The overall conclusions of this study – sulfate-dependent AOM by ANME archaea in the seep sediments, and slow organic matter remineralization at the background sites – are plausible, but the manuscript as a whole makes a somewhat improvised impression and many datasets are not presented to their best advantage. The metagenomic and transcriptomic analysis remains at a fairly generic level and does not comment on interesting results (for example the Chloroflexota dsr genes); the phylogenetic data are used to very limited effect. Much more is possible by going beyond phylum-level generalizations; it is very hard to say anything meaningful if the analysis and discussion remain stuck on this level. Phylogenies exist, and they should be used and explored.

To summarize, the manuscript needs more work and additional analyses to sharpen the conclusions.

Manuscript comments

Unpleasant surprises in the Introduction:

Line 35 and ff. The presence of marker genes reveals potential, not activity. For the latter, you would need transcriptional data.

Lines 102 to 105: Chloroflexota are sulfate reducers? Sulfate reducers (Desulfobacterota) use simple and halogenated HCs as electron acceptors (not sulfate?). The involvement of the Asgards in methane and HC degradation is also highly debatable and very likely applies to specialized lineages only; generalizations across a phylum (or superphylum, see Eme et al. 2023 Nature) are not helpful. These lines need to be rewritten and disentangled, to avoid nonsensical statements.

Line 290: Did you obtain and analyze gene transcripts (mRNA) ? Apparently yes, but this needs to be introduced more clearly.

Lines 306/7: Why are these groups not “statistically significant”? How was this tested?

Concerning figures 2 and 3, I would bring the original porewater profiles from the supplements back into the main manuscript, to demonstrate that the profiles are in fact quite linear. Or make a suitable selection from the original profiles in a nicely designed full-page figure (“Representative profiles from …”) so that the reader can see them without having to check the supplements. There is no reason to hide the real data in the supplements. After all, the central argument of this manuscript depends on these profiles!

The Metagenomics section bumps into some very interesting questions, and more or less ignores them. Line 410 ff: Are the Chloroflexota dsr genes functioning in the oxidative or reductive direction? Do they function in assimilatory or dissimilatory sulfate reduction? To which group of the Chloroflexota (a huge and highly diverse phylum!) do they belong? Assimilatory sulfate reduction is documented for Chloroflexota: Zheng R, Wang C, Sun C. 2024. Deep-sea in situ and laboratory multi-omics provide insights into the sulfur assimilation of a deep-sea Chloroflexotabacterium. mBio 15:e00004-24. https://doi.org/10.1128/mbio.00004-24

The Figure 6. Carbon dioxide per se is not abundant in normally buffered seawater; DIC occurs mostly as bicarbonate and some carbonate anions, and also microbially produced CO2 will enter the carbonate equilibrium (and thus magically disappear if you look for CO2 only). This is the likely reason why the HC-rich samples and the background sediments do not show a significant difference in CO2 concentrations.

Figure 7. Are you using “alkalinity flux” in the sense of total DIC flux (CO2 plus bicarbonate plus carbonate), or are you thinking of seawater alkalinity (which includes buffering contributions from other seawater compounds as well)?

Figure 8 is unreadable. It is impossible to tell which colors belong to archaea and to bacteria (for example, ANME has the same shade of red as Nitrospirota). Try another solution, either by providing contrasting colors to bacteria and archaea (blue, green and yellow for bacteria; red and purple for archaea), or design separate plots for bacteria and archaea.

Line 470: Do you have 13C data for methane to distinguish biogenic from deeply-sourced (thermogenic) methane? Is there anything usable about Barents Sea methane in the literature?

Line 480: How does sulfide production stimulate the Asgards? This is again an example of overly generic statements that really mean nothing. Remember that the Asgards consist of multiple phylum-level lineages (Eme et al. 2023 Nature); is it likely that they will all behave in the same way when tickled by sulfide?

Line 485: What are the geochemical features that can be used to tell apart seep and background sites? The discussion always returns to sulfate in its various guises (penetration depth, SMTZ …). Low organic carbon content is discussed, not as a diagnostic feature but as a factor that helps HC detection by keeping the heterotrophic background down.

Line 500 ff: the enormous variability of SMTZ depth in continental margin sediments argues against using this criterion for identifying slow seep areas

Lines 510 ff: these discussion paragraphs come across as somewhat generic. For example, a more careful analysis of the metagenomic data could help to identify whether DIC removal or sulfate recycling via sulfide oxidation are more likely. For example, do you have dsr genes that function in the oxidative direction? These can be told apart from their reductive cousins (Dahl et al. 2005. J Bacteriol. 187(4):1392-404). Also, what is known about the redox status of these Barents Sea sediments? Oxygen, nitrate, metals – anything that could serve as an electron sink for sulfur oxidation?

A concluding note about the phylogenetic gene trees in the supplements – they are barely mentioned in the manuscript and not really used for anything. However, they could demonstrate phylogenetic affinity for particular groups and lineages within major phyla, and thus they could sharpen the discussion beyond generic phylum-only generalizations. Check carefully where exactly you are in phylospace, and do not rely on Genbank-only annotation.

--- end of review ---

Schnabel and co-authors explored the impact of minor HC seepage on sediment sulfur cycling through generating a variety of datasets, including porewater geochemistry, reaction rate calculation, metagenomic and metatranscriptomic data. I admit that considerable efforts have been made to collect so many sediment cores and produce such comprehensive datasets. However, the present paper fails to formulate an informative and sound story; it reads more like a report rather than a scientific article. With that said, the data have not been deeply digested and synthesized. As a non-expert on microbiology, I can see that the microbiological data seem not to be well utilized and explored. The other flaw lies in the very limited novel perspectives provided by this study.

Specific comments:

1. The title reads a little confusing. I don’t see any discussion on the implication for HC reservoir detection.
2. The Abstract is a little verbose; it is usually limited to one paragraph.
3. HC positive site? How about HC bearing site?
4. Line 94: The depth of SMTZ at ~100 m in the continental margin sediments is uncommon.
5. Line 154: what molecular analyses?
6. Figures 2&3: I would suggest plotting the measured data instead of the extrapolated profiles. Maybe pick several representative profiles rather than the whole dataset.
7. Figure 4 can be combined with Figures 2&3.
8. Figure 6. As CH4 and CO2 are not key data and they are subject to sampling artefact, they can be moved to the supplement.
9. Line 448: Higher occurrence? Weird wording. Also I don’t see higher sulfate reduction rate in the deeper sediments from Fig. 5.
10. Line 508: No sulfate profile in Figure 5.
11. Line 514: It is more likely that the lower alkalinity flux is attributed to HCO3- removal by authigenic carbonate precipitation, which can be demonstrated by Ca and Mg data if they are available.
12. Line 523: I think the latter hypothesis is more likely. I doubt the reoxidation of sulfide can contribute large variation of sulfate flux. My understanding is that sulfate derived from the reoxidation of sulfide is rapidly used via sulfate reduction rather than remaining in the porewater.