome features' the study focuses on. The introduction fails to clearly outline the current knowledge about the genetic basis of the pathogenic lifestyle: What is known, what remains unknown, what constitutes a correlation, and what has been demonstrated? This lack of clarity makes reading difficult.
We thank the reviewer for pointing this out. We have now included in the introduction a list of genomic traits we focus on. We also tried to be more precise about demonstrated pathogenic traits and other correlated traits in the introduction.
(b) Page 3. « Various features of the genome have been implicated in the evolution of the pathogenic lifestyle. » The cited studies did not genuinely link genome features to lifestyle, so the authors can't use « implicated in » - correlation does not imply causation.
This sentence also somehow contradicts the one at the end of the paragraph: « we still have limited knowledge of which genomic features are specific to pathogenic lifestyle
We thank the reviewer for this comment. We added a phrase “correlated with or implicated in” and changed the last sentence of the paragraph into “Yet we still have limited knowledge of how important and frequent different genomic processes are in the evolution of pathogenicity across phylogenetically distinct groups of fungi and whether we can use genomic signatures left by some of these processes as predictors of pathogenic state.”.
(c) Page 3: « Fungal pathogen genomes, and in particular fungal plant pathogen genomes have been often linked to large sizes with expansions of TEs, and a unique presence of a compartmentalized genome with fast and slow evolving regions or chromosomes » Do the authors really need to say « often »? Do they really know how often?
We removed “often”.
(d) Such accessory genomic compartments were shown to facilitate the fast evolution of effectors (Dong, Raffaele, and Kamoun 2015) ». The cited paper doesn't « show » that genomic compartments facilitate the fast evolution of effectors. It's just an observation that there might be a correlation. It's an opinion piece, not a research manuscript.
We changed the sentence to “Such accessory genomic compartments could facilitate the fast evolution of effectors”.
(e) even though such architecture can facilitate pathogen evolution, it is currently recognized more as a side effect of a species evolutionary history rather than a pathogenicity related trait ». This sentence somehow contradicts the following one: « Such accessory genomic compartments were shown to facilitate the fast evolution of effectors".
Here we wanted to point out that even though accessory genome compartments and TE expansions can facilitate pathogen evolution the origin of such architecture is not linked to pathogenicity. We reformulated the sentence to “Even though such architecture can facilitate pathogen evolution, it is currently recognized that its origin is more likely a side effect of a species evolutionary history rather than being caused by pathogenicity”.
(f) As the number of genes is strongly correlated with fungal genome size (Stajich 2017), such expansions could be a major contributor to fungal genome size. » This sentence suggests that pathogens might have bigger genomes because they have more effectors. This is contradictory to the sentence right after « At the end of the spectrum are the endoparasites Microsporidia, which have among the smallest known fungal genomes ».
The authors state that pathogens have bigger genomes and then they take an example of a pathogen that has a minimal genome. I know it's probably because they lost genes following the transition to endoparasitism and not related to their capacity to cause disease. I just want to point out that their writing could be more precise. I invite authors to think of young scholars who are new to the field of fungal evolutionary genomics.
We thank the reviewer for prompting us to clarify the text. We rewrote this short extract as follows “Notably, not all pathogenic species experience genome or gene expansions, or show compartmentalized genome architecture. While gene family expansions are important for some pathogens, the contrary can be observed in others, such as Microsporidia. Due to transition to obligatory intracellular lifestyle these fungi show signatures of strong genome contractions and reduced gene repertoire (Katinka et al. 2001) without compromising their ability to induce disease in the host. This raises questions about universal genomic mechanisms of transition to pathogenic state.”
(g) I find it strange that the authors do not cite - and do not present the major results of two other studies that use the same type of approach and ask the same type of question in Sordariomycetes, although not focusing on pathogenicity:
Hensen et al.: https://pubmed.ncbi.nlm.nih.gov/37820761/
Shen et al.: https://pubmed.ncbi.nlm.nih.gov/33148650/
We thank the reviewer for pointing out this omission. We now added more information in the introduction to highlight the importance of the phylogenetic context in studying genome evolution as demonstrated by these studies. The following part was added to introduction: “Other phylogenomic studies investigating a wide range of Ascomycete species, while not explicitly focusing on the neutral evolution hypothesis, have found strong phylogenetic signals in genome evolution, reflected in distinct genome characteristics (e.g., genome size, gene number, intron number, repeat content) across lineages or families (Shen et al. 2020; Hensen et al. 2023). Variation in genome size has been shown to correlate with the activity of the repeat-induced point mutation (RIP) mechanism (Hensen et al. 2023; Badet and Croll 2025), by which repeated DNA is targeted and mutated. RIP can potentially lead to a slower rate of emergence of new genes via duplication (Galagan et al. 2003), and hinder TE proliferation limiting genome size expansion (Badet and Croll 2025). Variation in genome dynamics across lineages has also been suggested to result from environmental context and lifestyle strategies (Shen et al. 2020), with Saccharomycotina yeast fungi showing reductive genome evolution and Pezizomycotina filamentous fungi exhibiting frequent gene family expansions. Given the strong impact of phylogenetic membership, demographic history (Ne) and host-specific adaptations of pathogens on their genomes, we reasoned that further examination of genomic sequences in groups of species with various lifestyles can generate predictions regarding the architecture of pathogenic genomes.”
(h) Genome defense mechanisms against repeated elements, such as RIP, are not mentioned while they could have a major impact on genome size (Hensen et al cited above; Badet and Croll https://www.biorxiv.org/content/10.1101/2025.01.10.632494v1.full).
This citation is added in the text above.
(i) Should the reader assume that the genome features to be examined are those mentioned in the first paragraph or those in the penultimate one?
In the last paragraph of the introduction we included the complete list of investigated genomic traits.
(j) The insect-associated lifestyle is mentioned only in the research questions on page 4, but not earlier in the introduction. Why should we care about insect-associated fungi?
We apologize for this omission. We added a sentence explaining how neutral evolution hypotheses can explain patterns of genome evolution in endoparasites and species with specialized vectors (traits present in insect-associated species) and added a sentence in the last paragraph that this is the reason why we have selected this trait for analysis.
(2) Why use concatenation to infer phylogeny?
(a) Kapli et al. https://pubmed.ncbi.nlm.nih.gov/32424311/ « Analyses of both simulated and empirical data suggest that full likelihood methods are superior to the approximate coalescent methods and to concatenation »
(b) It also seems that a homogeneous model was used, and not a partitioned model, while the latter are more powerful. Why?
We thank the reviewer for the comment. When we were reconstructing the phylogenetic tree we were not aware of the publication and we followed common practices from literature for phylogenetic tree reconstruction even though currently they are not regarded as most optimal. In fact, in the first round of submission, we have included both concatenation as well as a multispecies coalescent method based on 1000 busco sequences and a concatenation method with different partitions for 250 busco sequences. All three methods produced similar topologies. Since the results were concordant, we chose to omit these analyses from the manuscript to streamline the presentation and focus on the most important results.
(3) Other comments:
Is there a table listing lifestyles?
Yes, lifestyles (pathogenicity and insect-association) are listed in Supplementary Table S1.
(4) Summary:
(a) seemingly similar pathogens »: meaning unclear; on what basis are they similar? why « seemingly »?
We removed “seemingly” from the sentence.
(b) Page 4: what's the difference between genome feature and genome trait?
There is no difference. We apologize for the confusion. We changed “feature” to “trait” whenever it refers to the specific 13 genomic traits analyzed in this study.
(c) Page 22: Braker, not Breaker
corrected
What do the authors mean when they write that genes were predicted with Augustus and Braker? Do they mean that the two sets of gene models were combined? Gene counts are based on Augustus (P24): why not Braker?
We only meant here that gene annotation was performed using Braker pipeline, which uses a particular version of Augustus. We corrected the sentence.
(d) Figure 2B and 2C:
'Undetermined sign' or 'Positive/Negative' would be better than « YES » or it's just impossible to understand the figure without reading the legend.
We changed “YES” to “UNDETERMINED SIGN” as suggested by the reviewer.
