Central carbon metabolism switching in lytic versus temperate coral reef viral communities
Curation statements for this article:-
Curated by eLife
eLife Assessment
This study presents an important conceptual advance linking viral lifestyle (temperate versus lytic) to patterns of central carbon metabolism gene content across coral reef viromes, proposing that these associations may reinforce feedback loops influencing reef microbialization state. The revised manuscript better aligns its claims with the underlying data by separating observed metabolic gene associations from the hypothesized downstream effects on infection, host metabolism, and reef state, but the evidence for the central mechanistic claims remains incomplete, as no new analyses were added to establish that the identified metabolic genes are mechanistically linked to temperate viral genomes, to rule out compositional effects, or to validate key gene annotations. This work will be of interest to viral ecologists and to researchers studying coral reef microbial dynamics, providing a foundation for future studies testing these proposed relationships directly.
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
Abstract
Coral reefs are declining globally due in part to bacterial overgrowth, a process known as microbialization. However, the role of bacteriophages that may inhibit microbialization by infecting and killing these bacteria remains poorly understood, especially their metabolic impacts on bacterial proliferation. To address this, we analyzed central carbon metabolism gene frequencies in viral communities from healthy (lytic-dominated) and degraded (temperate-dominated) Central Pacific coral reefs. We found that viral communities shifted broadly from enrichment in genes associated with reactions that build up pools of central intermediates on degraded, temperate-dominated reefs (‘anaplerotic’ reactions) to enrichment in genes associated with reactions that consume these pools to generate metabolic precursors for virion construction on healthy, lytic-dominated reefs (‘cataplerotic’ reactions). This switch was shown by the over-representation of Entner-Doudoroff (ED) glycolysis genes on degraded, temperate-dominated reefs and, on healthy, lytic-dominated reefs, by a cross-pathway enrichment of cataplerotic genes, including transaldolase and NADH-producing oxidoreductases. To explore the ecological implications of these observations, we incorporated them into a conceptual model with coral, algae, bacterial, and viral ‘populations’ and qualitative ‘high’ and ‘low’ rates linking them based on our findings and existing knowledge of healthy and degraded reefs. The model suggests two contrasting ecosystem scenarios: one in which lytic viral metabolism may enhance viral production and bacterial control, and another in which temperate viral metabolism may favor bacterial proliferation. This framework identifies viral metabolism as an important and underexplored dimension of coral reef ecology and a promising avenue for future conservation research.
Article activity feed
-
-
-
eLife Assessment
This study presents an important conceptual advance linking viral lifestyle (temperate versus lytic) to patterns of central carbon metabolism gene content across coral reef viromes, proposing that these associations may reinforce feedback loops influencing reef microbialization state. The revised manuscript better aligns its claims with the underlying data by separating observed metabolic gene associations from the hypothesized downstream effects on infection, host metabolism, and reef state, but the evidence for the central mechanistic claims remains incomplete, as no new analyses were added to establish that the identified metabolic genes are mechanistically linked to temperate viral genomes, to rule out compositional effects, or to validate key gene annotations. This work will be of interest to viral ecologists and …
eLife Assessment
This study presents an important conceptual advance linking viral lifestyle (temperate versus lytic) to patterns of central carbon metabolism gene content across coral reef viromes, proposing that these associations may reinforce feedback loops influencing reef microbialization state. The revised manuscript better aligns its claims with the underlying data by separating observed metabolic gene associations from the hypothesized downstream effects on infection, host metabolism, and reef state, but the evidence for the central mechanistic claims remains incomplete, as no new analyses were added to establish that the identified metabolic genes are mechanistically linked to temperate viral genomes, to rule out compositional effects, or to validate key gene annotations. This work will be of interest to viral ecologists and to researchers studying coral reef microbial dynamics, providing a foundation for future studies testing these proposed relationships directly.
-
Reviewer #2 (Public review):
Summary:
The manuscript by Kelman and coauthors investigates how viral communities differ in the genes they encode in healthy and degraded coral reef ecosystems. Across 19 viral metagenomes from Central Pacific reefs, the authors assess the frequency of integration/excision genes as a proxy for viral community temperateness and ask whether genes associated with central carbon metabolism covary with signatures of temperateness. The main finding is that viral communities with more temperate-related genes encode more genes from the Entner-Doudoroff pathway and other reactions interpreted as anaplerotic, whereas more lytic-associated viral communities show greater representation of some pentose phosphate pathway, TCA, and redox-associated genes interpreted as cataplerotic. The authors propose a model based on …
Reviewer #2 (Public review):
Summary:
The manuscript by Kelman and coauthors investigates how viral communities differ in the genes they encode in healthy and degraded coral reef ecosystems. Across 19 viral metagenomes from Central Pacific reefs, the authors assess the frequency of integration/excision genes as a proxy for viral community temperateness and ask whether genes associated with central carbon metabolism covary with signatures of temperateness. The main finding is that viral communities with more temperate-related genes encode more genes from the Entner-Doudoroff pathway and other reactions interpreted as anaplerotic, whereas more lytic-associated viral communities show greater representation of some pentose phosphate pathway, TCA, and redox-associated genes interpreted as cataplerotic. The authors propose a model based on these patterns in which viruses may reinforce feedback loops that influence reef microbialization state.
In the revised manuscript, the authors have refined the separation between the observations and hypothesized effects on infection, host metabolism, and reef state, which align the claims much closer to what the data support. No new analyses were added to buttress whether the metabolic genes are linked to temperate viral genomes, to address compositional effects, or to validate the key annotations, so the evidence for the central mechanistic claims is largely unchanged.
Overall, the study addresses an interesting and potentially important concept - that viral auxiliary metabolic genes are important components of microbial communities and can affect ecosystem functioning. Linking viral metabolism to coral reef microbialization is a creative conceptual advance. The manuscript is clearly written, and the contrasting metabolic gene profiles of these viromes is an interesting pattern that could motivate future work on how viral metabolic potential varies across reef states.
Strengths:
(1) The study connects viral lifestyle, central carbon metabolism, bacterial overgrowth, and reef degradation in a framework that could be useful for future studies of coral reef ecosystems and viral ecology. This is an interesting synthesis that links viral auxiliary metabolism to broader questions about microbialization and reef state.
(2) The manuscript is generally clearly organized around a testable prediction: viral metabolic gene content should vary along a lytic-to-temperate viral community gradient. The reported enrichment of anaplerotic genes in viromes with a larger fraction of temperate viruses is an intriguing result.
(3) The authors highlight several metabolic genes that may not have been previously reported in viral datasets or genomes. With further validation, these observations could expand the known repertoire of viral metabolic potential.
(4) The modelling helps clarify the feedbacks the authors propose may connect viral lifestyle, bacterial metabolism, and coral reef degradation. It provides a foundation for generating hypotheses about how viral metabolic genes could influence reef microbial dynamics.
(5) The revised manuscript is clearer about scope. The methods section now describes the read-based framework and its limitations, and the model is presented as a tool for generating hypotheses rather than as evidence.
Weaknesses:
(1) The main limitation remains that the evidence for several key claims remains indirect. The core analysis is based on correlations between metabolic gene frequencies and integration/excision-related genes. This does not demonstrate that the metabolic genes occur in temperate viral genomes, are physically linked to lysogeny genes, are expressed during infection, or alter host metabolism. The authors did not to attempt contig-level linkage. Thus, the data support an association between VLP-associated metabolic annotations and a community-level temperateness proxy, but not a direct link between temperate phages and these metabolic functions. The manuscript does not discuss alternative explanations for these correlations. More broadly, community-level gene frequencies should not be equated with genome-level or infection-level metabolic programs. A virome may contain more anaplerotic genes overall, but this does not demonstrate that individual viruses reprogram their hosts in an anaplerotic manner or that infection produces a net anaplerotic effect. Individual viruses may encode both anaplerotic and cataplerotic genes, and a smaller number of cataplerotic genes could have stronger metabolic consequences depending on expression, enzyme efficiency, pathway position, and host context.
(2) The revised framing is more careful, but some language still goes beyond the data. The ecological interpretation assumes that viral infection is strong enough to influence reef-scale bacterial population dynamics. However, the study does not directly measure infection frequency, lysis rates, viral production, burst size, lysogeny frequency, prophage induction, gene expression, or bacterial mortality. If viral mortality or lysogenic conversion were rare in these systems, the observed gene-frequency patterns could have limited ecosystem-level consequence. The discussion includes statements such as healthy reefs being "guarded by lytic viruses" and nutrient interventions could promote "re-viralization" for reef restoration. These are reasonable hypotheses, but they depend on infection rates, gene expression, and metabolic effects that were not measured nor discussed.
(3) Because gene frequencies are expressed as proportions of annotated genes, the large decline in genes assigned to phages (from about half of annotated genes in lytic viromes to about 5% in temperate viromes) can force most other gene categories upward with temperateness. The authors now acknowledge this compositional effect but do not test whether the main associations hold once it is accounted for. The many correlations examined are also not corrected for multiple comparisons. As a result, individual gene-level associations, particularly the temperate-associated ones, should be treated as tentative.
(4) The anaplerotic/cataplerotic classification still relies on sequence similarity annotations, so may be vulnerable to misannotation, or functional misinterpretation (notable examples of definitively annotated, but functionally ambiguous genes such as phoH and cobS have been well documented in viruses). The authors did not to include any further domain-based or phylogenetic checks for the genes important the main conclusions.
Overall, the manuscript presents a valuable hypothesis and highlights new ecological patterns in coral reef viral metagenomes. The revision substantially improves the discussion of the claims, but the evidence linking temperate viruses to specific metabolic functions and reef-scale outcomes remains incomplete. The work would be strengthened by analyses that directly link metabolic genes to viral genomes or lysogeny markers, address compositional effects, validate key annotations, and more clearly distinguish observed gene-frequency associations from hypothesized effects on infection, host metabolism, and reef state.
-
Author response:
The following is the authors’ response to the original reviews.
eLife Assessment:
The worldwide decline in the health of coral reefs is well documented, and overgrowth by microbial consortia can be a contributing factor. Kelman and colleagues used metagenomic analysis to interrogate potential changes in phage-associated genes predicted to be involved in central carbon metabolism. The study addresses the hypothesis that metabolic genes associated with carbon metabolism that are encoded by viruses reflect the health of the corals. The study contributes a valuable perspective on the potential role of phages in coral health, although limitations of the data and analyses offer an exploratory examination rather than a definitive result. Overall, the evidence supporting the major findings is incomplete, in part because the …
Author response:
The following is the authors’ response to the original reviews.
eLife Assessment:
The worldwide decline in the health of coral reefs is well documented, and overgrowth by microbial consortia can be a contributing factor. Kelman and colleagues used metagenomic analysis to interrogate potential changes in phage-associated genes predicted to be involved in central carbon metabolism. The study addresses the hypothesis that metabolic genes associated with carbon metabolism that are encoded by viruses reflect the health of the corals. The study contributes a valuable perspective on the potential role of phages in coral health, although limitations of the data and analyses offer an exploratory examination rather than a definitive result. Overall, the evidence supporting the major findings is incomplete, in part because the conceptual model relies on qualitative assumptions rather than empirical data.
We want to first thank both reviewers and the editor for their comments. They were incredibly constructive and thoughtful, and provided a roadmap to a much stronger paper for us to follow. Guided by the comments, we have sought not only to address the individual points raised below, but also to expand upon them where doing so improved the clarity and accessibility of the work. The paper is now stronger, more focused, more conservative, and more robust because of the guidance provided by the reviewers. Thank you.
In addition to more point-by-point issues, the manuscript has been strengthened in three main ways:
(1) Clarified the distinction between observations, inferences, and broader ecological implications, ensuring that empirical results, conceptual synthesis, and model predictions are clearly separated throughout the manuscript.
(2) Reframed and expanded the presentation of the conceptual model, clarifying its purpose as a qualitative framework for integrating our observations with established reef ecology and for generating experimentally testable hypotheses, rather than as an additional source of empirical evidence.
(3) Clarified the analytical framework and statistical interpretation, including expanding the Methods to better explain the inferential pipeline, the interpretation of community-level metabolic associations, and the use of bootstrapped confidence intervals to quantify uncertainty around ecological effect sizes.
Although the reviewers approached the manuscript from different perspectives, we found their recommendations to be remarkably complementary. Reviewer 1 consistently encouraged us to improve the clarity, transparency, and accessibility of the manuscript, while Reviewer 2 challenged us to more carefully align the strength of our conclusions with the strength of the available evidence. Rather than addressing these concerns independently, we used them as the basis for a broader restructuring that simultaneously improved the manuscript's organization, transparency, clarity, and evidentiary discipline. We believe the result is a substantially stronger paper.
Public Reviews:
Reviewer #1 (Public review):
Summary:
Microbialization (bacterial overgrowth) is a recognized component of degraded, eutrophied coral reefs where there is a shift from coral to algal dominance on the benthos. In addition, previous work has demonstrated that virus communities shift from a lytic strategy dominated (kill-the-winner) to a temperate (lysogenic) strategy dominated with reef microbialization. Kelman et al. sought to leverage previously published virus metagenomes produced from the water column of healthy and degraded coral reefs to assess virus community metabolic shifts. The authors also produce a conceptual model to demonstrate the potential impact of the observed metabolism shifts on reef fates.
Strengths:
The main strength of the manuscript is the findings from their metagenomic analyses and results. The virus metagenomes were produced using established approaches in the field and yield sufficient data per sample for their analyses. Interesting results regarding the shift in the types of genes from anaplerotic to cataplerotic provide the foundation for testable hypotheses to determine the magnitude of impact virus strategies have on reef health. The introduction is also well written and sets up the scene very well.
Weaknesses:
(1) The methods text currently omits important information related to the sampling design. It is not clear how many metagenomes are from healthy and degraded communities. This impacts the interpretability and robustness of the statistical results. Furthermore, it is unclear if analyses are based on assembled contigs or read-based alignments. Improving the clarity and organization of the Methods is essential for reproducibility.
(2) Regarding the bioinformatics approach, normalization using the "per cent known" approach within samples may not fully account for discovery bias related to sequencing depth. While Supplementary Table 1 shows variability in read counts, the lack of community-level metadata makes it difficult to determine if sequencing depth covaries with community type (healthy vs. degraded). The study would benefit from a rarefaction analysis or subsampling to ensure that gene frequency trends and Spearman correlations are biological signals rather than artifacts of sequencing effort.
(3) The qualitative model in Figure 5 is positioned as evidence for the role of viruses in reef health, but it does not provide independent support for the authors' hypotheses. Since the model is parameterized using "arbitrary units" to reflect the authors' assumptions rather than being derived from the empirical metagenomic data, it serves as a helpful illustration of a hypothesis but not as a validation of the findings.
(4) Results and discussion require revisions to improve readability and connectivity across sections. Ensuring a clear distinction between empirical data and model-based speculation would help the audience better appreciate the science.
We thank the reviewer for their thoughtful and constructive review. Their comments were particularly helpful in identifying opportunities to improve the clarity, organization, and reproducibility of the manuscript. In response, we substantially expanded and reorganized the Methods, streamlined the Results and their connection to the Discussion, and clarified the role of the conceptual model as a framework for integrating our observations and generating testable hypotheses rather than as independent empirical evidence. We respond to the specific comments below.
Reviewer #2 (Public review):
Summary:
The manuscript by Kelman and coauthors investigates how viral communities differ in the genes they encode in healthy and degraded coral reef ecosystems. Across 19 viral metagenomes from Central Pacific reefs, the authors assess the frequency of integration/excision genes as a proxy for viral community temperateness and ask whether genes associated with central carbon metabolism covary with signatures of temperateness. The main finding is that viral communities with more temperate-related genes encode more genes from the Entner-Doudoroff pathway and other reactions interpreted as anaplerotic, whereas more lytic-associated viral communities show greater representation of some pentose phosphate pathway, TCA, and redox-associated genes interpreted as cataplerotic. The authors propose a model based on these patterns in which lytic viral metabolism helps suppress bacterial overgrowth on healthy reefs, while temperate viral metabolism may promote microbialization on degraded reefs. The study addresses an interesting and potentially important concept - that viral auxiliary metabolic genes are important components of microbial communities and can affect ecosystem functioning. Linking viral metabolism to coral reef microbialization is a creative conceptual advance. The manuscript is clearly written, and the reported enrichment of anaplerotic genes in temperate-associated viromes is an interesting pattern that could motivate future work on how viral metabolic potential varies across reef states.
Strengths:
(1) The study connects viral lifestyle, central carbon metabolism, bacterial overgrowth, and reef degradation in a framework that could be useful for future studies of coral reef ecosystems and viral ecology. This is an interesting synthesis that links viral auxiliary metabolism to broader questions about microbialization and reef state.
(2) The manuscript is generally clearly organized around a testable prediction: viral metabolic gene content should vary along a lytic-to-temperate viral community gradient. The reported enrichment of anaplerotic genes in viromes with a larger fraction of temperate viruses is a compelling result.
(3) The authors highlight several virus-encoded metabolic genes that may not have been previously reported in viral datasets or genomes. If supported by further validation, these observations could expand the known repertoire of viral metabolic potential.
(4) The modeling helps clarify the feedbacks the authors propose may connect viral lifestyle, bacterial metabolism, and coral reef degradation. It provides a foundation for generating hypotheses about how viral metabolic genes could influence reef microbial dynamics.
Weaknesses:
(1) The main limitation is that the evidence for several key claims remains indirect. The core analysis is based on correlations between metabolic gene frequencies and integration/excision-related genes. This does not demonstrate that the metabolic genes occur in temperate viral genomes, are physically linked to lysogeny genes, are expressed during infection, or alter host metabolism. Thus, the data support an association between VLP-associated metabolic annotations and a community-level temperateness proxy, but not a direct link between temperate phages and these metabolic functions.
(2) It is important not to equate community-level gene frequencies with genome-level or infection-level metabolic programs. A virome may contain more anaplerotic genes overall, but that does not demonstrate that individual viruses reprogram their hosts in an anaplerotic manner nor that infection produces a net anaplerotic effect. Individual viruses may encode both anaplerotic and cataplerotic genes, and a smaller number of cataplerotic genes could have stronger metabolic consequences depending on expression, enzyme efficiency, pathway position, and host context. This is an important limitation that should be acknowledged and, if possible, addressed with contig- or genome-level analyses.
(3) The ecological interpretation assumes that viral infection is strong enough to influence reef-scale bacterial population dynamics. However, the study does not directly measure infection frequency, lysis rates, viral production, burst size, lysogeny frequency, prophage induction, gene expression, or bacterial mortality. If viral mortality or lysogenic conversion were rare in these systems, the observed gene-frequency patterns could have limited ecosystem-level consequences. This makes claims about viral metabolism suppressing bacterial overgrowth, accelerating microbialization, or acting as a conservation lever more speculative than suggested.
(4) There are statistical limitations related to the use of relative gene frequencies. Because genes are normalized as percentages of known genes, the data are compositional. Apparent increases in some categories may partly reflect decreases in others. Bootstrapped Spearman correlations are useful for assessing the robustness of these associations, but they do not address compositionality or multiple testing.
(5) The anaplerotic/cataplerotic classification is central to the manuscript's conclusions and would benefit from more support. The framework is useful, but it depends on both annotation confidence and biochemical context. Sequence-similarity annotations alone may be vulnerable to misannotation, especially for central metabolic enzymes that share conserved domains across functionally distinct proteins. Stronger evidence that key genes contain key functional domains and/or are phylogenetically related to characterized enzymes would help support the proposed functions. In addition, many central carbon enzymes are reversible or context-dependent, so a clearer rationale for each classification would strengthen the interpretation.
Overall, the manuscript presents a valuable hypothesis and highlights new ecological patterns in coral reef viral metagenomes, but falls short of the evidence needed for the strongest claims. The work would be strengthened by analyses that directly link metabolic genes to viral genomes or lysogeny markers, address compositional effects, validate key annotations, and more clearly distinguish observed gene-frequency associations from hypothesized effects on infection, host metabolism, and reef state.
We thank the reviewer for their careful and constructive assessment. Their comments were particularly valuable in sharpening the evidentiary boundaries of the study. In response, we revised the manuscript throughout to distinguish more clearly between the community-level gene-frequency associations we observe and genome-level linkage, infection-level metabolic effects, and ecosystem-level consequences that remain hypotheses for future testing. We also revised the statistical presentation and discussion of methodological limitations to better reflect the strength and scope of the available evidence. We respond to the specific comments below.
Recommendations for the authors:
Reviewer #1 (Recommendations for the authors):
(1) Significance
General comments
Well written and highlights the findings of the study. However, the final sentence needs rewording to be clearer regarding the claims being made (see below).
Specific comments:
Lines 86-88 - I suggest revising the final sentence to better match the strength of the data presented. As currently written, the sentence makes broad causal claims that are not directly supported by the results.
For example, the phrase "viral metabolism can either reinforce reef resilience or exacerbate collapse" implies a level of mechanistic certainty that the current data do not demonstrate. The results show shifts in viral community metabolism, measured through gene abundances, that are likely associated with changes in the environment, host assemblages, and virus community diversity. These shifts may potentiate the DDAM feedback loop, but additional data and experimental work would be needed to determine whether viruses actively maintain resilience or contribute to collapse.
Similarly, the phrase "making it a hidden driver of ecosystem fate" feels too strong for the results presented. A more appropriate framing would be to present viral community metabolism as a promising and important avenue for future research into reef trajectories and ecosystem change.
Finally, the phrase "a potential target for conservation strategies" is somewhat unclear. It is not obvious whether viral metabolism is being framed as something conservation efforts should directly manipulate, or as an indicator/marker of reef condition and recovery. Clarifying this would strengthen the significance statement.
We thank the reviewer for this thoughtful and constructive suggestion. We agreed that the Significance statement occasionally conveyed greater mechanistic certainty than was supported by the observational nature of the data. Guided by these comments, we revised the section throughout to more clearly distinguish between our empirical observations, the ecological interpretations they support, and the broader hypotheses and implications they motivate. In particular, we softened causal language regarding the role of viral metabolism in reef resilience and decline, reframed the discussion of conservation to emphasize future opportunities rather than immediate application, and clarified that the manuscript provides a conceptual ecological framework for future investigation rather than direct evidence of ecosystem-scale mechanisms. We believe these revisions now more accurately reflect both the strengths and the limitations of the study while preserving its central ecological significance (lines 78-93).
(2) Abstract
Specific comments
Lines 97-102 - Add commas to improve readability:
Original: We found that viral metabolism shifted broadly from being dominated by metabolism that builds up pools of central intermediates on degraded reefs dominated by temperate viral infection ('anaplerotic' reactions) to metabolism that consumes these pools to prioritize production of metabolic precursors for virion construction on healthy reefs dominated by lytic infection ('cataplerotic' reactions).
Revised: We found that viral metabolism shifted broadly from being dominated by metabolism that builds up pools of central intermediates on degraded reefs, dominated by temperate viral infection ('anaplerotic' reactions), to metabolism that consumes these pools to prioritize production of metabolic precursors for virion construction on healthy reefs, dominated by lytic infection ('cataplerotic' reactions).
Thank you. We agreed that this sentence needed revision for clarity. In addition to improving its readability, we revised it to describe the observed pattern explicitly at the level of viral-community metabolic gene content rather than implying a demonstrated shift in metabolism itself (lines 102-107).
(3) Introduction
General comments
Very well written introduction that sets up an exciting scene!
Specific comments
Lines 134-136 - Minor revision to improve readability:
Original: "Viral predation could, in principle, preclude microbialization by killing proliferating microbes, as lytic infection and lysis rates are predicted by established predator-prey models to rise with host densities and growth rates"
Revised: "Viral predation could, in principle, preclude microbialization by killing proliferating microbes, as lytic infection and lysis rates are predicted to rise with host densities and growth rates by established predator-prey models."
Thank you. We are glad the Introduction was engaging. We have adopted this edit (lines 140-142).
(4) Methods
General comments
The methods text needs to be revised to increase readability and transparency. I would love to really appreciate the work done and the results produced, but some aspects are unclear, which leaves me with a cloudy understanding of how the data was processed and the results were generated. I am also concerned that the qualitative model makes many assumptions regarding ecological interactions between the components. It might be good to clearly state these assumptions and their impact on the outcome in a table or explicitly within the text.
We thank the reviewer for these thoughtful comments. We agreed that the original Methods section did not always provide sufficient detail to allow readers to fully understand the analytical framework or the inferential pathway from the underlying sequence data to the ecological conclusions presented. Rather than addressing these comments individually in isolation, we substantially reorganized and expanded the Methods to improve transparency, readability, and reproducibility throughout. In particular, we clarified the study design and sampling framework, explicitly distinguished between the previously generated viral metagenomic datasets and the new analyses performed in this study, and expanded our description of the bioinformatic and statistical workflows. We also more clearly explain our decision to retain the community-level, read-based analytical framework of the previous studies while extending those analyses using the MG-RAST annotation platform. By avoiding a fundamentally different analytical approach, we maintain conceptual continuity and direct comparability. Finally, we clarified the assumptions underlying the conceptual model and its intended role as a qualitative framework for hypothesis generation rather than empirical validation. Collectively, these revisions provide a clearer, more transparent, and more reproducible account of the analytical framework used throughout the study (lines 206-390; this spans the entire Methods section).
Specific comments (reordered for clarity where needed)
Line 189-190 - Sentence needs rewording. Currently, it reads a bit awkwardly, and it should be clearer that these metagenomes were previously produced and sourced from a different manuscript. It reads as if data were generated from new samples and processed as previously described, rather than a reanalysis of a dataset.
Line 189-212 - Information related to the sample design (e.g., how many metagenomes from how many reefs/reef health states) is missing and should be added in this section to provide more context for the reader.
We thank the reviewer for this helpful suggestion. We have substantially revised the opening of the Methods to clarify that this study is an analysis of previously generated virus-like particle (VLP) metagenomes, most of which had been used in prior publications, from Pacific coral reefs rather than newly generated samples. We also expanded the description of the sampling design, including the number of metagenomes analyzed and their relationship to previous studies, and explain that we retained a bioinformatic framework closely aligned with those earlier analyses to maintain methodological continuity and enable direct comparison with the published work. We agree that these additions improve the clarity, transparency, and reproducibility of the study (Intro lines 173-176, Methods lines 206-211, Supplementary Table 1).
Line 208-210 - This sentence produced concern upon reading it. If contamination is a worry and would have a direct impact on the results (especially given that results seem to be based on read alignments to a reference database), I suggest a bacterial read removal step with Kraken or alignment to a bacterial reference be performed prior to functional annotation. Given that there were multiple steps in the VLP DNA processing/extraction, there is unlikely to be a large amount of bacterial sequence reads; however, it would be good to confirm.
Lines 213-218 - It is unclear if the reads were assembled into a contig prior to annotation. In addition, in the Supplementary Table 1, please clarify if "sequences" is equal to reads or assembled contigs.
We thank the reviewer for these helpful comments. We have clarified that the analyses are based on read-level annotations rather than assembled contigs, consistent with the original virome studies and with our aim of maintaining methodological continuity with those published analyses. We also expanded the Methods to describe the multiple safeguards used to minimize cellular contamination, including 0.45-µm filtration and chloroform treatment, cesium chloride purification, DNase treatment, microscopy confirming the absence of cells, and PCR confirmation of the absence of detectable 16S rDNA amplification. While no purification protocol can completely exclude the possibility of trace contamination, these independent purification and validation steps substantially reduce this possibility. We have therefore retained the existing read-level analytical framework while making both its basis and its limitations clearer in the revised manuscript (lines 205-252).
Line 219-224 - I am concerned that allowing alignment lengths as short at 15 bp will lead to false positives. Considering the minimum read length is 100 bp, a 50 bp alignment length is a more reasonable cutoff.
We thank the reviewer for drawing our attention to this point. In reviewing the MG-RAST parameters, we identified that the original manuscript incorrectly described the minimum alignment-length setting as 15 bp. The MG-RAST parameter used in our analysis is a minimum alignment length of 15 amino acids, and we have corrected the Methods accordingly. We have also clarified that these are the default MG-RAST annotation settings and that, because they are permissive, the resulting read-level annotations are used here for relative comparisons of broad functional representation among identically processed viromes rather than to establish the function of any individual read (lines 243-252).
Lines 231-232 - The described normalization approach does not account for sequencing depth per sample since, based on how I interpret the text, you are calculating the relative abundance of genes within a sample. This does not control for the potential increase in the number of detected genes within a sample driven by higher sequencing depth. To confirm that sequencing depth does not influence results, I suggest performing a rarefaction analysis or having a subsampling step to test for the sensitivity of read depth on results.
We thank the reviewer for this thoughtful comment. Our objective was to compare the relative frequencies of annotated functional categories within viral communities rather than estimate absolute gene abundance or gene richness. We therefore normalized gene frequencies within each virome as the proportion of annotated ("known") genes, allowing direct comparison of relative community composition among metagenomes with differing sequencing depths. This normalization strategy is also consistent with our previous community-level analyses of these viromes. To make this analytical framework clearer, we have substantially expanded the Methods to explain the rationale for this normalization approach, how higher-order functional categories were constructed, and how gene frequencies were compared across samples (lines 296-316).
We also clarify a related consequence of this normalization. Because gene frequencies are expressed as proportions of a common total of annotated genes, the data are compositional, and an increase in the relative frequency of one category is necessarily accompanied by decreases in others. We have now added an explicit statement of this limitation to the Methods and clarify that the values reported here represent relative community composition rather than absolute gene abundance (lines 253–257).
Line 240 - Define "EC" -> "... Enzyme Commission (EC) numbers...". All abbreviations should be clearly defined prior to use throughout the manuscript.
Line 262 - Minor: Define the acronym "CIs" here - "Confidence intervals (CIs)..."
We thank the reviewer for identifying these omissions. We now define Enzyme Commission (EC) upon first use in the manuscript in the Methods and Results sections (lines 261 and 423). Similarly, we now define confidence intervals (CIs) when introduced in the Methods and Figure 1 legend (lines 283–286 and 777–778).
Lines 264-265 - This sentence is repeated later in the manuscript, and it's not clear what is meant as is. I would revise to make it more clear it is meant to signal that the CIs will be within brackets within the text. Or, it may be best to just have CIs like so:
Example from line 372:
Original = "(Figure 1; ρs = -0.81; [-0.95, -0.58])"
Suggested revision throughout the manuscript = "(Figure 1; ρs = -0.81; CIs: -0.95, -0.58)"
We thank the reviewer for this helpful suggestion. To improve clarity, we now explicitly state in the Methods that 95% confidence intervals are reported in square brackets following each Spearman's rho estimate (lines 285-286), and we have changed Figure 1 to be consistent with that (lines 777-778). We retained this notation because it provides a compact and consistent presentation of the effect size and its associated uncertainty.
Lines 270-273 - Unclear what is being measured, diversity of genes within samples? Or is taxonomic diversity present within samples?
Thank you for pointing this out. We have now clarified that diversity is the Shannon index (H’) using the relative frequencies of annotated functional genes in each sample (lines 290-293).
Line 276-279 - Add "to" in "...was defined aid our..."
We thank the reviewer for identifying this issue. This sentence has been revised as part of our broader effort to improve the clarity of the manuscript (lines 319-322).
(5) Results
General comments
I believe this section would benefit from a bit of reorganization to help the flow of the manuscript. Certain sections, like the "Functional homologues as viral lifestyle markers", are not descriptive enough, and it is unclear how "the divergent correlations in these genes across a viral temperateness spectrum may reflect viral life strategies" is justified. Also, several results are not discussed in the discussion.
We thank the reviewer for this helpful observation. We agree that the Results would benefit from clearer organization and stronger links between individual observations and their broader interpretation. In response, we substantially reorganized the Results to improve their logical progression. We moved explanatory material describing the analytical framework to the Methods and reorganized the remaining Results from broad community-level patterns to increasingly focused analyses of central carbon metabolism. We also revised the Discussion to ensure that the principal observations presented in the Results are revisited and interpreted within the broader ecological framework developed in the manuscript. More broadly, the revised framing now consistently presents individual observations as community-level associations and their ecological implications as hypotheses rather than definitive functional interpretations (Introduction lines 190-195, Methods lines 296-316, 317-327 and 328-330; Results lines 392-518; Discussion, lines 525-529).
We agree that the original “Functional homologues as viral lifestyle markers” section overinterpreted these community-level associations. Because the confidence intervals for the two fructose-bisphosphate aldolase homologues both span zero, we have removed this subsection rather than present these patterns as evidence for viral lifestyle markers. The individual estimates remain reported in Table 1 for completeness.
Specific comments
Lines 346-369 - Overall, this "Hallmarks of viral temperateness" text seems more appropriate for the methods section and/or a figure/table caption rather than a results section. It is unclear to me whether there are any results being presented here; instead, it reads more like a description of how to interpret upcoming results. The text should be revised to make the results more clearly presented or moved to a more appropriate section of the manuscript.
We agree, and have moved the explanatory material into the Methods/Analytical Framework and rewritten the opening of the Results so that this section now presents observations rather than instructions for interpreting subsequent analyses (Methods, lines 296-316; Results, lines 392-403).
Line 353-355 - This sentence needs revision. It feels like two similarly themed, but mismatched, sentence fragments. The first part sets up for a methods statement, but the second part defines a metric?
Sentence: "As a measure of how consistently or monotonously gene frequency and sample temperateness covary, Spearman Rho (ρs) values reflect the strength and sign of their correlation."
Maybe it is just how I am reading it, but I feel it should be revised to (or something similar):
"As a measure of how consistently or monotonously gene frequency and sample temperateness covary, a Spearman's Rank Correlation analysis was performed, generating Spearman Rho (ρs) values reflecting the strength and sign of their correlation."
We thank the reviewer for this helpful observation. We agree that the opening of the Results conflated methodological explanation with the presentation of results, resulting in awkward phrasing and reduced clarity. Rather than simply revising the sentence, we substantially reorganized this portion of the manuscript. The analytical framework, including the rationale for the temperateness proxy, the interpretation of Spearman's ρ, and the relationship between Figures 1–5, has now been moved to a dedicated Analytical framework subsection of the Methods. This allows the Results to focus directly on the empirical observations while presenting the methodological context in its more appropriate location. We believe this restructuring substantially improves the clarity and flow of the manuscript (lines 272-316 and 392-403).
Line 357 - Missing a closing parenthesis on "(ps < 0."
Line 357-358 - In this sentence, it says the 95% confidence intervals are listed in square brackets (I am assuming in Figure 1). But Figure 1 has confidence intervals in parentheses, and this is stated within the figure caption as well. Needs to be made consistent across components of the manuscript. Revisiting this comment as I read the next section of the results, I understand now what was meant regarding the brackets. This should be clarified here (and previously).
Corrected. The missing parenthesis was added during the revision and restructuring of this section, and the presentation of confidence intervals has been standardized throughout the manuscript using square brackets for consistency and clarity (lines 285-286, 394-397, and the Figure 1 legend, line 777-778).
Lines 465-500 - The text here presents model results as valid biological outcomes. There is also a bit of discussion text in this section. Given this is a qualitative model with many strong assumptions, it's best to illustrate hypotheses rather than being framed as evidence.
We thank the reviewer for this important observation. We agree that the original presentation did not always sufficiently distinguish between the empirical observations presented in this study and the qualitative hypotheses generated by the conceptual model. We therefore substantially revised both the Results and the Discussion to more clearly separate these two levels of inference. Throughout the manuscript, the conceptual model is now consistently presented as a qualitative framework for integrating our observations with established reef ecology and for generating experimentally testable hypotheses, rather than as independent empirical evidence. Correspondingly, we adopted more conservative language throughout this section to ensure that model outputs are interpreted as conceptual scenarios illustrating possible ecological consequences of the observed community-level metabolic associations, rather than as demonstrated biological outcomes. We believe these revisions more accurately reflect both the purpose of the model and the observational nature of the underlying data (Methods, lines 317-330; Results, lines 483-518; Discussion, eg line 527).
(6) Discussion
General comments
There are a lot of results explicitly mentioned in the Results section left out of the discussion. Revisions can be made to increase the connectivity between sections and streamline the story of the manuscript to help the reader appreciate the findings and implications.
We thank the reviewer for this suggestion. In revising the manuscript, we substantially reorganized the Discussion to focus on the principal findings and their broader ecological implications. The revised Discussion now revisits the major observations underlying our conclusions while distinguishing more clearly between empirical observations, ecological inferences, and conceptual hypotheses. We elected not to discuss every individual observation in detail, instead focusing the Discussion on those findings most central to the manuscript's overarching conclusions.
(7) Figures
Figure 1 - A previous comment was made regarding the mismatch between the CIs in brackets and the figure containing them in parentheses (not a fatal error, just a quick fix in the text). An additional recommendation for this figure is to make it clearer which gene correlations are significant. In the main text and within the figure caption, it states that:
"Genes where the 95 % confidence intervals exclude zero are significant at the p < 0.05 level."
Within the figure, there are a few genes that don't seem to be significant given this guideline (Cell division & cycle, Stress response, Protein metabolism). I suggest some sort of visual aid that would allow the reader to determine which are significant and which are not. Adding an asterisk to all significant correlations (or removal of non-significant) could help interpretability for the reader.
The authors may also want to consider some visual way to signal temperature-associated vs lytic-associated genes in the figure. One way may be to organize the panels into lytic- and temperate-associated groups.
Thank you for these suggestions. Confidence intervals are now presented consistently in square brackets throughout the manuscript. We substantially revised the statistical presentation throughout the manuscript (lines 285-286, 397, 777-778). Rather than emphasizing null hypothesis significance testing, we now consistently present bootstrapped median Spearman’s ρ estimates together with 95% confidence intervals to quantify uncertainty around each estimate (lines 287-289). We believe this presentation more clearly communicates the magnitude, direction, and uncertainty of the observed community-level associations. Because these confidence intervals are not adjusted for multiplicity, we elected not to add asterisks to Figure 1 or to use confidence-interval exclusion of zero as the primary basis for our conclusions. The Figure 1 legend instead states that the confidence intervals quantify uncertainty and do not represent multiple-testing-adjusted significance (lines 785-786). Asterisks retained in Table 1 and Figure 2 are provided only as visual guides and are not interpreted as formal discoveries (lines 765-766, 797-799). Instead, our interpretation emphasizes the magnitude, uncertainty, direction, and coherence of associations across predefined pathways and functional categories.
In terms of figure layout, we retained mean gene frequency ordering of the functional categories to facilitate navigation of the figure. Because Figure 1 serves primarily as a reference across a broad range of SEED Level 1 functional categories, we felt that mean gene frequency ordering provides the clearest and most accessible presentation to help contextualize our results. We now clarify the meaning of positive and negative correlations in the figure legends for Figures 1 and 2: “Positive Spearman’s ρ values indicate temperate-associated gene categories, whereas negative ρ values indicate lytic-associated gene categories” (lines 778-780, 792-794).
(8) Data availability
I tried to access data on MG-RAST servers and was unable to find them due to a server error.
We thank the reviewer for bringing this to our attention. We have also occasionally encountered intermittent server availability on MG-RAST. To ensure continued accessibility of the underlying data, the FASTA files for all metagenomes, together with the project code, annotations, and simulation files, are also available through our GitHub repository (https://github.com/hopefulmonstersucla/viral-metabolism; lines 583-590). We continue to provide the MG-RAST accession numbers as the primary archival record of the datasets, while the GitHub repository provides an alternative access route should MG-RAST be temporarily unavailable.
To end, we again thank the reviewer for their thoughtful and constructive review. Throughout the revision, their comments consistently encouraged us to improve the clarity, organization, and accessibility of the manuscript. We believe the resulting paper is substantially clearer, more transparent, and more accessible to readers, and we are grateful for the roadmap the reviewer provided.
Reviewer #2 (Recommendations for the authors):
(1) My largest concern is that the language in the manuscript moves from gene frequency correlations to claims about viral metabolism driving reef decline, supporting bacterial proliferation, and offering a conservation lever. The abstract says the findings position viral metabolism as both a driver of reef decline and a conservation lever, with "re-viralization" as a restoration strategy. I do not think the current data support that level of inference. I believe the authors should consistently and clearly distinguish between conclusions that the data support (Viral communities are associated with different metabolic gene profiles. Those genes may indicate different metabolic potentials.) and those that are speculative, not demonstrated outcomes (Those potentials may influence host metabolism during infection. Those infection-level processes may alter reef-scale microbialization.) in their writing.
We thank the reviewer for identifying what we also came to view as the central issue in revising the original manuscript. We agree that the original framing did not always clearly distinguish between our empirical observations, the ecological inferences they support, and the broader hypotheses that they motivate. Addressing this concern became a major focus of the revision. Throughout the manuscript, we now consistently distinguish between community-level observations, conceptual interpretation, and broader ecological implications. We substantially revised the Significance statement, Abstract, Introduction, Results, Discussion, and conceptual model to ensure that the strength of our conclusions reflects the strength of the underlying evidence. In particular, we now present the observed metabolic associations as community-level patterns, while framing their potential consequences for infection, host metabolism, microbialization, and reef resilience as hypotheses generated by our conceptual framework rather than demonstrated outcomes. We believe this distinction now more accurately reflects both the scope of the data and the purpose of the study.
(2) A broader issue is the presentation and interpretation of the many correlation analyses. Figures 1 and 2 contain a large number (20+) of small panels, and it is difficult to assess which relationships are most important, what the direction of each relationship is, and which results are well supported. Because the manuscript's conclusions rely on recognizing broader patterns across many categories and genes, I think these figures should do more to guide interpretation. Some combination of color and symbol could improve readability. For example, using different colors for positive vs negative correlations and a separate indicator for whether confidence intervals exclude zero or whether results remain significant after multiple-testing correction. A compact summary panel or heatmap may also help readers synthesize the overall pattern more efficiently. Relatedly, the manuscript appears to evaluate a large number of separate correlations, but I do not see a correction for multiple testing. Interpreting each correlation based on whether its 95% bootstrap confidence interval excludes zero does not address false discovery across the full set of comparisons. I recommend that the authors report the number of tests performed and apply an appropriate false discovery rate correction, such as Benjamini-Hochberg. This would make it much easier to distinguish the most robust results from more tentative patterns.
We thank the reviewer for this thoughtful suggestion. We agreed that the original presentation required clearer guidance to help readers interpret the broader patterns across the many community-level correlations. In response, we substantially revised both the analytical framework and the figure framing. The manuscript now explicitly explains the interpretation of positive and negative Spearman’s ρ values within the Methods, and the legends for Figures 1 and 2 now clarify that positive correlations indicate temperate-associated gene categories whereas negative correlations indicate lytic-associated gene categories (lines 778-780, 792-794). Figure 1 retains all functional categories and presents their bootstrapped ρ estimates and 95% confidence intervals without asterisks. Its legend states that the confidence intervals quantify uncertainty and do not represent multiple-testing-adjusted significance (lines 785-786). Table 1 and Figure 2 retain asterisks only as visual guides to intervals that exclude zero, and their legends explicitly state that these do not represent multiple-testing-adjusted significance (lines 765-766, 797-799).
With respect to multiple testing, our objective was not to identify a small subset of individually significant genes through formal hypothesis testing, but rather to characterize broad community-level metabolic patterns across predefined functional categories and pathways. Accordingly, we emphasize bootstrapped effect sizes together with their associated confidence intervals to quantify uncertainty around each estimate, rather than relying primarily on null hypothesis significance testing. We therefore retained this estimation-focused analytical framework throughout the manuscript (eg lines 287-289). Finally, because Figures 3 and 4 already synthesize the pathway- and functional-level patterns emerging from the individual correlations shown in Figures 1 and 2, we elected not to add an additional summary heatmap, as we felt this would largely duplicate information already presented.
We also thank the reviewer for raising compositionality, which we agree we had not engaged with adequately in the original manuscript. Because gene frequencies are expressed as proportions of a common total of annotated genes, the data are compositional, and an increase in the relative frequency of one category is necessarily accompanied by decreases in others. We have now added an explicit statement of this limitation to the Methods and clarify that the values reported here represent relative community composition rather than absolute gene abundance (lines 253–257).
(3) The manuscript would be strengthened by additional scrutiny of the gene annotations and the viral origin of the metabolic genes. The conclusions depend on the interpretation that specific central carbon metabolism genes are encoded by viral communities and vary with viral lifestyle. However, the current analysis appears to rely primarily on read-level annotation against SEED/MG-RAST categories. This is useful for broad screening, but I do not think it is sufficient on its own to support the stronger claims about viral metabolic strategy. I recommend that the authors add additional validation steps for the key metabolic genes that drive the manuscript's conclusions. For example, metagenomic assembly could be used to determine whether central carbon metabolism genes occur on contigs with viral hallmark genes, especially integrases, excisionases, structural genes, or other viral replication genes. This would help establish whether the metabolic genes are physically linked to viral genomes (likely, but good to independently test) and whether they are associated specifically with temperate viral genome fragments (currently unclear). If assemblies are limited by read depth or read length, the authors should report this clearly and present the analysis as a limitation. I also recommend additional functional validation of the key enzyme annotations. For the genes that support the main anaplerotic/cataplerotic conclusions, the authors could use protein-domain-based annotation, reciprocal best-hit approaches, HMM searches, or phylogenetic analyses to confirm that these reads or contigs encode the proposed enzyme functions. These additional analyses do not need to validate every annotated gene in the dataset, but they should focus on the subset of genes that carry the main conclusions and those that are the most novel. This would make the manuscript's central claims much more convincing and would help distinguish robust viral metabolic signals from more tentative read-level functional assignments.
We thank the reviewer for this thoughtful suggestion. We agree that contig-level analyses, protein-domain validation, and phylogenetic confirmation would provide valuable complementary evidence regarding the genomic context and evolutionary history of individual metabolic genes. However, these analyses address a different question from the one considered here. Our objective was to characterize community-level associations between viral metabolic gene frequencies and a previously established proxy for viral community temperateness using the same read-based analytical framework as the earlier studies on which this work builds. To make this scope clearer, we substantially expanded the Methods to explain our rationale for retaining this analytical framework and revised the manuscript throughout to ensure that our conclusions are explicitly limited to community-level metabolic associations rather than genome-level linkage or infection-level function (eg 206-208, 296-316). We now consistently present these observations as generating hypotheses regarding viral metabolic strategies that future genome-resolved studies can test directly (eg lines 185-195, 317-330, 525-529). We believe these revisions more accurately align the strength and scale of our conclusions with the evidence presented.
(4) The anaplerotic/cataplerotic classification in Table 1 is a core part of the manuscript, but the rationale for several gene assignments is not sufficiently clear. I appreciate that the authors define these terms more broadly than their conventional TCA-cycle usage. However, many of the enzymes listed in Table 1 are reversible or participate in multiple pathways, so their classification depends on assumed flux direction and physiological context rather than gene presence alone. I recommend that the authors add a supplementary table explaining the rationale for each anaplerotic/cataplerotic assignment, including the assumed reaction direction and the relevant substrate and product pools.
We thank the reviewer for this thoughtful suggestion. We agree that anaplerotic and cataplerotic classifications depend on biochemical context and do not imply that every enzyme involved is intrinsically irreversible. These terms are also commonly applied within the TCA cycle to reversible reactions when their predominant net flux, under a given physiological context, replenishes central metabolite pools (anaplerotic) or draws from those pools toward energy production or specialized biosynthetic precursors (cataplerotic). We use the terms in this broader functional sense rather than restricting them to TCA-cycle reactions alone (lines 102-107, 262-271). Accordingly, the assignments in Table 1 reflect the expected predominant role of each reaction within the pathway context represented in Figure 3 (see line 271), rather than a demonstrated in vivo flux direction during infection. Figure 3 explicitly distinguishes reversible reactions with double-headed arrows from irreversible reactions with single-headed arrows, while Table 1 presents each classification alongside its pathway, gene identity, and enzyme annotation. We therefore felt that an additional supplementary table would largely duplicate information already presented across the Methods, Table 1, and Figure 3.
(5) Minor comment: I suggest revising phrases such as "the conceptual model reveals" to something like "the numerical model suggests" (e.g. line 179).
We thank the reviewer for this helpful suggestion. Throughout the revised manuscript, we adopted more cautious language when describing the conceptual model and its outputs. We now consistently use language such as suggests, illustrates, explores, and predicts rather than wording that could imply direct empirical demonstration. These changes permeate the entire revised manuscript. These revisions are consistent with our broader effort to present the model as a qualitative conceptual framework for integrating our observations with established reef ecology and for generating experimentally testable hypotheses, rather than as an independent source of empirical evidence (lines 111-118, 185-195, 317-330, 483-518, and 569-582).
To end, we again thank the reviewer for their thoughtful and insightful review. Their comments consistently challenged us to better align the strength of our conclusions with the strength of the available evidence. As a result, we substantially strengthened the manuscript by more clearly distinguishing between empirical observations, ecological inferences, and conceptual hypotheses. We believe these revisions have made the manuscript more rigorous, more transparent, and ultimately a stronger contribution, and we are grateful for the careful consideration the reviewer gave to our work.
-
-
-
eLife Assessment
The worldwide decline in the health of coral reefs is well documented, and overgrowth by microbial consortia can be a contributing factor. Kelman and colleagues used metagenomic analysis to interrogate potential changes in phage-associated genes predicted to be involved in central carbon metabolism. The study addresses the hypothesis that metabolic genes associated with carbon metabolism that are encoded by viruses reflect the health of the corals. The study contributes a valuable perspective on the potential role of phages in coral health, although limitations of the data and analyses offer an exploratory examination rather than a definitive result. Overall, the evidence supporting the major findings is incomplete, in part because the conceptual model relies on qualitative assumptions rather than empirical data.
-
Reviewer #1 (Public review):
Summary:
Microbialization (bacterial overgrowth) is a recognized component of degraded, eutrophied coral reefs where there is a shift from coral to algal dominance on the benthos. In addition, previous work has demonstrated that virus communities shift from a lytic strategy dominated (kill-the-winner) to a temperate (lysogenic) strategy dominated with reef microbialization. Kelman et al. sought to leverage previously published virus metagenomes produced from the water column of healthy and degraded coral reefs to assess virus community metabolic shifts. The authors also produce a conceptual model to demonstrate the potential impact of the observed metabolism shifts on reef fates.
Strengths:
The main strength of the manuscript is the findings from their metagenomic analyses and results. The virus metagenomes …
Reviewer #1 (Public review):
Summary:
Microbialization (bacterial overgrowth) is a recognized component of degraded, eutrophied coral reefs where there is a shift from coral to algal dominance on the benthos. In addition, previous work has demonstrated that virus communities shift from a lytic strategy dominated (kill-the-winner) to a temperate (lysogenic) strategy dominated with reef microbialization. Kelman et al. sought to leverage previously published virus metagenomes produced from the water column of healthy and degraded coral reefs to assess virus community metabolic shifts. The authors also produce a conceptual model to demonstrate the potential impact of the observed metabolism shifts on reef fates.
Strengths:
The main strength of the manuscript is the findings from their metagenomic analyses and results. The virus metagenomes were produced using established approaches in the field and yield sufficient data per sample for their analyses. Interesting results regarding the shift in the types of genes from anaplerotic to cataplerotic provide the foundation for testable hypotheses to determine the magnitude of impact virus strategies have on reef health. The introduction is also well written and sets up the scene very well.
Weaknesses:
(1) The methods text currently omits important information related to the sampling design. It is not clear how many metagenomes are from healthy and degraded communities. This impacts the interpretability and robustness of the statistical results. Furthermore, it is unclear if analyses are based on assembled contigs or read-based alignments. Improving the clarity and organization of the Methods is essential for reproducibility.
(2) Regarding the bioinformatics approach, normalization using the "percent known" approach within samples may not fully account for discovery bias related to sequencing depth. While Supplementary Table 1 shows variability in read counts, the lack of community-level metadata makes it difficult to determine if sequencing depth covaries with community type (healthy vs. degraded). The study would benefit from a rarefaction analysis or subsampling to ensure that gene frequency trends and Spearman correlations are biological signals rather than artifacts of sequencing effort.
(3) The qualitative model in Figure 5 is positioned as evidence for the role of viruses in reef health, but it does not provide independent support for the authors' hypotheses. Since the model is parameterized using "arbitrary units" to reflect the authors' assumptions rather than being derived from the empirical metagenomic data, it serves as a helpful illustration of a hypothesis but not as a validation of the findings.
(4) Results and discussion require revisions to improve readability and connectivity across sections. Ensuring a clear distinction between empirical data and model-based speculation would help the audience better appreciate the science.
-
Reviewer #2 (Public review):
Summary:
The manuscript by Kelman and coauthors investigates how viral communities differ in the genes they encode in healthy and degraded coral reef ecosystems. Across 19 viral metagenomes from Central Pacific reefs, the authors assess the frequency of integration/excision genes as a proxy for viral community temperateness and ask whether genes associated with central carbon metabolism covary with signatures of temperateness. The main finding is that viral communities with more temperate-related genes encode more genes from the Entner-Doudoroff pathway and other reactions interpreted as anaplerotic, whereas more lytic-associated viral communities show greater representation of some pentose phosphate pathway, TCA, and redox-associated genes interpreted as cataplerotic. The authors propose a model based on …
Reviewer #2 (Public review):
Summary:
The manuscript by Kelman and coauthors investigates how viral communities differ in the genes they encode in healthy and degraded coral reef ecosystems. Across 19 viral metagenomes from Central Pacific reefs, the authors assess the frequency of integration/excision genes as a proxy for viral community temperateness and ask whether genes associated with central carbon metabolism covary with signatures of temperateness. The main finding is that viral communities with more temperate-related genes encode more genes from the Entner-Doudoroff pathway and other reactions interpreted as anaplerotic, whereas more lytic-associated viral communities show greater representation of some pentose phosphate pathway, TCA, and redox-associated genes interpreted as cataplerotic. The authors propose a model based on these patterns in which lytic viral metabolism helps suppress bacterial overgrowth on healthy reefs, while temperate viral metabolism may promote microbialization on degraded reefs. The study addresses an interesting and potentially important concept - that viral auxiliary metabolic genes are important components of microbial communities and can affect ecosystem functioning. Linking viral metabolism to coral reef microbialization is a creative conceptual advance. The manuscript is clearly written, and the reported enrichment of anaplerotic genes in temperate-associated viromes is an interesting pattern that could motivate future work on how viral metabolic potential varies across reef states.
Strengths:
(1) The study connects viral lifestyle, central carbon metabolism, bacterial overgrowth, and reef degradation in a framework that could be useful for future studies of coral reef ecosystems and viral ecology. This is an interesting synthesis that links viral auxiliary metabolism to broader questions about microbialization and reef state.
(2) The manuscript is generally clearly organized around a testable prediction: viral metabolic gene content should vary along a lytic-to-temperate viral community gradient. The reported enrichment of anaplerotic genes in viromes with a larger fraction of temperate viruses is a compelling result.
(3) The authors highlight several virus-encoded metabolic genes that may not have been previously reported in viral datasets or genomes. If supported by further validation, these observations could expand the known repertoire of viral metabolic potential.
(4) The modeling helps clarify the feedbacks the authors propose may connect viral lifestyle, bacterial metabolism, and coral reef degradation. It provides a foundation for generating hypotheses about how viral metabolic genes could influence reef microbial dynamics.
Weaknesses:
(1) The main limitation is that the evidence for several key claims remains indirect. The core analysis is based on correlations between metabolic gene frequencies and integration/excision-related genes. This does not demonstrate that the metabolic genes occur in temperate viral genomes, are physically linked to lysogeny genes, are expressed during infection, or alter host metabolism. Thus, the data support an association between VLP-associated metabolic annotations and a community-level temperateness proxy, but not a direct link between temperate phages and these metabolic functions.
(2) It is important not to equate community-level gene frequencies with genome-level or infection-level metabolic programs. A virome may contain more anaplerotic genes overall, but that does not demonstrate that individual viruses reprogram their hosts in an anaplerotic manner nor that infection produces a net anaplerotic effect. Individual viruses may encode both anaplerotic and cataplerotic genes, and a smaller number of cataplerotic genes could have stronger metabolic consequences depending on expression, enzyme efficiency, pathway position, and host context. This is an important limitation that should be acknowledged and, if possible, addressed with contig- or genome-level analyses.
(3) The ecological interpretation assumes that viral infection is strong enough to influence reef-scale bacterial population dynamics. However, the study does not directly measure infection frequency, lysis rates, viral production, burst size, lysogeny frequency, prophage induction, gene expression, or bacterial mortality. If viral mortality or lysogenic conversion were rare in these systems, the observed gene-frequency patterns could have limited ecosystem-level consequences. This makes claims about viral metabolism suppressing bacterial overgrowth, accelerating microbialization, or acting as a conservation lever more speculative than suggested.
(4) There are statistical limitations related to the use of relative gene frequencies. Because genes are normalized as percentages of known genes, the data are compositional. Apparent increases in some categories may partly reflect decreases in others. Bootstrapped Spearman correlations are useful for assessing the robustness of these associations, but they do not address compositionality or multiple testing.
(5) The anaplerotic/cataplerotic classification is central to the manuscript's conclusions and would benefit from more support. The framework is useful, but it depends on both annotation confidence and biochemical context. Sequence-similarity annotations alone may be vulnerable to misannotation, especially for central metabolic enzymes that share conserved domains across functionally distinct proteins. Stronger evidence that key genes contain key functional domains and/or are phylogenetically related to characterized enzymes would help support the proposed functions. In addition, many central carbon enzymes are reversible or context-dependent, so a clearer rationale for each classification would strengthen the interpretation.
Overall, the manuscript presents a valuable hypothesis and highlights new ecological patterns in coral reef viral metagenomes, but falls short of the evidence needed for the strongest claims. The work would be strengthened by analyses that directly link metabolic genes to viral genomes or lysogeny markers, address compositional effects, validate key annotations, and more clearly distinguish observed gene-frequency associations from hypothesized effects on infection, host metabolism, and reef state.
-
-
Central carbon metabolism switching in lytic versus temperate coral reef viral communities https://www.biorxiv.org/content/10.1101/2025.09.25.676695v1
-
Central carbon metabolism switching in lytic versus temperate coral reef viral communities https://www.biorxiv.org/content/10.1101/2025.09.25.676695v1
-